From pgsql-performance-owner@postgresql.org Thu Jan 1 13:15:00 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id B5334D1BB71 for ; Thu, 1 Jan 2004 17:14:58 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 24547-04 for ; Thu, 1 Jan 2004 13:14:10 -0400 (AST) Received: from mail2000-4.so-net.net.tw (mail2000-4.so-net.net.tw [61.64.127.20]) by svr1.postgresql.org (Postfix) with SMTP id 70590D1B498 for ; Thu, 1 Jan 2004 13:14:08 -0400 (AST) Received: from 10.1.9.28 by mail2000-4.so-net.net.tw with Mail2000 ESMTP Server V2.71S(31478:0:AUTH_RELAY) Fri, 02 Jan 2004 01:14:06 +0800 (CST); (envelope-from ) Received: By OpenMail Mailer;Fri, 02 Jan 2004 01:14:02 +0800 (CST) From: "cnliou" Reply-To: cnliou@so-net.net.tw Subject: pg_restore makes disk busy Message-ID: <1072977242.87202.cnliou@so-net.net.tw> To: "" Date: Fri, 02 Jan 2004 01:14:02 +0800 (CST) MIME-Version: 1.0 Content-Type: text/plain X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=0.0 tagged_above=0.0 required=5.0 tests= X-Spam-Level: X-Archive-Number: 200401/1 X-Sequence-Number: 5176 Happy new year! When performing "pg_restore -L list --disable-triggers -d db1 -v my_archive" , my hard disk for Linux box (with 96MB RAM) becomes extremely busy. One example is that it takes more than 5 miniutes to restore for a table from 7800 rows. Each row has less than 117 bytes in length with total of 6 columns. Hence I think the amount of the to-restore data is not the cause of performance problem. The swap size is only 68K. Therefore, I don't think small amount of RAM is a problem, either. kjournald uses 2% CPU and postmaster uses 10%. CPU is about 95% idle. What makes the restore so slow? How do I speed it up? Regards, CN From pgsql-performance-owner@postgresql.org Thu Jan 1 23:34:49 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 6ED80D1B436 for ; Fri, 2 Jan 2004 03:34:46 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 90045-04 for ; Thu, 1 Jan 2004 23:34:01 -0400 (AST) Received: from pomeray.duluoz.net (12-241-192-70.client.attbi.com [12.241.192.70]) by svr1.postgresql.org (Postfix) with ESMTP id A7B59D1B48E for ; Thu, 1 Jan 2004 23:33:56 -0400 (AST) Received: (qmail 1976 invoked by uid 1000); 2 Jan 2004 03:34:01 -0000 Date: Thu, 1 Jan 2004 19:34:01 -0800 From: Mike Glover To: pgsql-performance@postgresql.org Subject: Very slow update + not using clustered index Message-Id: <20040101193401.6640420b.mpg4@duluoz.net> X-Mailer: Sylpheed version 0.9.0claws (GTK+ 1.2.10; i686-pc-linux-gnu) Mime-Version: 1.0 Content-Type: multipart/signed; protocol="application/pgp-signature"; micalg="pgp-sha1"; boundary="=.lOrWFPH0Z)WqCU" X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/2 X-Sequence-Number: 5177 --=.lOrWFPH0Z)WqCU Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit I have these two tables: Table "de.summary" Column | Type | Modifiers --------------+-----------------------------+--------------- isbn | character varying(10) | not null source | character varying(20) | not null condition | smallint | availability | smallint | price_list | numeric(11,2) | price_min | numeric(11,2) | last_update | timestamp without time zone | default now() Indexes: "summary_pkey" primary key, btree (isbn, source) Table "de.inventory" Column | Type | Modifiers --------------+-----------------------+----------- isbn | character varying(10) | condition | integer | availability | integer | price | numeric(9,2) | Indexes: "inventory_isbn_idx" btree (isbn) Both tables are clustered on their respective indexes. The entire database has been freshly VACUUM FULL'd and ANALYZE'd (after clustering). I want to run the following query, but it takes a *very* long time. Like this: bookshelf=> explain analyze update summary set price_min=0, availability=2, condition=9 where isbn = inventory.isbn and price_min = inventory.price; QUERY PLAN ----------------------------------------------------------------- ----------------------------------------------------------- Merge Join (cost=496170.66..517271.50 rows=5051 width=51) (actual time=226940.723..292247.643 rows=419277 loops=1) Merge Cond: (("outer".price_min = "inner".price) AND ("outer"."?column7?" = "inner"."?column3?")) -> Sort (cost=366877.05..371990.05 rows=2045201 width=61) (actual time=162681.929..177216.158 rows=2045200 loops=1) Sort Key: summary.price_min, (summary.isbn)::text -> Seq Scan on summary (cost=0.00..44651.01 rows=2045201 width=61) (actual time=8.139..22179.379 rows=2045201 loops=1) -> Sort(cost=129293.61..131499.09 rows=882192 width=25) (actual time=64213.663..67563.175 rows=882192 loops=1) Sort Key: inventory.price, (inventory.isbn)::text -> Seq Scan on inventory(cost=0.00..16173.92 rows=882192 width=25)(actual time=5.773..21548.942 rows=882192 loops=1) Total runtime: 3162319.477 ms(9 rows) Running what I believe to be the comparable select query is more reasonable: bookshelf=> explain analyze select s.* from summary s, inventory i where s.isbn = i.isbn and s.price_min = i.price; QUERY PLAN ----------------------------------------------------------------------- Merge Join (cost=495960.66..517061.50 rows=5051 width=59) (actual time=194048.974..215835.982 rows=419277 loops=1) Merge Cond: (("outer".price_min = "inner".price) AND ("outer"."?column8?" ="inner"."?column3?")) -> Sort (cost=366667.05..371780.05 rows=2045201 width=59) (actual time=147678.109..149945.170 rows=2045200 loops=1) Sort Key: s.price_min, (s.isbn)::text -> Seq Scan on summary s (cost=0.00..49431.01 rows=2045201 width=59) (actual time=0.056..9304.803 rows=2045201 loops=1) -> Sort (cost=129293.61..131499.09 rows=882192 width=25) (actual time=46338.696..47183.739 rows=882192 loops=1) Sort Key: i.price, (i.isbn)::text -> Seq Scan on inventory i (cost=0.00..16173.92 rows=882192 width=25) (actual time=0.089..2419.187 rows=882192 loops=1) Total runtime: 216324.171 ms I had figured that the tables would get sorted on isbn, because of the clustering. I understand why price might get chosen (fewer matches), but the planner seems to be making the wrong choice: bookshelf=> explain analyze select s.* from summary s, inventory i where s.isbn = i.isbn; QUERY PLAN ----------------------------------------------------------------------- Merge Join (cost=489500.66..512953.69 rows=882192 width=59) (actual time=152247.741..174408.812 rows=882192 loops=1) Merge Cond: ("outer"."?column8?" = "inner"."?column2?") -> Sort (cost=366667.05..371780.05 rows=2045201 width=59) (actual time=118562.097..120817.894 rows=2045146 loops=1) Sort Key:(s.isbn)::text -> Seq Scan on summary s (cost=0.00..49431.01 rows=2045201 width=59) (actual time=0.062..8766.683 rows=2045201 loops=1) -> Sort (cost=122833.61..125039.09 rows=882192 width=14)(actual time=33685.455..34480.190 rows=882192 loops=1) Sort Key:(i.isbn)::text -> Seq Scan on inventory i (cost=0.00..16173.92 rows=882192 width=14) (actual time=0.088..1942.173 rows=882192 loops=1) Total runtime: 174926.115 ms So, my first question is: why is the planner still sorting on price when isbn seems (considerably) quicker, and how can I force it to sort by isbn(if I even should)? The second question is: why, oh why does the update take such and obscenely long time to complete? The 175s (and even 216s) for the select seems reasonable given the size of the tables, but not 3000s to update the same rows. The processor (AMD 1.3GHz) is 90%+ utilization for most of the execution time. I can post more information if it would be helpful, but this post is long enough already. TIA, and happy new year. -mike -- Mike Glover Key ID BFD19F2C --=.lOrWFPH0Z)WqCU Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.3 (GNU/Linux) iD8DBQE/9OapZrNpxr/RnywRAqCnAJ9A70H01+oIh9mSldPhtpbliawQawCcD4lB n9Tg01I8ywLkBeQFjiSIupk= =VcSD -----END PGP SIGNATURE----- --=.lOrWFPH0Z)WqCU-- From pgsql-performance-owner@postgresql.org Fri Jan 2 00:07:20 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id CDBBED1B453 for ; Fri, 2 Jan 2004 04:07:18 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 89890-08 for ; Fri, 2 Jan 2004 00:06:28 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 2B157D1B436 for ; Fri, 2 Jan 2004 00:06:28 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i0246B19002967; Thu, 1 Jan 2004 23:06:11 -0500 (EST) To: Mike Glover Cc: pgsql-performance@postgresql.org Subject: Re: Very slow update + not using clustered index In-reply-to: <20040101193401.6640420b.mpg4@duluoz.net> References: <20040101193401.6640420b.mpg4@duluoz.net> Comments: In-reply-to Mike Glover message dated "Thu, 01 Jan 2004 19:34:01 -0800" Date: Thu, 01 Jan 2004 23:06:11 -0500 Message-ID: <2966.1073016371@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/3 X-Sequence-Number: 5178 Mike Glover writes: > I want to run the following query, but it takes a *very* long time. > Like this: > bookshelf=> explain analyze update summary set price_min=0, > availability=2, condition=9 where isbn = inventory.isbn and price_min = > inventory.price; > ... > Total runtime: 3162319.477 ms(9 rows) > Running what I believe to be the comparable select query is more > reasonable: > bookshelf=> explain analyze select s.* from summary s, inventory i where > s.isbn = i.isbn and s.price_min = i.price; > ... > Total runtime: 216324.171 ms AFAICS these plans are identical, and therefore the difference in runtime must be ascribed to the time spent actually doing the updates. It seems unlikely that the raw row inserts and updating the single index could be quite that slow --- perhaps you have a foreign key or trigger performance problem? > So, my first question is: why is the planner still sorting on price when > isbn seems (considerably) quicker, and how can I force it to sort by > isbn(if I even should)? Is this PG 7.4? It looks to me like the planner *should* consider both possible orderings of the mergejoin sort keys. I'm not sure that it knows enough to realize that the key with more distinct values should be put first, however. A quick experiment shows that if the planner does not have any reason to prefer one ordering over another, the current coding will put the last WHERE clause first: regression=# create table t1(f1 int, f2 int); CREATE TABLE regression=# create table t2(f1 int, f2 int); CREATE TABLE regression=# explain select * from t1,t2 where t1.f1=t2.f1 and t1.f2=t2.f2; QUERY PLAN ------------------------------------------------------------------------- Merge Join (cost=139.66..154.91 rows=25 width=16) Merge Cond: (("outer".f2 = "inner".f2) AND ("outer".f1 = "inner".f1)) -> Sort (cost=69.83..72.33 rows=1000 width=8) Sort Key: t1.f2, t1.f1 -> Seq Scan on t1 (cost=0.00..20.00 rows=1000 width=8) -> Sort (cost=69.83..72.33 rows=1000 width=8) Sort Key: t2.f2, t2.f1 -> Seq Scan on t2 (cost=0.00..20.00 rows=1000 width=8) (8 rows) regression=# explain select * from t1,t2 where t1.f2=t2.f2 and t1.f1=t2.f1; QUERY PLAN ------------------------------------------------------------------------- Merge Join (cost=139.66..154.91 rows=25 width=16) Merge Cond: (("outer".f1 = "inner".f1) AND ("outer".f2 = "inner".f2)) -> Sort (cost=69.83..72.33 rows=1000 width=8) Sort Key: t1.f1, t1.f2 -> Seq Scan on t1 (cost=0.00..20.00 rows=1000 width=8) -> Sort (cost=69.83..72.33 rows=1000 width=8) Sort Key: t2.f1, t2.f2 -> Seq Scan on t2 (cost=0.00..20.00 rows=1000 width=8) (8 rows) and so you could probably improve matters just by switching the order of your WHERE clauses. Of course this answer will break as soon as anyone touches any part of the related code, so I'd like to try to fix it so that there is actually a principled choice made. Could you send along the pg_stats rows for these columns? > The second question is: why, oh why does the update take such and > obscenely long time to complete? See above --- the problem is not within the plan, but must be sought elsewhere. regards, tom lane From pgsql-performance-owner@postgresql.org Sat Jan 3 02:50:38 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id DE1D4D1C956 for ; Sat, 3 Jan 2004 06:50:35 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 31700-05 for ; Sat, 3 Jan 2004 02:49:47 -0400 (AST) Received: from relay01.kbs.net.au (relay01.kbs.net.au [203.220.32.149]) by svr1.postgresql.org (Postfix) with ESMTP id C94C9D1C919 for ; Sat, 3 Jan 2004 02:49:45 -0400 (AST) Received: from [203.221.247.106] (helo=familyhealth.com.au) by relay01.kbs.net.au with esmtp (Exim 3.36 #2) id 1AcHN3-0007wH-00; Fri, 02 Jan 2004 15:57:14 +1100 Message-ID: <3FF4FA25.7050802@familyhealth.com.au> Date: Fri, 02 Jan 2004 12:57:09 +0800 From: Christopher Kings-Lynne User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.6b) Gecko/20031205 Thunderbird/0.4 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Vivek Khera Cc: pgsql-performance@postgresql.org Subject: Re: deferred foreign keys References: In-Reply-To: Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/15 X-Sequence-Number: 5190 > One more question: does the FK checker know to skip checking a > constraint if the column in question did not change during an update? > > That is, if I have a user table that references an owner_id in an > owners table as a foreign key, but I update fields other than owner_id > in the user table, will it still try to verify that owner_id is a > valid value even though it did not change? > > I'm using PG 7.4. As of 7.4, yes the check is skipped. Chris From pgsql-performance-owner@postgresql.org Fri Jan 2 02:17:28 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D54DFD1B499 for ; Fri, 2 Jan 2004 06:17:26 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 03672-04 for ; Fri, 2 Jan 2004 02:16:37 -0400 (AST) Received: from pomeray.duluoz.net (12-241-192-70.client.attbi.com [12.241.192.70]) by svr1.postgresql.org (Postfix) with ESMTP id E8D3ED1B4A1 for ; Fri, 2 Jan 2004 02:16:35 -0400 (AST) Received: (qmail 2133 invoked by uid 1000); 2 Jan 2004 06:16:34 -0000 Date: Thu, 1 Jan 2004 22:16:30 -0800 From: Mike Glover To: Tom Lane Cc: pgsql-performance@postgresql.org Subject: Re: Very slow update + not using clustered index Message-Id: <20040101221630.4b001c35.mpg4@duluoz.net> In-Reply-To: <2966.1073016371@sss.pgh.pa.us> References: <20040101193401.6640420b.mpg4@duluoz.net> <2966.1073016371@sss.pgh.pa.us> X-Mailer: Sylpheed version 0.9.0claws (GTK+ 1.2.10; i686-pc-linux-gnu) Mime-Version: 1.0 Content-Type: multipart/signed; protocol="application/pgp-signature"; micalg="pgp-sha1"; boundary="=.:X5DcK3,R.IM5H" X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/4 X-Sequence-Number: 5179 --=.:X5DcK3,R.IM5H Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Tom- Thanks for the quick response. More details are inline. -mike On Thu, 01 Jan 2004 23:06:11 -0500 Tom Lane wrote: > Mike Glover writes: > AFAICS these plans are identical, and therefore the difference in > runtime must be ascribed to the time spent actually doing the updates. > It seems unlikely that the raw row inserts and updating the single > index could be quite that slow --- perhaps you have a foreign key > or trigger performance problem? There are no foreign keys or triggers for either of the tables. > Is this PG 7.4? Yes, PG 7.4 > > A quick experiment shows that if the planner does not have any reason > to prefer one ordering over another, the current coding will put the > last WHERE clause first: [snip]> > and so you could probably improve matters just by switching the order > of your WHERE clauses. Of course this answer will break as soon as > anyone touches any part of the related code, so I'd like to try to fix > it so that there is actually a principled choice made. Could you send > along the pg_stats rows for these columns? > It looks like the planner is already making a principled choice: bookshelf=> explain select s.* from summary s, inventory i where s.isbn = i.isbn and s.price_min = i.price; QUERY PLAN ----------------------------------------------------------------------- Merge Join (cost=491180.66..512965.72 rows=9237 width=58) Merge Cond: (("outer".price_min = "inner".price) AND ("outer"."?column8?" = "inner"."?column3?")) -> Sort (cost=361887.05..367000.05 rows=2045201 width=58) Sort Key: s.price_min, (s.isbn)::text -> Seq Scan on summary s (cost=0.00..44651.01 rows=2045201 width=58) -> Sort (cost=129293.61..131499.09 rows=882192 width=25) Sort Key: i.price, (i.isbn)::text -> Seq Scan on inventory i (cost=0.00..16173.92 rows=882192 width=25) (8 rows) bookshelf=> explain select s.* from summary s, inventory i where s.price_min = i.price and s.isbn = i.isbn; QUERY PLAN ----------------------------------------------------------------------- Merge Join (cost=491180.66..512965.72 rows=9237 width=58) Merge Cond: (("outer".price_min = "inner".price) AND ("outer"."?column8?" ="inner"."?column3?")) -> Sort (cost=361887.05..367000.05 rows=2045201 width=58) Sort Key: s.price_min, (s.isbn)::text -> Seq Scan on summary s (cost=0.00..44651.01 rows=2045201 width=58) -> Sort(cost=129293.61..131499.09 rows=882192 width=25) Sort Key: i.price, (i.isbn)::text -> Seq Scan on inventory i (cost=0.00..16173.92 rows=882192 width=25) (8 rows) Here are the pg_stats rows: bookshelf=> select * from pg_stats where schemaname='de' and tablename='inventory' and attname='isbn'; schemaname | tablename | attname | null_frac | avg_width | n_distinct | most_common_vals | most_common_freqs | histogram_bounds | correlation ------------+-----------+---------+-----------+-----------+------------ +------------------+-------------------+------------------------------- ----------------------------------------------------------------------- ----------------------+------------- de | inventory | isbn | 0 | 14 | -1 | | | {0002551543,0198268211,0375507299,0486231305,0673395197,0767901576,0810 304430,0865738890,0931595029,1574160052,9971504014} | 1(1 row) bookshelf=> select * from pg_stats where schemaname='de' and tablename='inventory' and attname='price'; schemaname | tablename | attname | null_frac | avg_width | n_distinct | most_common_vals | most_common_freqs | histogram_bounds | correlation ------------+-----------+---------+-----------+-----------+------------ +--------------------------------------------------------------+------- ----------------------------------------------------------------------- -----------------------+----------------------------------------------- --------------------------+------------- de | inventory | price | 0 | 11 | 1628 | {59.95,0.00,54.88,53.30,60.50,64.25,73.63,49.39,50.02,53.37} | {0.259667,0.00633333,0.00533333,0.00466667,0.00466667,0.00466667,0.0046 6667,0.00433333,0.004,0.004} | {49.16,52.06,55.53,59.56,63.78,68.90,76.90,88.53,106.16,143.75,1538.88} | 0.149342(1 row) bookshelf=> select * from pg_stats where schemaname='de' and tablename='summary' and attname='isbn'; schemaname | tablename | attname | null_frac | avg_width | n_distinct | most_common_vals | most_common_freqs | histogram_bounds | correlation ------------+-----------+---------+-----------+-----------+------------ +------------------+-------------------+------------------------------- ----------------------------------------------------------------------- ----------------------+------------- de | summary | isbn | 0 | 14 | -1 | | | {0001984209,020801912X,0395287693,055214911X,0722525915,0787630896,0822 218100,0883856263,1413900275,1843910381,9999955045} | 1(1 row) bookshelf=> select * from pg_stats where schemaname='de' and tablename='summary' and attname='price_min'; schemaname | tablename | attname | null_frac | avg_width | n_distinct | most_common_vals | most_common_freqs | histogram_bounds | correlation ------------+-----------+-----------+-----------+-----------+---------- --+---------------------------------------------------------+---------- ----------------------------------------------------------------------- -------------------+--------------------------------------------------- ------------------+------------- de | summary | price_min | 0 | 10 | 1532 | {0.00,59.95,6.95,6.00,4.07,10.17,11.53,10.85,4.75,8.81} | {0.425333,0.029,0.0193333,0.00533333,0.00333333,0.00333333,0.00333333,0 .003,0.00266667,0.00266667} | {0.05,7.11,10.30,14.28,19.54,27.86,50.47,61.25,76.44,104.79,744.73} | 0.0546667(1 row) (mangled a bit by the auto-linewrap, I'm afraid) > > The second question is: why, oh why does the update take such and > > obscenely long time to complete? > > See above --- the problem is not within the plan, but must be sought > elsewhere. > > regards, tom lane > > ---------------------------(end of > broadcast)--------------------------- TIP 5: Have you checked our > extensive FAQ? > > http://www.postgresql.org/docs/faqs/FAQ.html -- Mike Glover Key ID BFD19F2C --=.:X5DcK3,R.IM5H Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.3 (GNU/Linux) iD8DBQE/9QzCZrNpxr/RnywRAkdvAJ9IzbSx+3uybk/oivFH/fFKhPmyoACfTmKp M8Ulg4+/e00DFg9+Ev3GsuM= =kKHh -----END PGP SIGNATURE----- --=.:X5DcK3,R.IM5H-- From pgsql-performance-owner@postgresql.org Fri Jan 2 05:44:04 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 459C2D1B499 for ; Fri, 2 Jan 2004 09:44:02 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 23320-02 for ; Fri, 2 Jan 2004 05:43:15 -0400 (AST) Received: from mailer.elma.loc (mail.elma.fr [213.41.14.138]) by svr1.postgresql.org (Postfix) with ESMTP id 87F7AD1B49B for ; Fri, 2 Jan 2004 05:43:12 -0400 (AST) Received: from mailer.elma.loc (localhost [127.0.0.1]) by localhost (Postfix) with ESMTP id 0E730EC018 for ; Fri, 2 Jan 2004 10:43:17 +0100 (CET) Received: from zoot.elma.fr (herve.elma.fr [10.0.1.2]) by mailer.elma.loc (Postfix) with ESMTP id E874BEC00F for ; Fri, 2 Jan 2004 10:43:16 +0100 (CET) From: =?iso-8859-15?q?Herv=E9=20Piedvache?= Organization: Elma =?iso-8859-15?q?Ing=E9nierie?= Informatique To: pgsql-performance@postgresql.org Subject: Why memory is not used ? Why vacuum so slow ? Date: Fri, 2 Jan 2004 10:42:57 +0100 User-Agent: KMail/1.5.4 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-15" Content-Transfer-Encoding: 8bit Content-Disposition: inline Message-Id: <200401021042.57714.herve@elma.fr> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/5 X-Sequence-Number: 5180 Hi, I have tried to tune a database that I'm using only for statistical access ... I mean that I'm importing a dump of my production database each night, but preserving some aggregat tables, and statistics ones ... (that I'm calculating after the importation of the dump). This database is only used by few people but make some big requests, tables have mixed sizes between 200 000 rows up to 10 000 000 records. This server's got 2Gb memory, and 100 Gb RAID 5 Hard disk, is a woody Debian, and I'm using a self compiled version of PotsgreSQL v7.3.4. My postgresql.conf file looks like this : # # Shared Memory Size # shared_buffers = 31000 # min max_connections*2 or 16, 8KB each max_fsm_relations = 1000 # min 10, fsm is free space map, ~40 bytes max_fsm_pages = 10000 # min 1000, fsm is free space map, ~6 bytes #max_locks_per_transaction = 64 # min 10 wal_buffers = 32 # min 4, typically 8KB each # # Non-shared Memory Sizes # sort_mem = 32768 # min 64, size in KB vacuum_mem = 32768 # min 1024, size in KB #checkpoint_segments = 3 # in logfile segments, min 1, 16MB each checkpoint_timeout = 160 # range 30-3600, in seconds effective_cache_size = 400000 # typically 8KB each random_page_cost = 1.5 # units are one sequential page fetch cost Before my effective_cache_size was 1000 ... and reading some tuning pages and comments telling : "effective_cache_size: You should adjust this according to the amount of free memory you have." ... I grow it to 400000 ... Then ... first point I'm only using 5% of my memory (all linux system,and software) ... and no swap (good point for this) ... Why I don't use more memory ... ?? Second point ... after importing my dump ... I make a vacuum full analyze of my base (in same time because of my caculation of the day before for my aggregats and stats tables about 200 000 row deleted and/or inserted for more than 20 tables (each)) ... but It takes about 5 hours ... Example of a (for me) really slow vacuum ... more than 85 min for a table with only 9105740 records ... INFO: �--Relation public.hebcnt-- INFO: �Pages 175115: Changed 0, reaped 3309, Empty 0, New 0; Tup 9105740: Vac 175330, Keep/VTL 0/0, UnUsed 0, MinLen 148, MaxLen 148; Re-using: Free/Avail. Space 46265980/26336600; EndEmpty/Avail. Pages 0/3310. ��������CPU 6.75s/1.67u sec elapsed 91.41 sec. INFO: �Index ix_hebcnt_idc: Pages 40446; Tuples 9105740: Deleted 175330. ��������CPU 2.94s/6.17u sec elapsed 222.34 sec. INFO: �Index ix_hebcnt_cweek: Pages 229977; Tuples 9105740: Deleted 175330. ��������CPU 9.64s/3.14u sec elapsed 1136.01 sec. INFO: �Index ix_hebcnt_cpte: Pages 72939; Tuples 9105740: Deleted 175330. ��������CPU 4.86s/9.13u sec elapsed 398.73 sec. INFO: �Index ix_hebcnt_idctweek: Pages 66014; Tuples 9105740: Deleted 175330. ��������CPU 3.87s/8.61u sec elapsed 163.26 sec. INFO: �Rel hebcnt: Pages: 175115 --> 171807; Tuple(s) moved: 175330. ��������CPU 16.49s/52.04u sec elapsed 1406.34 sec. INFO: �Index ix_hebcnt_idc: Pages 40446; Tuples 9105740: Deleted 175330. ��������CPU 1.76s/5.65u sec elapsed 124.98 sec. INFO: �Index ix_hebcnt_cweek: Pages 230690; Tuples 9105740: Deleted 175330. ��������CPU 10.07s/2.60u sec elapsed 1095.17 sec. INFO: �Index ix_hebcnt_cpte: Pages 72940; Tuples 9105740: Deleted 175330. ��������CPU 4.51s/8.90u sec elapsed 353.45 sec. INFO: �Index ix_hebcnt_idcweek: Pages 66015; Tuples 9105740: Deleted 175330. ��������CPU 3.96s/8.58u sec elapsed 147.64 sec. INFO: �--Relation pg_toast.pg_toast_76059978-- INFO: �Pages 0: Changed 0, reaped 0, Empty 0, New 0; Tup 0: Vac 0, Keep/VTL 0/0, UnUsed 0, MinLen 0, MaxLen 0; Re-using: Free/Avail. Space 0/0; EndEmpty/ Avail. Pages 0/0. ��������CPU 0.00s/0.00u sec elapsed 0.02 sec. INFO: �Index pg_toast_76059978_index: Pages 1; Tuples 0. ��������CPU 0.00s/0.00u sec elapsed 0.00 sec. INFO: �Analyzing public.hebcnt Structure of this table : frstats=# \d hebcnt Table "public.hebcnt" Column | Type | Modifiers ------------------+-----------------------------+------------------------ id_c | integer | not null contrat | text | not null arrete_week | text | not null cpte | text | not null is_active | boolean | not null year | text | not null use | integer | not null use_priv | integer | not null use_ind | integer | not null passback | integer | not null resa | integer | not null noshow | integer | not null nbc | integer | not null dureecnt | integer | not null dureecpt | integer | not null anciennete2 | integer | not null c_week | text | not null blacklist | integer | not null dcrea | timestamp without time zone | not null default now() dmaj | timestamp without time zone | Indexes: ix_hebcnt_cweek btree (c_week), ix_hebcnt_cpte btree (cpte), ix_hebcnt_idc btree (id_c), ix_hebcnt_idcweek btree (id_c, c_week) Any idea ? Regards, -- Herv� Piedvache Elma Ing�nierie Informatique 6 rue du Faubourg Saint-Honor� F-75008 - Paris - France Pho. 33-144949901 Fax. 33-144949902 From pgsql-performance-owner@postgresql.org Fri Jan 2 09:04:32 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 37955D1B43D for ; Fri, 2 Jan 2004 13:04:31 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 38807-05 for ; Fri, 2 Jan 2004 09:03:44 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id 6D522D1B45C for ; Fri, 2 Jan 2004 09:03:41 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id i02D3eU6031441 for ; Fri, 2 Jan 2004 13:03:40 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id i02Cb1Mn006567 for pgsql-performance@postgresql.org; Fri, 2 Jan 2004 12:37:01 GMT From: Christopher Browne X-Newsgroups: comp.databases.postgresql.performance Subject: Tuning Techniques To Avoid? Date: Fri, 02 Jan 2004 07:33:46 -0500 Organization: cbbrowne Computing Inc Lines: 16 Message-ID: References: <200401021042.57714.herve@elma.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Complaints-To: usenet@news.hub.org X-message-flag: Outlook is rather hackable, isn't it? X-Home-Page: http://www.cbbrowne.com/info/ X-Affero: http://svcs.affero.net/rm.php?r=cbbrowne User-Agent: Gnus/5.1003 (Gnus v5.10.3) XEmacs/21.4 (Reasonable Discussion, linux) Cancel-Lock: sha1:DmG7XsX4zLYDs85FtnMSIc7wiaw= To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/6 X-Sequence-Number: 5181 Here's a scheme for query optimization that probably needs to be avoided in that it would run afoul of a patent held by Oracle... It looks like what they have patented is pretty much a "greedy search" heuristic, starting by finding the table in a join that has the greatest selectivity (e.g. - where the number of entries selected is cut the most by the selection criteria), and then describes how to search for the "best" approach to joining in the other tables. -- output = reverse("gro.mca" "@" "enworbbc") http://www.ntlug.org/~cbbrowne/nonrdbms.html "If I could find a way to get [Saddam Hussein] out of there, even putting a contract out on him, if the CIA still did that sort of a thing, assuming it ever did, I would be for it." -- Richard M. Nixon From pgsql-performance-owner@postgresql.org Fri Jan 2 10:43:26 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 75183D1B4A6 for ; Fri, 2 Jan 2004 14:43:25 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 44686-10 for ; Fri, 2 Jan 2004 10:42:40 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id E4E09D1B4CE for ; Fri, 2 Jan 2004 10:42:30 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i02EgS19005631; Fri, 2 Jan 2004 09:42:28 -0500 (EST) To: =?iso-8859-15?q?Herv=E9=20Piedvache?= Cc: pgsql-performance@postgresql.org Subject: Re: Why memory is not used ? Why vacuum so slow ? In-reply-to: <200401021042.57714.herve@elma.fr> References: <200401021042.57714.herve@elma.fr> Comments: In-reply-to =?iso-8859-15?q?Herv=E9=20Piedvache?= message dated "Fri, 02 Jan 2004 10:42:57 +0100" Date: Fri, 02 Jan 2004 09:42:28 -0500 Message-ID: <5630.1073054548@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/7 X-Sequence-Number: 5182 =?iso-8859-15?q?Herv=E9=20Piedvache?= writes: > Second point ... after importing my dump ... I make a vacuum full analyze of > my base (in same time because of my caculation of the day before for my > aggregats and stats tables about 200 000 row deleted and/or inserted for more > than 20 tables (each)) ... but It takes about 5 hours ... Don't do vacuum full. You should not need it in ordinary circumstances, if you are doing plain vacuums on a reasonable schedule and you have the FSM parameters set high enough. (You do not BTW ... with 175000 pages in this table alone, 10000 FSM pages for the whole database is surely too low.) regards, tom lane From pgsql-performance-owner@postgresql.org Fri Jan 2 12:52:52 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E41BCD1D988 for ; Fri, 2 Jan 2004 16:52:48 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 49201-09 for ; Fri, 2 Jan 2004 12:52:03 -0400 (AST) Received: from www.postgresql.com (www.postgresql.com [200.46.204.209]) by svr1.postgresql.org (Postfix) with ESMTP id 0BB79D1D958 for ; Fri, 2 Jan 2004 12:51:54 -0400 (AST) Received: from mailer.elma.loc (unknown [213.41.14.138]) by www.postgresql.com (Postfix) with ESMTP id 3E3B9CF8A0D for ; Fri, 2 Jan 2004 12:47:58 -0400 (AST) Received: from mailer.elma.loc (localhost [127.0.0.1]) by localhost (Postfix) with ESMTP id 10D56EC018; Fri, 2 Jan 2004 16:18:47 +0100 (CET) Received: from zoot.elma.fr (herve.elma.fr [10.0.1.2]) by mailer.elma.loc (Postfix) with ESMTP id DBA3AEC00D; Fri, 2 Jan 2004 16:18:46 +0100 (CET) From: =?iso-8859-15?q?Herv=E9=20Piedvache?= Organization: Elma =?iso-8859-15?q?Ing=E9nierie?= Informatique To: Tom Lane Subject: Re: Why memory is not used ? Why vacuum so slow ? Date: Fri, 2 Jan 2004 16:18:28 +0100 User-Agent: KMail/1.5.4 Cc: pgsql-performance@postgresql.org References: <200401021042.57714.herve@elma.fr> <5630.1073054548@sss.pgh.pa.us> In-Reply-To: <5630.1073054548@sss.pgh.pa.us> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-15" Content-Transfer-Encoding: 8bit Content-Disposition: inline Message-Id: <200401021618.28300.herve@elma.fr> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/8 X-Sequence-Number: 5183 Hi Tom, Le Vendredi 2 Janvier 2004 15:42, Tom Lane a �crit : > =?iso-8859-15?q?Herv=E9=20Piedvache?= writes: > > Second point ... after importing my dump ... I make a vacuum full analyze > > of my base (in same time because of my caculation of the day before for > > my aggregats and stats tables about 200 000 row deleted and/or inserted > > for more than 20 tables (each)) ... but It takes about 5 hours ... > > Don't do vacuum full. You should not need it in ordinary circumstances, > if you are doing plain vacuums on a reasonable schedule and you have the > FSM parameters set high enough. (You do not BTW ... with 175000 pages in > this table alone, 10000 FSM pages for the whole database is surely too > low.) Ok for this ... I have now configured the FSM pages to 300 000 ... then when I have started the database I get a message about my SHMMAX too low ... it was set to : more /proc/sys/kernel/shmmax 262111232 Then I put 300000000 ... PostgreSQL accepted to start ... What can be maximum value for this ? To be usufull to the entire configuration ... ? Like this during during the vacuum full this is my used memory ... total used free shared buffers cached Mem: 2069608 2059052 10556 0 8648 1950672 -/+ buffers/cache: 99732 1969876 Swap: 2097136 16080 2081056 Seems that's I'm really using 5% of my memory ??? no ? or I missed something ? Now difficult to test again ... I will have to wait tomorrow morning to see the result ... because I have already vacuumed the base to day ... But I have done again a full vacuum to see if I have quick visible difference ... and I have also saw that the full vacuum for pg_atribute seems to be so slow ... more than 1 min for 7256 tupples ? Is this is normal ? INFO: --Relation pg_catalog.pg_attribute-- INFO: Pages 119: Changed 0, reaped 1, Empty 0, New 0; Tup 7256: Vac 0, Keep/ VTL 0/0, UnUsed 3, MinLen 128, MaxLen 128; Re-using: Free/Avail. Space 14664/504; EndEmpty/Avail. Pages 0/1. CPU 0.00s/0.00u sec elapsed 0.08 sec. INFO: Index pg_attribute_relid_attnam_index: Pages 21082; Tuples 7256: Deleted 0. CPU 0.83s/0.13u sec elapsed 59.32 sec. INFO: Index pg_attribute_relid_attnum_index: Pages 5147; Tuples 7256: Deleted 0. CPU 0.26s/0.03u sec elapsed 8.79 sec. INFO: Rel pg_attribute: Pages: 119 --> 119; Tuple(s) moved: 0. CPU 0.00s/0.00u sec elapsed 0.00 sec. INFO: Analyzing pg_catalog.pg_attribute Thanks for your help ... Regards, -- Herv� Piedvache Elma Ing�nierie Informatique 6 rue du Faubourg Saint-Honor� F-75008 - Paris - France Pho. 33-144949901 Fax. 33-144949902 From pgsql-performance-owner@postgresql.org Fri Jan 2 18:29:38 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D6E8AD1D61F for ; Fri, 2 Jan 2004 22:29:33 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 65442-01 for ; Fri, 2 Jan 2004 18:28:51 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id A3BB6D1DE50 for ; Fri, 2 Jan 2004 18:16:15 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i02Fj719005887; Fri, 2 Jan 2004 10:45:07 -0500 (EST) To: Mike Glover Cc: pgsql-performance@postgresql.org Subject: Re: Very slow update + not using clustered index In-reply-to: <20040101221630.4b001c35.mpg4@duluoz.net> References: <20040101193401.6640420b.mpg4@duluoz.net> <2966.1073016371@sss.pgh.pa.us> <20040101221630.4b001c35.mpg4@duluoz.net> Comments: In-reply-to Mike Glover message dated "Thu, 01 Jan 2004 22:16:30 -0800" Date: Fri, 02 Jan 2004 10:45:07 -0500 Message-ID: <5886.1073058307@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/10 X-Sequence-Number: 5185 Mike Glover writes: > Tom Lane wrote: >> It seems unlikely that the raw row inserts and updating the single >> index could be quite that slow --- perhaps you have a foreign key >> or trigger performance problem? > There are no foreign keys or triggers for either of the tables. Hmph. It's clear that it is the update overhead that's taking the time (since you show 292 seconds actual time in the update's top plan node --- that's the time to find the rows and compute their new values, and all the rest of the elapsed 3162 sec has to be update overhead). Maybe you just have a slow disk. Just out of curiosity, how much time does the update take if you don't have any index on the summary table? Try create temp table tsummary as select * from summary; vacuum analyze tsummary; explain analyze update tsummary set ... ; >> A quick experiment shows that if the planner does not have any reason >> to prefer one ordering over another, the current coding will put the >> last WHERE clause first: > [snip]> > It looks like the planner is already making a principled choice: After a little bit of experimentation I was reminded that the planner does account for the possibility that a merge join can stop short of full execution when the first mergejoin columns have different data ranges. In this case it's preferring to put price first because there is a greater discrepancy in the ranges of s.price_min and i.price than there is in the ranges of the isbn columns. I'm not sure that it's wrong. You could try increasing the statistics target on the price columns (and re-ANALYZing) to see if finer-grain data changes that estimate at all. In any case, the fact that the chosen plan doesn't make use of your index on isbn doesn't mean that such a plan wasn't considered. It was, but this plan was estimated to be less expensive. You could check out alternative plans and see if the estimate is accurate by fooling with enable_seqscan and enable_sort. regards, tom lane From pgsql-performance-owner@postgresql.org Fri Jan 2 18:29:28 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 3CC0CD1D17D for ; Fri, 2 Jan 2004 22:29:24 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 65516-02 for ; Fri, 2 Jan 2004 18:28:40 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id A403CD1DE41 for ; Fri, 2 Jan 2004 18:16:12 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i02FpD19005931; Fri, 2 Jan 2004 10:51:13 -0500 (EST) To: =?iso-8859-15?q?Herv=E9=20Piedvache?= Cc: pgsql-performance@postgresql.org Subject: Re: Why memory is not used ? Why vacuum so slow ? In-reply-to: <200401021618.28300.herve@elma.fr> References: <200401021042.57714.herve@elma.fr> <5630.1073054548@sss.pgh.pa.us> <200401021618.28300.herve@elma.fr> Comments: In-reply-to =?iso-8859-15?q?Herv=E9=20Piedvache?= message dated "Fri, 02 Jan 2004 16:18:28 +0100" Date: Fri, 02 Jan 2004 10:51:13 -0500 Message-ID: <5930.1073058673@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/9 X-Sequence-Number: 5184 =?iso-8859-15?q?Herv=E9=20Piedvache?= writes: > and I have also saw that the full vacuum for pg_atribute seems > to be so slow ... more than 1 min for 7256 tupples ? Is this is normal ? > INFO: --Relation pg_catalog.pg_attribute-- > INFO: Pages 119: Changed 0, reaped 1, Empty 0, New 0; Tup 7256: Vac 0, Keep/ > VTL 0/0, UnUsed 3, MinLen 128, MaxLen 128; Re-using: Free/Avail. Space > 14664/504; EndEmpty/Avail. Pages 0/1. > CPU 0.00s/0.00u sec elapsed 0.08 sec. > INFO: Index pg_attribute_relid_attnam_index: Pages 21082; Tuples 7256: > Deleted 0. > CPU 0.83s/0.13u sec elapsed 59.32 sec. > INFO: Index pg_attribute_relid_attnum_index: Pages 5147; Tuples 7256: Deleted > 0. > CPU 0.26s/0.03u sec elapsed 8.79 sec. You're suffering from index bloat (21000 pages in an index for a 119-page table!?). Updating to 7.4 would probably fix this, but if that's not practical consider reindexing pg_attribute. regards, tom lane From pgsql-performance-owner@postgresql.org Fri Jan 2 22:54:20 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A76C2D1D0F2 for ; Sat, 3 Jan 2004 02:54:18 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 94828-04 for ; Fri, 2 Jan 2004 22:53:36 -0400 (AST) Received: from smtp.istop.com (dci.doncaster.on.ca [66.11.168.194]) by svr1.postgresql.org (Postfix) with ESMTP id 130B9D1D450 for ; Fri, 2 Jan 2004 22:18:48 -0400 (AST) Received: from stark.dyndns.tv (gsstark.mtl.istop.com [66.11.160.162]) by smtp.istop.com (Postfix) with ESMTP id 58B32375C3; Fri, 2 Jan 2004 21:18:45 -0500 (EST) Received: from localhost ([127.0.0.1] helo=stark.dyndns.tv ident=foobar) by stark.dyndns.tv with smtp (Exim 3.36 #1 (Debian)) id 1AcbNF-0004fR-00; Fri, 02 Jan 2004 21:18:45 -0500 To: Christopher Browne Cc: pgsql-performance@postgresql.org Subject: Re: Tuning Techniques To Avoid? References: <200401021042.57714.herve@elma.fr> In-Reply-To: From: Greg Stark Organization: The Emacs Conspiracy; member since 1992 Date: 02 Jan 2004 21:18:45 -0500 Message-ID: <8765ft22re.fsf@stark.dyndns.tv> Lines: 11 User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/11 X-Sequence-Number: 5186 Christopher Browne writes: > Here's a scheme for query optimization that probably needs to be > avoided in that it would run afoul of a patent held by Oracle... What does this have to do with Herv� Piedvache's post "Why memory is not used?" ? -- greg From pgsql-performance-owner@postgresql.org Sat Jan 3 01:38:47 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1BAE7D1B44E for ; Sat, 3 Jan 2004 05:38:43 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 23222-08 for ; Sat, 3 Jan 2004 01:37:53 -0400 (AST) Received: from ms-smtp-04.tampabay.rr.com (ms-smtp-04-smtplb.tampabay.rr.com [65.32.5.134]) by svr1.postgresql.org (Postfix) with ESMTP id AD6F4D1B446 for ; Sat, 3 Jan 2004 01:37:39 -0400 (AST) Received: from unity.basin.cultured.net (24161231hfc249.tampabay.rr.com [24.161.231.249]) by ms-smtp-04.tampabay.rr.com (8.12.10/8.12.7) with ESMTP id i035bWu7003598 for ; Sat, 3 Jan 2004 00:37:33 -0500 (EST) Subject: "fun with multipart primary keys" hobby kit From: Chris Trawick To: pgsql-performance@postgresql.org Content-Type: text/plain Message-Id: <1073108262.733.27.camel@unity.basin.cultured.net> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Sat, 03 Jan 2004 00:37:43 -0500 Content-Transfer-Encoding: 7bit X-Virus-Scanned: Symantec AntiVirus Scan Engine X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/12 X-Sequence-Number: 5187 I tried searching the archives to find something like this. The search function doesn't like me much, and believe me the feeling is mutual. So I'm forced to pollute your inboxes with yet another "why the hell isn't this thing using my index" email. I apologize in advance. I have a many-to-many relation table with a multipart primary key: siren=# \d listcontact Table "public.listcontact" Column | Type | Modifiers ----------------+---------+----------- contactlistid | integer | not null contactid | bigint | not null Indexes: listcontact_pkey primary key btree (contactlistid, contactid) (There were some FKs in there too, but I stripped off everything I could during my investigation and they didn't survive.) I'm doing some performance testing so I loaded it with a few elephant piles: siren=# select count(*) from listcontact; count --------- 1409196 (1 row) And packed it down good: siren=# vacuum full analyze; VACUUM I didn't get the performance I expected. I took one of our queries and mutilated it and found some curious behavior on this table. I started running queries on just this table and couldn't explain what I was seeing. I tried this: siren=# EXPLAIN ANALYZE SELECT * FROM ListContact WHERE contactListID=-1 AND contactID=91347; QUERY PLAN --------------------------------------------------------------------------------------------------------- Seq Scan on listcontact (cost=0.00..29427.94 rows=1 width=12) (actual time=893.15..5079.52 rows=1 loops=1) Filter: ((contactlistid = -1) AND (contactid = 91347)) Total runtime: 5079.74 msec (3 rows) A seqscan... Fair enough, there's lots of memory on this box. I didn't want to see a seqscan though, I wanted to see an index. So, I disabled seqscan and tried it again: QUERY PLAN ------------------------------------------------------------------------------------------------------------------------------ Index Scan using listcontact_pkey on listcontact (cost=0.00..58522.64 rows=1 width=12) (actual time=402.73..9419.77 rows=1 loops=1) Index Cond: (contactlistid = -1) Filter: (contactid = 91347) Total runtime: 9419.97 msec (4 rows) Am I reading this right? Is it only using half of the fully-qualified pk index? How do I diagnose this? Has anyone seen this before? postgresql 7.3.1 linux 2.6.0 quad xeon 450 chris From pgsql-performance-owner@postgresql.org Sat Jan 3 01:58:53 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 8FCFDD1B45A for ; Sat, 3 Jan 2004 05:58:52 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 28950-05 for ; Sat, 3 Jan 2004 01:58:04 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 38F68D1B4AE for ; Sat, 3 Jan 2004 01:58:03 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i035vs19019197; Sat, 3 Jan 2004 00:57:54 -0500 (EST) To: Chris Trawick Cc: pgsql-performance@postgresql.org Subject: Re: "fun with multipart primary keys" hobby kit In-reply-to: <1073108262.733.27.camel@unity.basin.cultured.net> References: <1073108262.733.27.camel@unity.basin.cultured.net> Comments: In-reply-to Chris Trawick message dated "Sat, 03 Jan 2004 00:37:43 -0500" Date: Sat, 03 Jan 2004 00:57:54 -0500 Message-ID: <19196.1073109474@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/13 X-Sequence-Number: 5188 Chris Trawick writes: > contactid | bigint | not null ^^^^^^ > Am I reading this right? Is it only using half of the fully-qualified > pk index? How do I diagnose this? Has anyone seen this before? Surely you've been around here long enough to remember the must-cast-bigint-constants problem. regards, tom lane From pgsql-performance-owner@postgresql.org Sat Jan 3 02:08:02 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1CB81D1B45E for ; Sat, 3 Jan 2004 06:08:01 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 31118-04 for ; Sat, 3 Jan 2004 02:07:11 -0400 (AST) Received: from ms-smtp-02.tampabay.rr.com (ms-smtp-02-smtplb.tampabay.rr.com [65.32.5.132]) by svr1.postgresql.org (Postfix) with ESMTP id B0780D1B4CA for ; Sat, 3 Jan 2004 02:07:10 -0400 (AST) Received: from unity.basin.cultured.net (24161231hfc249.tampabay.rr.com [24.161.231.249]) by ms-smtp-02.tampabay.rr.com (8.12.10/8.12.7) with ESMTP id i03674qX005076; Sat, 3 Jan 2004 01:07:04 -0500 (EST) Subject: Re: "fun with multipart primary keys" hobby kit From: Chris Trawick To: Tom Lane Cc: pgsql-performance@postgresql.org In-Reply-To: <19196.1073109474@sss.pgh.pa.us> References: <1073108262.733.27.camel@unity.basin.cultured.net> <19196.1073109474@sss.pgh.pa.us> Content-Type: text/plain Message-Id: <1073110034.733.42.camel@unity.basin.cultured.net> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Sat, 03 Jan 2004 01:07:14 -0500 Content-Transfer-Encoding: 7bit X-Virus-Scanned: Symantec AntiVirus Scan Engine X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/14 X-Sequence-Number: 5189 Actually, it would appear that I was born yesterday. I had no idea. Added the cast and it fell right in. Thanks! chris <-- feeling pretty dumb right now On Sat, 2004-01-03 at 00:57, Tom Lane wrote: > Chris Trawick writes: > > contactid | bigint | not null > ^^^^^^ > > > Am I reading this right? Is it only using half of the fully-qualified > > pk index? How do I diagnose this? Has anyone seen this before? > > Surely you've been around here long enough to remember the > must-cast-bigint-constants problem. > > regards, tom lane From pgsql-performance-owner@postgresql.org Sat Jan 3 21:19:23 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 79172D1B442 for ; Sun, 4 Jan 2004 01:19:22 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 09025-05 for ; Sat, 3 Jan 2004 21:18:39 -0400 (AST) Received: from sccrmhc13.comcast.net (sccrmhc13.comcast.net [204.127.202.64]) by svr1.postgresql.org (Postfix) with ESMTP id 41A69D1B454 for ; Sat, 3 Jan 2004 21:18:27 -0400 (AST) Received: from lorenso.com (c-24-1-26-144.client.comcast.net[24.1.26.144]) by comcast.net (sccrmhc13) with ESMTP id <20040104011835016008tgk8e>; Sun, 4 Jan 2004 01:18:35 +0000 Message-ID: <3FF769EA.7090508@lorenso.com> Date: Sat, 03 Jan 2004 19:18:34 -0600 From: "D. Dante Lorenso" User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.6b) Gecko/20031205 Thunderbird/0.4 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: Indexing a Boolean or Null column? Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/16 X-Sequence-Number: 5191 I've been debating with a collegue who argues that indexing a boolean column is a BAD idea and that is will actually slow down queries. My plan is to have a table with many rows sharing 'versions' (version/archive/history) of data where the most current row is the one where 'is_active' contains a true value. If the table begins to look like this: data_id(pk) | data_lookup_key | data_is_active | ... ------------+-----------------+----------------+-------- 1 | banana | false | ... 2 | banana | false | ... 3 | banana | false | ... 4 | banana | false | ... 5 | banana | false | ... 6 | banana | false | ... 7 | banana | false | ... 8 | banana | false | ... 9 | banana | true | ... 10 | apple | true | ... 11 | pear | false | ... 12 | pear | false | ... 13 | pear | false | ... 14 | pear | false | ... 15 | pear | false | ... ... 1000000 | pear | true | ... Will an index on the 'data_is_active' column be used or work as I expect? I'm assuming that I may have a million entries sharing the same 'data_lookup_key' and I'll be using that to search for the active version of the row. SELECT * FROM table WHERE data_lookup_key = 'pear' AND data_is_active IS TRUE; Does it make sense to have an index on data_is_active? Now, I've read that in some databases the index on a column that has relatively even distribution of values over a small set of values will not be efficient. I bet this is in a FAQ somewhere. Can you point me in the right direction? Dante From pgsql-performance-owner@postgresql.org Sun Jan 4 00:27:46 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id B4431D1C526 for ; Sun, 4 Jan 2004 04:27:45 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 34794-01 for ; Sun, 4 Jan 2004 00:26:56 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 499D3D1B4A7 for ; Sun, 4 Jan 2004 00:26:55 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i044Qq19017337; Sat, 3 Jan 2004 23:26:53 -0500 (EST) To: "D. Dante Lorenso" Cc: pgsql-performance@postgresql.org Subject: Re: Indexing a Boolean or Null column? In-reply-to: <3FF769EA.7090508@lorenso.com> References: <3FF769EA.7090508@lorenso.com> Comments: In-reply-to "D. Dante Lorenso" message dated "Sat, 03 Jan 2004 19:18:34 -0600" Date: Sat, 03 Jan 2004 23:26:52 -0500 Message-ID: <17336.1073190412@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/17 X-Sequence-Number: 5192 "D. Dante Lorenso" writes: > Does it make sense to have an index on data_is_active? Hard to say. You weren't very clear about what fraction of the table rows you expect to have data_is_active = true. If that's a very small fraction, then an index might be worthwhile. However, I'd suggest using a partial index that merges the is_active test with some other useful behavior. For example, if this is a common pattern: > SELECT * > FROM table > WHERE data_lookup_key = 'pear' > AND data_is_active IS TRUE; then what you really want is CREATE INDEX myindex ON table (data_lookup_key) WHERE data_is_active IS TRUE; > I bet this is in a FAQ somewhere. Can you point me in the right > direction? See the docs on partial indexes. regards, tom lane From pgsql-performance-owner@postgresql.org Sun Jan 4 00:33:35 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D4E78D1B4BD for ; Sun, 4 Jan 2004 04:33:33 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 31429-09 for ; Sun, 4 Jan 2004 00:32:45 -0400 (AST) Received: from relay01.kbs.net.au (relay01.kbs.net.au [203.220.32.149]) by svr1.postgresql.org (Postfix) with ESMTP id 2F4EED1B4AE for ; Sun, 4 Jan 2004 00:32:44 -0400 (AST) Received: from [203.221.247.184] (helo=familyhealth.com.au) by relay01.kbs.net.au with esmtp (Exim 3.36 #2) id 1AczwP-0008AO-00; Sun, 04 Jan 2004 15:32:41 +1100 Message-ID: <3FF79764.6060805@familyhealth.com.au> Date: Sun, 04 Jan 2004 12:32:36 +0800 From: Christopher Kings-Lynne User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.6b) Gecko/20031205 Thunderbird/0.4 X-Accept-Language: en-us, en MIME-Version: 1.0 To: "D. Dante Lorenso" Cc: pgsql-performance@postgresql.org Subject: Re: Indexing a Boolean or Null column? References: <3FF769EA.7090508@lorenso.com> In-Reply-To: <3FF769EA.7090508@lorenso.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/18 X-Sequence-Number: 5193 > Will an index on the 'data_is_active' column be used or work > as I expect? I'm assuming that I may have a million entries > sharing the same 'data_lookup_key' and I'll be using that to > search for the active version of the row. An index just on a boolean column won't be 'selective enough'. eg. The index will only be able to choose 50% of the table - since it's faster to do a full table scan in that case, the index won't get used. A multi keyed index, however will work a bit better, eg an index over (data_lookup_key, data_is_active). That way, the index will first be able to find the correct key (which is nicely selective) and then will be able to halve the resulting search space to get the active ones. BTW, you shouldn't use 'banana', 'pear', etc as the data_lookup_key, you should make another table like this: id name 1 banana 2 apple 3 pear And then replace the data_lookup_key col with a column of integers that is a foreign key to the names table - waaaaay faster to process. Chris From pgsql-performance-owner@postgresql.org Sun Jan 4 01:25:35 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1EE7ED1B454 for ; Sun, 4 Jan 2004 05:25:35 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 40461-02 for ; Sun, 4 Jan 2004 01:24:45 -0400 (AST) Received: from sccrmhc12.comcast.net (sccrmhc12.comcast.net [204.127.202.56]) by svr1.postgresql.org (Postfix) with ESMTP id 6E349D1B479 for ; Sun, 4 Jan 2004 01:24:44 -0400 (AST) Received: from lorenso.com (c-24-1-26-144.client.comcast.net[24.1.26.144]) by comcast.net (sccrmhc12) with ESMTP id <2004010405244101200hrj9ce>; Sun, 4 Jan 2004 05:24:44 +0000 Message-ID: <3FF7A395.4050106@lorenso.com> Date: Sat, 03 Jan 2004 23:24:37 -0600 From: "D. Dante Lorenso" User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.6b) Gecko/20031205 Thunderbird/0.4 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Christopher Kings-Lynne Cc: pgsql-performance@postgresql.org Subject: Re: Indexing a Boolean or Null column? References: <3FF769EA.7090508@lorenso.com> <3FF79764.6060805@familyhealth.com.au> In-Reply-To: <3FF79764.6060805@familyhealth.com.au> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/19 X-Sequence-Number: 5194 Christopher Kings-Lynne wrote: > > Will an index on the 'data_is_active' column be used or work > > as I expect? I'm assuming that I may have a million entries > > sharing the same 'data_lookup_key' and I'll be using that to > > search for the active version of the row. > An index just on a boolean column won't be 'selective enough'. > eg. The index will only be able to choose 50% of the table - > since it's faster to do a full table scan in that case, the > index won't get used. Ok, so ...evenly distributed data on small set of values forces sequential scan since that's faster. I expected that based on what I've read so far. > A multi keyed index, however will work a bit better, eg an index > over (data_lookup_key, data_is_active). > > That way, the index will first be able to find the correct > key (which is nicely selective) and then will be able to > halve the resulting ? search space to get the active ones. I'm not using the 50% TRUE / 50% FALSE model. My model will be more like only ONE value IS TRUE for 'is_active' for each 'data_lookup_key' in my table. All the rest are FALSE. In this case for 100 rows all having the same 'data_lookup_key' we are looking at a 99% FALSE / 1% TRUE model ... and what I'll be searching for is the ONE TRUE. In this case, it WILL pay off to have the index on a boolean column, yes? Will I win my debate with my collegue? ;-) I think Tom Lanes suggestion of partial indexes is what I need to look into. > BTW, you shouldn't use 'banana', 'pear', etc as the data_lookup_key, > you should make another table like this: ... And then replace the > data_lookup_key col with a column of integers that is a foreign > key to the names table - waaaaay faster to process. Gotcha, yeah, I'm targeting as close to 3NF as I get get. Was just trying to be generic for my example ... bad example, oops. Dante From pgsql-performance-owner@postgresql.org Sun Jan 4 03:21:56 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 8E712D1B494 for ; Sun, 4 Jan 2004 07:21:54 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 42671-10 for ; Sun, 4 Jan 2004 03:21:05 -0400 (AST) Received: from relay01.kbs.net.au (relay01.kbs.net.au [203.220.32.149]) by svr1.postgresql.org (Postfix) with ESMTP id 33811D1BB71 for ; Sun, 4 Jan 2004 03:21:04 -0400 (AST) Received: from [203.221.247.184] (helo=familyhealth.com.au) by relay01.kbs.net.au with esmtp (Exim 3.36 #2) id 1Ad2ZP-00057I-00; Sun, 04 Jan 2004 18:21:07 +1100 Message-ID: <3FF7BEDD.7050208@familyhealth.com.au> Date: Sun, 04 Jan 2004 15:21:01 +0800 From: Christopher Kings-Lynne User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.6b) Gecko/20031205 Thunderbird/0.4 X-Accept-Language: en-us, en MIME-Version: 1.0 To: "D. Dante Lorenso" Cc: pgsql-performance@postgresql.org Subject: Re: Indexing a Boolean or Null column? References: <3FF769EA.7090508@lorenso.com> <3FF79764.6060805@familyhealth.com.au> <3FF7A395.4050106@lorenso.com> In-Reply-To: <3FF7A395.4050106@lorenso.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/20 X-Sequence-Number: 5195 > In this case, it WILL pay off to have the index on a boolean > column, yes? Will I win my debate with my collegue? ;-) > > I think Tom Lanes suggestion of partial indexes is what I need to > look into. Yes, given that it will be highly skewed towards false entries, Tom's suggestion is perfect. Chris From pgsql-performance-owner@postgresql.org Sun Jan 4 03:23:10 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 124B7D1C919 for ; Sun, 4 Jan 2004 07:23:09 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 45392-06 for ; Sun, 4 Jan 2004 03:22:21 -0400 (AST) Received: from relay01.kbs.net.au (relay01.kbs.net.au [203.220.32.149]) by svr1.postgresql.org (Postfix) with ESMTP id D972CD1B590 for ; Sun, 4 Jan 2004 03:22:19 -0400 (AST) Received: from [203.221.247.184] (helo=familyhealth.com.au) by relay01.kbs.net.au with esmtp (Exim 3.36 #2) id 1Ad2ac-00059i-00; Sun, 04 Jan 2004 18:22:23 +1100 Message-ID: <3FF7BF29.8070401@familyhealth.com.au> Date: Sun, 04 Jan 2004 15:22:17 +0800 From: Christopher Kings-Lynne User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.6b) Gecko/20031205 Thunderbird/0.4 X-Accept-Language: en-us, en MIME-Version: 1.0 To: "D. Dante Lorenso" Cc: pgsql-performance@postgresql.org Subject: Re: Indexing a Boolean or Null column? References: <3FF769EA.7090508@lorenso.com> <3FF79764.6060805@familyhealth.com.au> <3FF7A395.4050106@lorenso.com> In-Reply-To: <3FF7A395.4050106@lorenso.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/21 X-Sequence-Number: 5196 > Ok, so ...evenly distributed data on small set of values forces > sequential scan since that's faster. I expected that based on > what I've read so far. Actually, it's more a case of that fetching an item via and index is considered, say, four times slower than fetching something off a sequential scan (sort of). Hence, if you are selecting more than 25% of the table, then a sequential scan will be faster, even though it has to process more rows. Chris From pgsql-performance-owner@postgresql.org Sun Jan 4 03:49:50 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 6AFF7D1B46C for ; Sun, 4 Jan 2004 07:49:49 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 47228-08 for ; Sun, 4 Jan 2004 03:49:00 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id C6122D1B4AE for ; Sun, 4 Jan 2004 03:48:24 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i047mI19019150; Sun, 4 Jan 2004 02:48:18 -0500 (EST) To: Christopher Kings-Lynne Cc: "D. Dante Lorenso" , pgsql-performance@postgresql.org Subject: Re: Indexing a Boolean or Null column? In-reply-to: <3FF7BF29.8070401@familyhealth.com.au> References: <3FF769EA.7090508@lorenso.com> <3FF79764.6060805@familyhealth.com.au> <3FF7A395.4050106@lorenso.com> <3FF7BF29.8070401@familyhealth.com.au> Comments: In-reply-to Christopher Kings-Lynne message dated "Sun, 04 Jan 2004 15:22:17 +0800" Date: Sun, 04 Jan 2004 02:48:18 -0500 Message-ID: <19149.1073202498@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/22 X-Sequence-Number: 5197 Christopher Kings-Lynne writes: >> Ok, so ...evenly distributed data on small set of values forces >> sequential scan since that's faster. I expected that based on >> what I've read so far. > Actually, it's more a case of that fetching an item via and index is > considered, say, four times slower than fetching something off a > sequential scan (sort of). Hence, if you are selecting more than 25% of > the table, then a sequential scan will be faster, even though it has to > process more rows. Actually it's worse than that: if an indexscan is going to fetch more than a few percent of the table, the planner will think it slower than a sequential scan --- and usually it'll be right. The four-to-one ratio refers to the cost of fetching a whole page (8K) randomly versus sequentially. In a seqscan, you can examine all the rows on a page (dozens to hundreds usually) for the price of one page fetch. In an indexscan, one page fetch might bring in just one row that you care about. So the breakeven point is a lot worse than 4:1. There is constant debate about the values of these parameters; in particular the 4:1 page fetch cost ratio breaks down if you are able to cache a significant fraction of the table in RAM. See the list archives for details. But it's certainly true that an indexscan has to be a lot more selective than 25% before it's going to be a win over a seqscan. I'd say 1% to 5% is the right ballpark. regards, tom lane From pgsql-performance-owner@postgresql.org Sun Jan 4 18:05:20 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 39E7ED1B438 for ; Sun, 4 Jan 2004 22:05:19 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 33052-07 for ; Sun, 4 Jan 2004 18:04:32 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id B983ED1B8EA for ; Sun, 4 Jan 2004 18:04:13 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id i04M4DU6098633 for ; Sun, 4 Jan 2004 22:04:13 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id i04LiePq086489 for pgsql-performance@postgresql.org; Sun, 4 Jan 2004 21:44:40 GMT From: Christopher Browne X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Indexing a Boolean or Null column? Date: Sun, 04 Jan 2004 16:39:37 -0500 Organization: cbbrowne Computing Inc Lines: 65 Message-ID: References: <3FF769EA.7090508@lorenso.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Complaints-To: usenet@news.hub.org X-message-flag: Outlook is rather hackable, isn't it? X-Home-Page: http://www.cbbrowne.com/info/ X-Affero: http://svcs.affero.net/rm.php?r=cbbrowne User-Agent: Gnus/5.1003 (Gnus v5.10.3) XEmacs/21.4 (Reasonable Discussion, linux) Cancel-Lock: sha1:sQ6cN4j7q8GtV8LJbcJqdwlGnis= To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/23 X-Sequence-Number: 5198 After a long battle with technology, dante@lorenso.com ("D. Dante Lorenso"), an earthling, wrote: > I've been debating with a collegue who argues that indexing a > boolean column is a BAD idea and that is will actually slow > down queries. No, it would be expected to slow down inserts, but not likely queries. > Will an index on the 'data_is_active' column be used or work > as I expect? I'm assuming that I may have a million entries > sharing the same 'data_lookup_key' and I'll be using that to > search for the active version of the row. > SELECT * > FROM table > WHERE data_lookup_key = 'pear' > AND data_is_active IS TRUE; > > Does it make sense to have an index on data_is_active? Not really. > Now, I've read that in some databases the index on a column that has > relatively even distribution of values over a small set of values > will not be efficient. The problem is (and this is likely to be true for just about any database system that is 'page-based,' which is just about any of them, these days) that what happens, with the elements being so pervasive, throughout the table, queries will be quite likely to hit nearly every page of the table. If you're hitting practically every page, then it is more efficient to just walk thru the pages (Seq Scan) rather than to bother reading the index. The only improvement that could (in theory) be made is to cluster all the "true" values onto one set of pages, and all the "false" ones onto another set of pages, and have a special sort of index that knows which pages are "true" and "false". I _think_ that Oracle's notion of "cluster tables" function rather like this; it is rather debatable whether it would be worthwhile to do similar with PostgreSQL. A way of 'clustering' with PostgreSQL might be to have two tables table_active and table_inactive where a view, representing the 'join' of them, would throw in the 'data_is_active' value. By clever use of some rules/triggers, you could insert into the view, and have values get shuffled into the appropriate table. When doing a select on the view, if you asked for "data_is_active is TRUE", the select would only draw data from table_inactive, or vice-versa. Unfortunately, sometimes the query optimizer may not be clever enough when working with the resulting joins, though that may just be a Simple Matter Of Programming to make it more clever in future versions. :-) -- output = reverse("gro.gultn" "@" "enworbbc") http://www3.sympatico.ca/cbbrowne/spreadsheets.html Rules of the Evil Overlord #136. "If I build a bomb, I will simply remember which wire to cut if it has to be deactivated and make every wire red." From pgsql-performance-owner@postgresql.org Sun Jan 4 19:37:08 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 124E7D1B444 for ; Sun, 4 Jan 2004 23:37:08 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 50082-05 for ; Sun, 4 Jan 2004 19:36:20 -0400 (AST) Received: from sccrmhc11.comcast.net (sccrmhc11.comcast.net [204.127.202.55]) by svr1.postgresql.org (Postfix) with ESMTP id 69588D1B4A7 for ; Sun, 4 Jan 2004 19:36:17 -0400 (AST) Received: from [10.0.1.3] (h000393e3ce55.ne.client2.attbi.com[24.91.235.158]) by comcast.net (sccrmhc11) with SMTP id <20040104233619011006cu21e>; Sun, 4 Jan 2004 23:36:19 +0000 User-Agent: Microsoft-Entourage/10.1.1.2418 Date: Sun, 04 Jan 2004 18:36:16 -0500 Subject: Use my (date) index, darn it! From: John Siracusa To: Message-ID: Mime-version: 1.0 Reply-To: siracusa@mindspring.com Content-type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/24 X-Sequence-Number: 5199 I have a very large table (about a million rows) which I most frequently want to select a subset of rows from base on a date field. That date field is indexed, and when Postgres uses that index, queries are fast. But sometimes it decides not to use the index, resorting to a sequential scan instead. This is really, really slow. To try to convince it to use my date index, I turned off the sequential scan strategy in the planner. That worked on one copy of the db, but not on another, where it decided to use an index from another column entirely, which didn't help performance. I dropped the other index, leaving only the date index, and performance was good again. Obviously the planner is making some bad choices here. I know that it is trying to avoid random seeks or other scary things implied by a "correlation" statistic that is not close to 1 or -1, but it is massively overestimating the hit caused by those seeks and seemingly not taking into account the size of the table! This is Postgres 7.4 on Linux and Mac OS X, BTW. Anyway, to "fix" the situation, I clustered the table on the date column. But I fear that the data will slowly "drift" back to a state where the planner decides again that a sequential scan is a good idea. Blah. So, my question: how can I prevent this? Ideally, the planner should be smarter. Failing that, I'd like to be able to force it to use the index that I know will result in the fastest queries (3 seconds vs. 30 seconds in many cases). Suggestions? -John From pgsql-performance-owner@postgresql.org Mon Jan 5 02:56:35 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id F2143D1B4CE for ; Mon, 5 Jan 2004 06:56:34 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 15368-02 for ; Mon, 5 Jan 2004 02:55:46 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 9E5D7D1B456 for ; Mon, 5 Jan 2004 02:55:44 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i056th19022765; Mon, 5 Jan 2004 01:55:43 -0500 (EST) To: siracusa@mindspring.com Cc: pgsql-performance@postgresql.org Subject: Re: Use my (date) index, darn it! In-reply-to: References: Comments: In-reply-to John Siracusa message dated "Sun, 04 Jan 2004 18:36:16 -0500" Date: Mon, 05 Jan 2004 01:55:43 -0500 Message-ID: <22764.1073285743@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/25 X-Sequence-Number: 5200 John Siracusa writes: > Obviously the planner is making some bad choices here. A fair conclusion ... > I know that it is trying to avoid random seeks or other scary things > implied by a "correlation" statistic that is not close to 1 or -1, but > it is massively overestimating the hit caused by those seeks and > seemingly not taking into account the size of the table! You haven't given any evidence to support these conclusions, though. Could we see some table schemas, EXPLAIN ANALYZE output, and relevant pg_stats entries for the various cases? regards, tom lane From pgsql-performance-owner@postgresql.org Mon Jan 5 07:29:30 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 11B2AD1B48E for ; Mon, 5 Jan 2004 11:29:27 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 56587-01 for ; Mon, 5 Jan 2004 07:28:39 -0400 (AST) Received: from ns2.rox.net (ns2.rox.net [212.63.65.2]) by svr1.postgresql.org (Postfix) with ESMTP id 972A1D1B47D for ; Mon, 5 Jan 2004 07:28:36 -0400 (AST) Received: from localhost ([127.0.0.1]) by emma.rox.net with esmtp (Exim 4.22) id 1AdSuR-0006DM-CZ for pgsql-performance@postgresql.org; Mon, 05 Jan 2004 12:28:35 +0100 Received: from [195.135.143.205] (helo=[195.135.143.205]) by emma.rox.net with asmtp (Exim 4.22) id 1AdSuQ-0006D0-Om for pgsql-performance@postgresql.org; Mon, 05 Jan 2004 12:28:34 +0100 Mime-Version: 1.0 (Apple Message framework v609) Content-Transfer-Encoding: 7bit Message-Id: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> Content-Type: text/plain; charset=US-ASCII; format=flowed To: pgsql-performance@postgresql.org From: David Teran Subject: optimizing Postgres queries Date: Mon, 5 Jan 2004 12:28:32 +0100 X-Mailer: Apple Mail (2.609) X-Scanned-By: rockenstein AG X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/26 X-Sequence-Number: 5201 Hi, we are new to Postgres and we are evaluating Postgres 7.4 on MacOS X as an alternative to FrontBase 3.6.27. From the available features Postgres is the choice #1. We have some tests to check the performance and FrontBase is about 10 times faster than Postgres. We already played around with explain analyse select. It seems that for large tables Postgres does not use an index. We often see the scan message in the query plan. Were can we find more hints about tuning the performance? The database is about 350 MB large, without BLOB's. We tried to define every important index for the selects but it seems that something still goes wrong: FrontBase needs about 23 seconds for about 4300 selects and Postgres needs 4 minutes, 34 seconds. Any clues? regards David From pgsql-performance-owner@postgresql.org Mon Jan 5 07:37:48 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 0B859D1B4D7 for ; Mon, 5 Jan 2004 11:37:47 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 48120-09 for ; Mon, 5 Jan 2004 07:36:59 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id 5D230D1B48E for ; Mon, 5 Jan 2004 07:36:56 -0400 (AST) Received: from daithan.intranet.pspl.co.in shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.47 $ on Novell NetWare via secured & encrypted transport (TLS); Mon, 05 Jan 2004 04:36:53 -0700 From: Shridhar Daithankar To: pgsql-performance@postgresql.org Subject: Re: optimizing Postgres queries Date: Mon, 5 Jan 2004 17:05:55 +0530 User-Agent: KMail/1.5.4 References: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> In-Reply-To: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200401051705.55120.shridhar_daithankar@myrealbox.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/27 X-Sequence-Number: 5202 On Monday 05 January 2004 16:58, David Teran wrote: > We have some tests to check the performance and FrontBase is about 10 > times faster than Postgres. We already played around with explain > analyse select. It seems that for large tables Postgres does not use an > index. We often see the scan message in the query plan. Were can we > find more hints about tuning the performance? The database is about 350 > MB large, without BLOB's. We tried to define every important index for > the selects but it seems that something still goes wrong: FrontBase > needs about 23 seconds for about 4300 selects and Postgres needs 4 > minutes, 34 seconds. Check http://www.varlena.com/varlena/GeneralBits/Tidbits/perf.html http://www.varlena.com/varlena/GeneralBits/Tidbits/annotated_conf_e.html Are you sure you are using correct data types on indexes? e.g. if field1 is an int2 field, then following query would not use an index. select * from table where field1=2; However following will select * from table where field1=2::int2; It is called as typecasting and postgresql is rather strict about it when it comes to making a decision of index usage. I am sure above two tips could take care of some of the problems. Such kind of query needs more specific information. Can you post explain analyze output for queries and database schema. HTH Shridhar From pgsql-performance-owner@postgresql.org Mon Jan 5 08:05:53 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 34E4CD1B4A2 for ; Mon, 5 Jan 2004 12:05:52 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 56464-08 for ; Mon, 5 Jan 2004 08:05:05 -0400 (AST) Received: from ns2.rox.net (ns2.rox.net [212.63.65.2]) by svr1.postgresql.org (Postfix) with ESMTP id F1093D1B48E for ; Mon, 5 Jan 2004 08:05:01 -0400 (AST) Received: from localhost ([127.0.0.1]) by emma.rox.net with esmtp (Exim 4.22) id 1AdTTm-0006Zi-9I; Mon, 05 Jan 2004 13:05:06 +0100 Received: from [195.135.143.205] (helo=[195.135.143.205]) by emma.rox.net with asmtp (Exim 4.22) id 1AdTTl-0006ZT-M4; Mon, 05 Jan 2004 13:05:05 +0100 In-Reply-To: <200401051705.55120.shridhar_daithankar@myrealbox.com> References: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> <200401051705.55120.shridhar_daithankar@myrealbox.com> Mime-Version: 1.0 (Apple Message framework v609) Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: <64C8DBBA-3F77-11D8-A528-000A95A6F0DC@cluster9.com> Content-Transfer-Encoding: 7bit Cc: pgsql-performance@postgresql.org From: David Teran Subject: Re: optimizing Postgres queries Date: Mon, 5 Jan 2004 13:05:03 +0100 To: Shridhar Daithankar X-Mailer: Apple Mail (2.609) X-Scanned-By: rockenstein AG X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/28 X-Sequence-Number: 5203 Hi Shridhar, > Are you sure you are using correct data types on indexes? > Did not know about this... > e.g. if field1 is an int2 field, then following query would not use an > index. > our fk have the type bigint, when i try one simple select like this: explain analyze SELECT --columns-- FROM KEY_VALUE_META_DATA t0 WHERE t0.ID_FOREIGN_TABLE = 21110; i see that no index is being used whereas when i use explain analyze SELECT --columns-- FROM KEY_VALUE_META_DATA t0 WHERE t0.ID_FOREIGN_TABLE = 21110::bigint; an index is used. Very fine, the performance is about 10 to 100 times faster for the single select. I am using WebObjects with JDBC. I will now create a DB with integer instead of bigint and see how this performs. regards David From pgsql-performance-owner@postgresql.org Mon Jan 5 08:11:50 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 95820D1B473 for ; Mon, 5 Jan 2004 12:11:48 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 55973-08 for ; Mon, 5 Jan 2004 08:11:02 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id 1F3B4D1B44D for ; Mon, 5 Jan 2004 08:10:59 -0400 (AST) Received: from daithan.intranet.pspl.co.in shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.47 $ on Novell NetWare via secured & encrypted transport (TLS); Mon, 05 Jan 2004 05:11:02 -0700 From: Shridhar Daithankar To: pgsql-performance@postgresql.org Subject: Re: optimizing Postgres queries Date: Mon, 5 Jan 2004 17:40:05 +0530 User-Agent: KMail/1.5.4 References: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> <200401051705.55120.shridhar_daithankar@myrealbox.com> <64C8DBBA-3F77-11D8-A528-000A95A6F0DC@cluster9.com> In-Reply-To: <64C8DBBA-3F77-11D8-A528-000A95A6F0DC@cluster9.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200401051740.05006.shridhar_daithankar@myrealbox.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/29 X-Sequence-Number: 5204 On Monday 05 January 2004 17:35, David Teran wrote: > explain analyze SELECT --columns-- FROM KEY_VALUE_META_DATA t0 WHERE > t0.ID_FOREIGN_TABLE = 21110; > > i see that no index is being used whereas when i use > > explain analyze SELECT --columns-- FROM KEY_VALUE_META_DATA t0 WHERE > t0.ID_FOREIGN_TABLE = 21110::bigint; > > an index is used. Very fine, the performance is about 10 to 100 times > faster for the single select. > > I am using WebObjects with JDBC. I will now create a DB with integer > instead of bigint and see how this performs. The performance will likely to be the same. Its just that integer happens to be default integer type and hence it does not need an explicit typecast. ( I don't remember exactly which integer is default but it is either of int2,int4 and int8...:-)) The performance diffference is likely due to use of index, which is in turn due to typecasting. If you need bigint, you should use them. Just remember to typecast whenever required. Shridhar From pgsql-performance-owner@postgresql.org Mon Jan 5 08:18:56 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 47D6CD1B49B for ; Mon, 5 Jan 2004 12:18:54 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 60235-04 for ; Mon, 5 Jan 2004 08:18:07 -0400 (AST) Received: from ns2.rox.net (ns2.rox.net [212.63.65.2]) by svr1.postgresql.org (Postfix) with ESMTP id BA7A3D1B478 for ; Mon, 5 Jan 2004 08:18:04 -0400 (AST) Received: from localhost ([127.0.0.1]) by emma.rox.net with esmtp (Exim 4.22) id 1AdTgP-0006o6-8n; Mon, 05 Jan 2004 13:18:09 +0100 Received: from [195.135.143.205] (helo=[195.135.143.205]) by emma.rox.net with asmtp (Exim 4.22) id 1AdTgO-0006nf-Dl; Mon, 05 Jan 2004 13:18:08 +0100 In-Reply-To: <200401051740.05006.shridhar_daithankar@myrealbox.com> References: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> <200401051705.55120.shridhar_daithankar@myrealbox.com> <64C8DBBA-3F77-11D8-A528-000A95A6F0DC@cluster9.com> <200401051740.05006.shridhar_daithankar@myrealbox.com> Mime-Version: 1.0 (Apple Message framework v609) Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: <3796212E-3F79-11D8-A528-000A95A6F0DC@cluster9.com> Content-Transfer-Encoding: 7bit Cc: pgsql-performance@postgresql.org From: David Teran Subject: Re: optimizing Postgres queries Date: Mon, 5 Jan 2004 13:18:06 +0100 To: Shridhar Daithankar X-Mailer: Apple Mail (2.609) X-Scanned-By: rockenstein AG X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/30 X-Sequence-Number: 5205 Hi, > The performance will likely to be the same. Its just that integer > happens to > be default integer type and hence it does not need an explicit > typecast. ( I > don't remember exactly which integer is default but it is either of > int2,int4 > and int8...:-)) > The docs say int4 is much faster than int8, but i will check this. > The performance diffference is likely due to use of index, which is in > turn > due to typecasting. If you need bigint, you should use them. Just > remember to > typecast whenever required. This is my bigger problem: i am using EOF (OR mapping tool) which frees me more or less form writing a lot of SQL. If i need to typecast to use an index then i have to see how to do this with this framework. Regards David From pgsql-performance-owner@postgresql.org Mon Jan 5 08:39:41 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 61623D1B446 for ; Mon, 5 Jan 2004 12:39:40 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 59036-09 for ; Mon, 5 Jan 2004 08:38:54 -0400 (AST) Received: from smtp.pspl.co.in (www.pspl.co.in [202.54.11.65]) by svr1.postgresql.org (Postfix) with ESMTP id C2195D1B43A for ; Mon, 5 Jan 2004 08:38:48 -0400 (AST) Received: (from root@localhost) by smtp.pspl.co.in (8.12.9/8.12.9) id i05CcvOw011144 for ; Mon, 5 Jan 2004 18:08:57 +0530 Received: from daithan.intranet.pspl.co.in (daithan.intranet.pspl.co.in [192.168.7.161]) (authenticated bits=0) by persistent.co.in (8.12.9/8.12.9) with ESMTP id i05CcuJN011131 for ; Mon, 5 Jan 2004 18:08:57 +0530 From: Shridhar Daithankar To: pgsql-performance@postgresql.org Subject: Re: optimizing Postgres queries Date: Mon, 5 Jan 2004 18:07:53 +0530 User-Agent: KMail/1.5.4 References: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> <200401051740.05006.shridhar_daithankar@myrealbox.com> <3796212E-3F79-11D8-A528-000A95A6F0DC@cluster9.com> In-Reply-To: <3796212E-3F79-11D8-A528-000A95A6F0DC@cluster9.com> MIME-Version: 1.0 Content-Disposition: inline Message-Id: <200401051806.27496.shridhar_daithankar@myrealbox.com> Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/31 X-Sequence-Number: 5206 On Monday 05 January 2004 17:48, David Teran wrote: > Hi, > > > The performance will likely to be the same. Its just that integer > > happens to > > be default integer type and hence it does not need an explicit > > typecast. ( I > > don't remember exactly which integer is default but it is either of > > int2,int4 > > and int8...:-)) > > The docs say int4 is much faster than int8, but i will check this. Well yes. That is correct as well. What I (really) meant to say that an index scan to pick few in4 tuples wouldn't be hell much faster than an index scan to pick same number of tuples with int8 definition. The initial boost you got from converting to index scan, would be probably best you can beat out of it.. Of course if you are scanning a few million of them sequentially, then it is different story. > This is my bigger problem: i am using EOF (OR mapping tool) which frees > me more or less form writing a lot of SQL. If i need to typecast to use > an index then i have to see how to do this with this framework. Well, you can direct your queries to a function rather than table, that would cast the argument appropriately and select. Postgresql support function overloading as well, in case you need different types of arguments with same name. Or you can write an instead rule on server side which will perform casting before touching the table. I am not sure of exact details it would take to make it work, but it should work, at least in theory. That way you can preserve the efforts invested in the mapping tool. Of course, converting everything to integer might be a simpler option after all..:-) Shridhar From pgsql-performance-owner@postgresql.org Mon Jan 5 11:16:23 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id F0878D1B488 for ; Mon, 5 Jan 2004 15:16:21 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 90977-06 for ; Mon, 5 Jan 2004 11:15:35 -0400 (AST) Received: from relay01.kbs.net.au (relay01.kbs.net.au [203.220.32.149]) by svr1.postgresql.org (Postfix) with ESMTP id 3A3DFD1B43D for ; Mon, 5 Jan 2004 11:15:31 -0400 (AST) Received: from [203.221.247.78] (helo=familyhealth.com.au) by relay01.kbs.net.au with esmtp (Exim 3.36 #2) id 1AdWSA-00071I-00; Tue, 06 Jan 2004 02:15:39 +1100 Message-ID: <3FF97F8F.6040305@familyhealth.com.au> Date: Mon, 05 Jan 2004 23:15:27 +0800 From: Christopher Kings-Lynne User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.6b) Gecko/20031205 Thunderbird/0.4 X-Accept-Language: en-us, en MIME-Version: 1.0 To: David Teran Cc: Shridhar Daithankar , pgsql-performance@postgresql.org Subject: Re: optimizing Postgres queries References: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> <200401051705.55120.shridhar_daithankar@myrealbox.com> <64C8DBBA-3F77-11D8-A528-000A95A6F0DC@cluster9.com> In-Reply-To: <64C8DBBA-3F77-11D8-A528-000A95A6F0DC@cluster9.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/32 X-Sequence-Number: 5207 > explain analyze SELECT --columns-- FROM KEY_VALUE_META_DATA t0 WHERE > t0.ID_FOREIGN_TABLE = 21110::bigint; > > an index is used. Very fine, the performance is about 10 to 100 times > faster for the single select. An alternative technique is to do this: ... t0.ID_FOREIGN_TABLE = '21110'; Chris From pgsql-performance-owner@postgresql.org Mon Jan 5 11:23:32 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E8F17D1B4A2 for ; Mon, 5 Jan 2004 15:23:31 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 95119-02 for ; Mon, 5 Jan 2004 11:22:45 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 79814D1B4AD for ; Mon, 5 Jan 2004 11:22:36 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i05FMW19009967; Mon, 5 Jan 2004 10:22:32 -0500 (EST) To: David Teran Cc: Shridhar Daithankar , pgsql-performance@postgresql.org Subject: Re: optimizing Postgres queries In-reply-to: <3796212E-3F79-11D8-A528-000A95A6F0DC@cluster9.com> References: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> <200401051705.55120.shridhar_daithankar@myrealbox.com> <64C8DBBA-3F77-11D8-A528-000A95A6F0DC@cluster9.com> <200401051740.05006.shridhar_daithankar@myrealbox.com> <3796212E-3F79-11D8-A528-000A95A6F0DC@cluster9.com> Comments: In-reply-to David Teran message dated "Mon, 05 Jan 2004 13:18:06 +0100" Date: Mon, 05 Jan 2004 10:22:32 -0500 Message-ID: <9966.1073316152@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/33 X-Sequence-Number: 5208 David Teran writes: > This is my bigger problem: i am using EOF (OR mapping tool) which frees > me more or less form writing a lot of SQL. If i need to typecast to use > an index then i have to see how to do this with this framework. It's worth pointing out that this problem is fixed (at long last) in CVS tip. Ypu probably shouldn't expend large amounts of effort on working around a problem that will go away in 7.5. If you don't anticipate going to production for six months or so, you could adopt CVS tip as your development platform, with the expectation that 7.5 will be released by the time you need a production system. I wouldn't recommend running CVS tip as a production database but it should be plenty stable enough for devel purposes. Another plan would be to use int4 columns for the time being with the intention of widening them to int8 when you move to 7.5. This would depend on how soon you anticipate needing values > 32 bits, of course. regards, tom lane From pgsql-performance-owner@postgresql.org Mon Jan 5 12:19:53 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 0AA2FD1B4C7 for ; Mon, 5 Jan 2004 16:19:51 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 07692-05 for ; Mon, 5 Jan 2004 12:18:59 -0400 (AST) Received: from cayenne.palmdigitalmedia.com (cayenne.palmdigitalmedia.com [63.110.43.230]) by svr1.postgresql.org (Postfix) with SMTP id 33DCFD1B47C for ; Mon, 5 Jan 2004 12:19:00 -0400 (AST) Received: (qmail 20440 invoked from network); 5 Jan 2004 16:18:54 -0000 Received: from unknown (HELO ?192.168.124.23?) (63.110.43.226) by 0 with SMTP; 5 Jan 2004 16:18:54 -0000 User-Agent: Microsoft-Entourage/10.1.1.2418 Date: Mon, 05 Jan 2004 11:19:04 -0500 Subject: Re: Use my (date) index, darn it! From: John Siracusa To: Postgres Performance Message-ID: In-Reply-To: <22764.1073285743@sss.pgh.pa.us> Mime-version: 1.0 Reply-To: siracusa@mindspring.com Content-type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/34 X-Sequence-Number: 5209 On 1/5/04 1:55 AM, Tom Lane wrote: > John Siracusa writes: >> Obviously the planner is making some bad choices here. > > A fair conclusion ... > >> I know that it is trying to avoid random seeks or other scary things >> implied by a "correlation" statistic that is not close to 1 or -1, but >> it is massively overestimating the hit caused by those seeks and >> seemingly not taking into account the size of the table! > > You haven't given any evidence to support these conclusions, though. Well here's what I was basing that theory on: before clustering, the correlation for the date column was around 0.3. After clustering, it was 1, and the index was always used. Does clustering change any other statistics other that correlation? I ran analyze immediately before and after the cluster operation. > Could we see some table schemas, EXPLAIN ANALYZE output, and relevant > pg_stats entries for the various cases? Well, the table is clustered now, so I can't reproduce the situation. Is there any way to "uncluster" a table? Should I just cluster it on a different column? -John From pgsql-performance-owner@postgresql.org Mon Jan 5 12:30:44 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 2130ED1B458 for ; Mon, 5 Jan 2004 16:30:43 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 05627-10 for ; Mon, 5 Jan 2004 12:29:52 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id A55B0D1B531 for ; Mon, 5 Jan 2004 12:29:52 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i05GTo19010347; Mon, 5 Jan 2004 11:29:50 -0500 (EST) To: siracusa@mindspring.com Cc: Postgres Performance Subject: Re: Use my (date) index, darn it! In-reply-to: References: Comments: In-reply-to John Siracusa message dated "Mon, 05 Jan 2004 11:19:04 -0500" Date: Mon, 05 Jan 2004 11:29:50 -0500 Message-ID: <10346.1073320190@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/35 X-Sequence-Number: 5210 John Siracusa writes: > Is there any way to "uncluster" a table? Should I just cluster it on a > different column? That should work, if you choose one that's uncorrelated with the previous clustering attribute. regards, tom lane From pgsql-performance-owner@postgresql.org Mon Jan 5 12:34:35 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id B7742D1B4BE for ; Mon, 5 Jan 2004 16:34:33 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 15294-01 for ; Mon, 5 Jan 2004 12:33:42 -0400 (AST) Received: from lorax.kcilink.com (lorax.kciLink.com [206.112.95.1]) by svr1.postgresql.org (Postfix) with ESMTP id B4AA5D1B511 for ; Mon, 5 Jan 2004 12:33:43 -0400 (AST) Received: from localhost (localhost [127.0.0.1]) by lorax.kcilink.com (Postfix) with ESMTP id 521DB3EA4 for ; Mon, 5 Jan 2004 11:33:42 -0500 (EST) Received: from lorax.kcilink.com ([127.0.0.1]) by localhost (lorax.kcilink.com [127.0.0.1]) (amavisd-new, port 10024) with LMTP id 08645-05 for ; Mon, 5 Jan 2004 11:33:41 -0500 (EST) Received: from lorax.kcilink.com (localhost [127.0.0.1]) by lorax.kcilink.com (Postfix) with ESMTP id 98CF43E0B for ; Mon, 5 Jan 2004 11:33:41 -0500 (EST) Received: (from news@localhost) by lorax.kcilink.com (8.12.9p2/8.12.9/Submit) id i05GXfxX021441 for pgsql-performance@postgresql.org; Mon, 5 Jan 2004 11:33:41 -0500 (EST) (envelope-from news) To: pgsql-performance@postgresql.org Path: not-for-mail From: Vivek Khera Newsgroups: ml.postgres.performance Subject: Re: deferred foreign keys Date: Mon, 05 Jan 2004 11:33:40 -0500 Organization: Khera Communications, Inc., Rockville, MD Lines: 18 Message-ID: References: <3FF4FA25.7050802@familyhealth.com.au> NNTP-Posting-Host: yertle.kcilink.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Trace: lorax.kcilink.com 1073320421 5889 65.205.34.180 (5 Jan 2004 16:33:41 GMT) X-Complaints-To: daemon@kciLink.com NNTP-Posting-Date: Mon, 5 Jan 2004 16:33:41 +0000 (UTC) User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, berkeley-unix) Cancel-Lock: sha1:EsrsYGpOQBiYOgOR8KaF8iTuglM= X-Virus-Scanned: by amavisd-new at kciLink.com X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/36 X-Sequence-Number: 5211 >>>>> "CK" == Christopher Kings-Lynne writes: >> One more question: does the FK checker know to skip checking a >> constraint if the column in question did not change during an update? CK> As of 7.4, yes the check is skipped. Thanks. Then it sorta makes it moot for me to try deferred checks, since the Pimary and Foreign keys never change once set. I wonder what is making the transactions appear to run lockstep, then... -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Vivek Khera, Ph.D. Khera Communications, Inc. Internet: khera@kciLink.com Rockville, MD +1-301-869-4449 x806 AIM: vivekkhera Y!: vivek_khera http://www.khera.org/~vivek/ From pgsql-performance-owner@postgresql.org Mon Jan 5 13:05:18 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C8E37D1B43A for ; Mon, 5 Jan 2004 17:05:16 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 15904-03 for ; Mon, 5 Jan 2004 13:04:25 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id 98FCED1B432 for ; Mon, 5 Jan 2004 13:04:26 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id i05H4PU6072488 for ; Mon, 5 Jan 2004 17:04:26 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id i05GkLsg066361 for pgsql-performance@postgresql.org; Mon, 5 Jan 2004 16:46:21 GMT From: Christopher Browne X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Use my (date) index, darn it! Date: Mon, 05 Jan 2004 11:45:54 -0500 Organization: cbbrowne Computing Inc Lines: 49 Message-ID: References: <22764.1073285743@sss.pgh.pa.us> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Complaints-To: usenet@news.hub.org X-message-flag: Outlook is rather hackable, isn't it? X-Home-Page: http://www.cbbrowne.com/info/ X-Affero: http://svcs.affero.net/rm.php?r=cbbrowne User-Agent: Gnus/5.1003 (Gnus v5.10.3) XEmacs/21.4 (Reasonable Discussion, linux) Cancel-Lock: sha1:mzQTwC+VNXYoITJceNozYJwAQpI= To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/37 X-Sequence-Number: 5212 After a long battle with technology, siracusa@mindspring.com (John Siracusa), an earthling, wrote: > On 1/5/04 1:55 AM, Tom Lane wrote: >> John Siracusa writes: >>> Obviously the planner is making some bad choices here. >> >> A fair conclusion ... >> >>> I know that it is trying to avoid random seeks or other scary things >>> implied by a "correlation" statistic that is not close to 1 or -1, but >>> it is massively overestimating the hit caused by those seeks and >>> seemingly not taking into account the size of the table! >> >> You haven't given any evidence to support these conclusions, though. > > Well here's what I was basing that theory on: before clustering, the > correlation for the date column was around 0.3. After clustering, it was 1, > and the index was always used. Does clustering change any other statistics > other that correlation? I ran analyze immediately before and after the > cluster operation. > >> Could we see some table schemas, EXPLAIN ANALYZE output, and relevant >> pg_stats entries for the various cases? > > Well, the table is clustered now, so I can't reproduce the situation. Is > there any way to "uncluster" a table? Should I just cluster it on a > different column? That would presumably work... It sounds to me as though the statistics that are being collected aren't "good enough." That tends to be a sign that the quantity of statistics (e.g. - bins in the histogram) are insufficient. This would be resolved by changing the number of bins (default of 10) via "ALTER TABLE FOO ALTER COLUMN BAR SET STATISTICS 100" (or some other value higher than 10). Clustering would rearrange the contents of the table, and perhaps make the histogram 'more representative.' Increasing the "SET STATISTICS" value will quite likely be even more helpful, and is a lot less expensive than clustering the table... -- If this was helpful, rate me http://www.ntlug.org/~cbbrowne/nonrdbms.html Rules of the Evil Overlord #158. "I will exchange the labels on my folder of top-secret plans and my folder of family recipes. Imagine the hero's surprise when he decodes the stolen plans and finds instructions for Grandma's Potato Salad." From pgsql-performance-owner@postgresql.org Mon Jan 5 14:17:14 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A962BD1B43D for ; Mon, 5 Jan 2004 18:17:10 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 27239-09 for ; Mon, 5 Jan 2004 14:16:20 -0400 (AST) Received: from cayenne.palmdigitalmedia.com (cayenne.palmdigitalmedia.com [63.110.43.230]) by svr1.postgresql.org (Postfix) with SMTP id B4F55D1B444 for ; Mon, 5 Jan 2004 14:16:04 -0400 (AST) Received: (qmail 31174 invoked from network); 5 Jan 2004 18:15:46 -0000 Received: from unknown (HELO ?192.168.124.23?) (63.110.43.226) by 0 with SMTP; 5 Jan 2004 18:15:46 -0000 User-Agent: Microsoft-Entourage/10.1.1.2418 Date: Mon, 05 Jan 2004 13:15:56 -0500 Subject: Re: Use my (date) index, darn it! From: John Siracusa To: Postgres Performance Message-ID: In-Reply-To: Mime-version: 1.0 Reply-To: siracusa@mindspring.com Content-type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/38 X-Sequence-Number: 5213 On 1/5/04 11:45 AM, Christopher Browne wrote: > It sounds to me as though the statistics that are being collected > aren't "good enough." That tends to be a sign that the quantity of > statistics (e.g. - bins in the histogram) are insufficient. > > This would be resolved by changing the number of bins (default of 10) > via "ALTER TABLE FOO ALTER COLUMN BAR SET STATISTICS 100" (or some > other value higher than 10). I did that, but I wasn't sure what value to use and what column to increase. I believe I increased the date column itself to 50 or something, but then I wasn't sure what to do next. I re-analyzed the table with the date column set to 50 but it didn't seem to help, so I resorted to clustering. > Clustering would rearrange the contents of the table, and perhaps make > the histogram 'more representative.' Increasing the "SET STATISTICS" > value will quite likely be even more helpful, and is a lot less > expensive than clustering the table... What column(s) should I increase? Do I have to do anything after increasing the statistics, or do I just wait for the stats collector to do its thing? -John From pgsql-performance-owner@postgresql.org Mon Jan 5 14:39:29 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5FB35D1B45E for ; Mon, 5 Jan 2004 18:39:28 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 36668-05 for ; Mon, 5 Jan 2004 14:38:38 -0400 (AST) Received: from wolff.to (wolff.to [66.93.249.74]) by svr1.postgresql.org (Postfix) with SMTP id C8525D1B44E for ; Mon, 5 Jan 2004 14:38:37 -0400 (AST) Received: (qmail 11079 invoked by uid 500); 5 Jan 2004 18:38:59 -0000 Date: Mon, 5 Jan 2004 12:38:59 -0600 From: Bruno Wolff III To: Vivek Khera Cc: pgsql-performance@postgresql.org Subject: Re: deferred foreign keys Message-ID: <20040105183859.GA28448@wolff.to> Mail-Followup-To: Vivek Khera , pgsql-performance@postgresql.org References: <3FF4FA25.7050802@familyhealth.com.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/39 X-Sequence-Number: 5214 On Mon, Jan 05, 2004 at 11:33:40 -0500, Vivek Khera wrote: > > Thanks. Then it sorta makes it moot for me to try deferred checks, > since the Pimary and Foreign keys never change once set. I wonder > what is making the transactions appear to run lockstep, then... I think this is probably the issue with foreign key checks needing an exclusive lock, since there is no shared lock that will prevent deletes. This problem has been discussed a number of times on the lists and you should be able to find out more information from the archives if you want to confirm that this is the root cause of your problems. From pgsql-performance-owner@postgresql.org Mon Jan 5 14:48:37 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9BDDDD1B473 for ; Mon, 5 Jan 2004 18:48:34 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 31306-09 for ; Mon, 5 Jan 2004 14:47:44 -0400 (AST) Received: from ns2.rox.net (ns2.rox.net [212.63.65.2]) by svr1.postgresql.org (Postfix) with ESMTP id B3FA0D1B43D for ; Mon, 5 Jan 2004 14:47:44 -0400 (AST) Received: from localhost ([127.0.0.1]) by emma.rox.net with esmtp (Exim 4.22) id 1AdZlD-0004lC-FT; Mon, 05 Jan 2004 19:47:31 +0100 Received: from [195.135.143.205] (helo=[195.135.143.205]) by emma.rox.net with asmtp (Exim 4.22) id 1AdZlC-0004l0-Rx; Mon, 05 Jan 2004 19:47:30 +0100 In-Reply-To: <9966.1073316152@sss.pgh.pa.us> References: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> <200401051705.55120.shridhar_daithankar@myrealbox.com> <64C8DBBA-3F77-11D8-A528-000A95A6F0DC@cluster9.com> <200401051740.05006.shridhar_daithankar@myrealbox.com> <3796212E-3F79-11D8-A528-000A95A6F0DC@cluster9.com> <9966.1073316152@sss.pgh.pa.us> Mime-Version: 1.0 (Apple Message framework v609) Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: <9CC6A792-3FAF-11D8-A528-000A95A6F0DC@cluster9.com> Content-Transfer-Encoding: 7bit Cc: Shridhar Daithankar , pgsql-performance@postgresql.org From: David Teran Subject: Re: optimizing Postgres queries Date: Mon, 5 Jan 2004 19:47:28 +0100 To: Tom Lane X-Mailer: Apple Mail (2.609) X-Scanned-By: rockenstein AG X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/40 X-Sequence-Number: 5215 Hi Tom, > It's worth pointing out that this problem is fixed (at long last) in > CVS tip. Ypu probably shouldn't expend large amounts of effort on > working around a problem that will go away in 7.5. > We have now changed the definition to integer, this will work for some time. We are currently evaluating and have several production database we might switch in some time. What we found out now is that a query with a single 'where' works fine, the query planer uses the index but when we have 'two' where clauses it does not use the index anymore: EXPLAIN ANALYZE SELECT columns... FROM "KEY_VALUE_META_DATA" t0 WHERE (t0."ID_VALUE" = 14542); performs fine, less than one millisecond. EXPLAIN ANALYZE SELECT columns... FROM "KEY_VALUE_META_DATA" t0 WHERE (t0."ID_VALUE" = 14542 OR t0."ID_VALUE" = 14550); performs bad: about 235 milliseconds. I tried to change the second one to use IN but this did not help at all. Am i doing something wrong? I have an index defined like this: CREATE INDEX key_value_meta_data__id_value__fk_index ON "KEY_VALUE_META_DATA" USING btree ("ID_VALUE"); Regards David From pgsql-performance-owner@postgresql.org Mon Jan 5 14:54:41 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 718D1D1B4CC for ; Mon, 5 Jan 2004 18:54:39 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 41752-02 for ; Mon, 5 Jan 2004 14:53:49 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id A8791D1B433 for ; Mon, 5 Jan 2004 14:53:03 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i05Iqf19016231; Mon, 5 Jan 2004 13:52:41 -0500 (EST) To: David Teran Cc: Shridhar Daithankar , pgsql-performance@postgresql.org Subject: Re: optimizing Postgres queries In-reply-to: <9CC6A792-3FAF-11D8-A528-000A95A6F0DC@cluster9.com> References: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> <200401051705.55120.shridhar_daithankar@myrealbox.com> <64C8DBBA-3F77-11D8-A528-000A95A6F0DC@cluster9.com> <200401051740.05006.shridhar_daithankar@myrealbox.com> <3796212E-3F79-11D8-A528-000A95A6F0DC@cluster9.com> <9966.1073316152@sss.pgh.pa.us> <9CC6A792-3FAF-11D8-A528-000A95A6F0DC@cluster9.com> Comments: In-reply-to David Teran message dated "Mon, 05 Jan 2004 19:47:28 +0100" Date: Mon, 05 Jan 2004 13:52:40 -0500 Message-ID: <16230.1073328760@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/41 X-Sequence-Number: 5216 David Teran writes: > What we found out now is that a query with a single 'where' works fine, > the query planer uses the index but when we have 'two' where clauses it > does not use the index anymore: > EXPLAIN ANALYZE SELECT columns... FROM "KEY_VALUE_META_DATA" t0 WHERE > (t0."ID_VALUE" = 14542); performs fine, less than one millisecond. > EXPLAIN ANALYZE SELECT columns... FROM "KEY_VALUE_META_DATA" t0 WHERE > (t0."ID_VALUE" = 14542 OR t0."ID_VALUE" = 14550); performs bad: about > 235 milliseconds. Please, when you ask this sort of question, show the EXPLAIN ANALYZE output. It is not a virtue to provide minimal information and see if anyone can guess what's happening. regards, tom lane From pgsql-performance-owner@postgresql.org Mon Jan 5 14:58:09 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 4BDA1D1B4D5 for ; Mon, 5 Jan 2004 18:58:07 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 43054-01 for ; Mon, 5 Jan 2004 14:57:17 -0400 (AST) Received: from megazone.bigpanda.com (megazone.bigpanda.com [64.147.171.210]) by svr1.postgresql.org (Postfix) with ESMTP id 538EFD1B43A for ; Mon, 5 Jan 2004 14:57:17 -0400 (AST) Received: by megazone.bigpanda.com (Postfix, from userid 1001) id 7A15B354A9; Mon, 5 Jan 2004 10:57:02 -0800 (PST) Received: from localhost (localhost [127.0.0.1]) by megazone.bigpanda.com (Postfix) with ESMTP id 74F803548C; Mon, 5 Jan 2004 10:57:02 -0800 (PST) Date: Mon, 5 Jan 2004 10:57:02 -0800 (PST) From: Stephan Szabo To: Bruno Wolff III Cc: Vivek Khera , pgsql-performance@postgresql.org Subject: Re: deferred foreign keys In-Reply-To: <20040105183859.GA28448@wolff.to> Message-ID: <20040105105401.C72490@megazone.bigpanda.com> References: <3FF4FA25.7050802@familyhealth.com.au> <20040105183859.GA28448@wolff.to> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/43 X-Sequence-Number: 5218 On Mon, 5 Jan 2004, Bruno Wolff III wrote: > On Mon, Jan 05, 2004 at 11:33:40 -0500, > Vivek Khera wrote: > > > > Thanks. Then it sorta makes it moot for me to try deferred checks, > > since the Pimary and Foreign keys never change once set. I wonder > > what is making the transactions appear to run lockstep, then... > > I think this is probably the issue with foreign key checks needing an > exclusive lock, since there is no shared lock that will prevent deletes. But, if he's updating the fk table but not the keyed column, it should no longer be doing the check and grabbing the locks. If he's seeing it grab the row locks still a full test case would be handy because it'd probably mean we missed something. From pgsql-performance-owner@postgresql.org Mon Jan 5 14:58:04 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 305B9D1B4D4 for ; Mon, 5 Jan 2004 18:58:03 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 42052-04 for ; Mon, 5 Jan 2004 14:57:12 -0400 (AST) Received: from yertle.kcilink.com (yertle.kcilink.com [65.205.34.180]) by svr1.postgresql.org (Postfix) with ESMTP id D51B1D1B43D for ; Mon, 5 Jan 2004 14:57:12 -0400 (AST) Received: from [192.168.7.103] (host-103.int.kcilink.com [192.168.7.103]) by yertle.kcilink.com (Postfix) with ESMTP id 7693C2178C for ; Mon, 5 Jan 2004 13:57:07 -0500 (EST) Mime-Version: 1.0 (Apple Message framework v609) In-Reply-To: <20040105183859.GA28448@wolff.to> References: <3FF4FA25.7050802@familyhealth.com.au> <20040105183859.GA28448@wolff.to> Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: Content-Transfer-Encoding: 7bit From: Vivek Khera Subject: Re: deferred foreign keys Date: Mon, 5 Jan 2004 13:57:07 -0500 To: pgsql-performance@postgresql.org X-Mailer: Apple Mail (2.609) X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/42 X-Sequence-Number: 5217 On Jan 5, 2004, at 1:38 PM, Bruno Wolff III wrote: > I think this is probably the issue with foreign key checks needing an > exclusive lock, since there is no shared lock that will prevent > deletes. > That was my original thought upon reading all the discussion of late regarding the FK checking locks. I figured if I deferred the checks to commit, I could save some contention time. However, if FK checks are skipped if the field in question is not updated, what locks would there be? Are they taken even if the checks are not performed on some sort of "be prepared" principle? Vivek Khera, Ph.D. +1-301-869-4449 x806 From pgsql-performance-owner@postgresql.org Mon Jan 5 15:02:52 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id DFD3AD1B496 for ; Mon, 5 Jan 2004 19:02:48 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 43563-01 for ; Mon, 5 Jan 2004 15:01:59 -0400 (AST) Received: from yertle.kcilink.com (yertle.kcilink.com [65.205.34.180]) by svr1.postgresql.org (Postfix) with ESMTP id 9E111D1B45D for ; Mon, 5 Jan 2004 15:01:59 -0400 (AST) Received: from [192.168.7.103] (host-103.int.kcilink.com [192.168.7.103]) by yertle.kcilink.com (Postfix) with ESMTP id 49E482178B for ; Mon, 5 Jan 2004 14:02:00 -0500 (EST) Mime-Version: 1.0 (Apple Message framework v609) In-Reply-To: <20040105105401.C72490@megazone.bigpanda.com> References: <3FF4FA25.7050802@familyhealth.com.au> <20040105183859.GA28448@wolff.to> <20040105105401.C72490@megazone.bigpanda.com> Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: Content-Transfer-Encoding: 7bit From: Vivek Khera Subject: Re: deferred foreign keys Date: Mon, 5 Jan 2004 14:02:00 -0500 To: pgsql-performance@postgresql.org X-Mailer: Apple Mail (2.609) X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/44 X-Sequence-Number: 5219 On Jan 5, 2004, at 1:57 PM, Stephan Szabo wrote: > But, if he's updating the fk table but not the keyed column, it should > no > longer be doing the check and grabbing the locks. If he's seeing it > grab > the row locks still a full test case would be handy because it'd > probably > mean we missed something. > I'm not *sure* it is taking any locks. The transactions appear to be running lock step (operating on different parts of the same pair of tables) and I was going to see if deferring the locks made the difference. It is my feeling now that it will not. However, if there is a way to detect if locks are being taken, I'll do that. I'd like to avoid dropping and recreating the foreign keys if I can since it takes up some bit of time on the table with 20+ million rows. Vivek Khera, Ph.D. +1-301-869-4449 x806 From pgsql-performance-owner@postgresql.org Mon Jan 5 15:02:55 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 6A1D7D1B4AD for ; Mon, 5 Jan 2004 19:02:51 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 37454-09 for ; Mon, 5 Jan 2004 15:02:02 -0400 (AST) Received: from ns2.rox.net (ns2.rox.net [212.63.65.2]) by svr1.postgresql.org (Postfix) with ESMTP id 0A72CD1B477 for ; Mon, 5 Jan 2004 15:02:02 -0400 (AST) Received: from localhost ([127.0.0.1]) by emma.rox.net with esmtp (Exim 4.22) id 1AdZzI-0004rm-0V; Mon, 05 Jan 2004 20:02:04 +0100 Received: from [195.135.143.205] (helo=[195.135.143.205]) by emma.rox.net with asmtp (Exim 4.22) id 1AdZzH-0004rP-Ay; Mon, 05 Jan 2004 20:02:03 +0100 In-Reply-To: <16230.1073328760@sss.pgh.pa.us> References: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> <200401051705.55120.shridhar_daithankar@myrealbox.com> <64C8DBBA-3F77-11D8-A528-000A95A6F0DC@cluster9.com> <200401051740.05006.shridhar_daithankar@myrealbox.com> <3796212E-3F79-11D8-A528-000A95A6F0DC@cluster9.com> <9966.1073316152@sss.pgh.pa.us> <9CC6A792-3FAF-11D8-A528-000A95A6F0DC@cluster9.com> <16230.1073328760@sss.pgh.pa.us> Mime-Version: 1.0 (Apple Message framework v609) Content-Type: text/plain; charset=ISO-8859-1; delsp=yes; format=flowed Message-Id: Content-Transfer-Encoding: quoted-printable Cc: Shridhar Daithankar , pgsql-performance@postgresql.org From: David Teran Subject: Re: optimizing Postgres queries Date: Mon, 5 Jan 2004 20:02:01 +0100 To: Tom Lane X-Mailer: Apple Mail (2.609) X-Scanned-By: rockenstein AG X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/45 X-Sequence-Number: 5220 Hi Tom, > David Teran writes: >> What we found out now is that a query with a single 'where' works=20=20 >> fine, >> the query planer uses the index but when we have 'two' where clauses=20= =20 >> it >> does not use the index anymore: > >> EXPLAIN ANALYZE SELECT columns... FROM "KEY_VALUE_META_DATA" t0 WHERE >> (t0."ID_VALUE" =3D 14542); performs fine, less than one millisecond. > >> EXPLAIN ANALYZE SELECT columns... FROM "KEY_VALUE_META_DATA" t0 WHERE >> (t0."ID_VALUE" =3D 14542 OR t0."ID_VALUE" =3D 14550); performs bad: about >> 235 milliseconds. > > Please, when you ask this sort of question, show the EXPLAIN ANALYZE > output. It is not a virtue to provide minimal information and see if > anyone can guess what's happening. > Sorry for that, i thought this is such a trivial question that the=20=20 answer is easy. explain result from first query: Index=A0Scan=A0using=A0key_value_meta_data__id_value__fk_index=A0on=A0"KEY_= VALUE_M=20 ETA_DATA"=A0t0=A0=A0(cost=3D0.00..1585.52=A0rows=3D467=A0width=3D1068)=A0(a= ctual=A0time=3D0.42=20 4..0.493=A0rows=3D13=A0loops=3D1) =A0=A0Index=A0Cond:=A0("ID_VALUE"=A0=3D=A021094) Total runtime: 0.608 ms explain result from second query: Seq=A0Scan=A0on=A0"KEY_VALUE_META_DATA"=A0t0=A0=A0(cost=3D0.00..2671.16=A0r= ows=3D931=A0width=20 =3D1068)=A0(actual=A0time=3D122.669..172.179=A0rows=3D25=A0loops=3D1) =A0=A0Filter:=A0(("ID_VALUE"=A0=3D=A021094)=A0OR=A0("ID_VALUE"=A0=3D=A02110= 3)) Total runtime: 172.354 ms I found out that its possible to disable seq scans with set=20=20 enable_seqscan to off; then the second query result looks like this: Index=A0Scan=A0using=A0key_value_meta_data__id_value__fk_index,=A0key_value= _meta=20 _data__id_value__fk_index=A0on=A0"KEY_VALUE_META_DATA"=A0t0=A0=A0(cost=3D0.= 00..3173.=20 35=A0rows=3D931=A0width=3D1068)=A0(actual=A0time=3D0.116..0.578=A0rows=3D25= =A0loops=3D1) =A0=A0Index=A0Cond:=A0(("ID_VALUE"=A0=3D=A021094)=A0OR=A0("ID_VALUE"=A0=3D= =A021103)) Total runtime: 0.716 ms But i read in the docs that its not OK to turn this off by default. I=20=20 really wonder if this is my fault or not, from my point of view this is=20= =20 such a simple select that the query plan should not result in a table=20=20 scan. Regards David From pgsql-performance-owner@postgresql.org Mon Jan 5 15:03:52 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 8A175D1B45E for ; Mon, 5 Jan 2004 19:03:51 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 41945-04 for ; Mon, 5 Jan 2004 15:03:02 -0400 (AST) Received: from cayenne.palmdigitalmedia.com (cayenne.palmdigitalmedia.com [63.110.43.230]) by svr1.postgresql.org (Postfix) with SMTP id 88875D1B4C2 for ; Mon, 5 Jan 2004 15:03:01 -0400 (AST) Received: (qmail 8356 invoked from network); 5 Jan 2004 19:03:02 -0000 Received: from unknown (HELO ?192.168.124.23?) (63.110.43.226) by 0 with SMTP; 5 Jan 2004 19:03:02 -0000 User-Agent: Microsoft-Entourage/10.1.1.2418 Date: Mon, 05 Jan 2004 14:03:12 -0500 Subject: Select max(foo) and select count(*) optimization From: John Siracusa To: Postgres Performance Message-ID: Mime-version: 1.0 Reply-To: siracusa@mindspring.com Content-type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/46 X-Sequence-Number: 5221 Speaking of special cases (well, I was on the admin list) there are two kinds that would really benefit from some attention. 1. The query "select max(foo) from bar" where the column foo has an index. Aren't indexes ordered? If not, an "ordered index" would be useful in this situation so that this query, rather than doing a sequential scan of the whole table, would just "ask the index" for the max value and return nearly instantly. 2. The query "select count(*) from bar" Surely the total number of rows in a table is kept somewhere convenient. If not, it would be nice if it could be :) Again, rather than doing a sequential scan of the entire table, this type of query could return instantly. I believe MySQL does both of these optimizations (which are probably a lot easier in that product, given its data storage system). These were the first areas where I noticed a big performance difference between MySQL and Postgres. Especially with very large tables, hearing the disks grind as Postgres scans every single row in order to determine the number of rows in a table or the max value of a column (even a primary key created from a sequence) is pretty painful. If the implementation is not too horrendous, this is an area where an orders-of-magnitude performance increase can be had. -John From pgsql-performance-owner@postgresql.org Mon Jan 5 15:06:57 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A3ABED1B531 for ; Mon, 5 Jan 2004 19:06:55 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 38276-10 for ; Mon, 5 Jan 2004 15:06:05 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id F140BD1B478 for ; Mon, 5 Jan 2004 15:05:49 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i05J5m19020541; Mon, 5 Jan 2004 14:05:48 -0500 (EST) To: David Teran Cc: Shridhar Daithankar , pgsql-performance@postgresql.org Subject: Re: optimizing Postgres queries In-reply-to: References: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> <200401051705.55120.shridhar_daithankar@myrealbox.com> <64C8DBBA-3F77-11D8-A528-000A95A6F0DC@cluster9.com> <200401051740.05006.shridhar_daithankar@myrealbox.com> <3796212E-3F79-11D8-A528-000A95A6F0DC@cluster9.com> <9966.1073316152@sss.pgh.pa.us> <9CC6A792-3FAF-11D8-A528-000A95A6F0DC@cluster9.com> <16230.1073328760@sss.pgh.pa.us> Comments: In-reply-to David Teran message dated "Mon, 05 Jan 2004 20:02:01 +0100" Date: Mon, 05 Jan 2004 14:05:48 -0500 Message-ID: <20540.1073329548@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/47 X-Sequence-Number: 5222 David Teran writes: > explain result from second query: > Seq Scan on "KEY_VALUE_META_DATA" t0 (cost=0.00..2671.16 rows=931 width > =1068) (actual time=122.669..172.179 rows=25 loops=1) > Filter: (("ID_VALUE" = 21094) OR ("ID_VALUE" = 21103)) The problem is evidently that the row estimate is so far off (931 estimate vs 25 actual). Have you done ANALYZE or VACUUM ANALYZE on this table recently? If you have, I'd be interested to see the pg_stats row for ID_VALUE. It might be that you need to increase the statistics target for this table. regards, tom lane From pgsql-performance-owner@postgresql.org Mon Jan 5 15:21:41 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 4C1F5D1B44D for ; Mon, 5 Jan 2004 19:21:40 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 45044-06 for ; Mon, 5 Jan 2004 15:20:50 -0400 (AST) Received: from ns2.rox.net (ns2.rox.net [212.63.65.2]) by svr1.postgresql.org (Postfix) with ESMTP id BC87ED1B4AF for ; Mon, 5 Jan 2004 15:20:49 -0400 (AST) Received: from localhost ([127.0.0.1]) by emma.rox.net with esmtp (Exim 4.22) id 1AdaHT-000599-VE; Mon, 05 Jan 2004 20:20:52 +0100 Received: from [195.135.143.205] (helo=[195.135.143.205]) by emma.rox.net with asmtp (Exim 4.22) id 1AdaHS-00058C-SQ; Mon, 05 Jan 2004 20:20:50 +0100 In-Reply-To: <20540.1073329548@sss.pgh.pa.us> References: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> <200401051705.55120.shridhar_daithankar@myrealbox.com> <64C8DBBA-3F77-11D8-A528-000A95A6F0DC@cluster9.com> <200401051740.05006.shridhar_daithankar@myrealbox.com> <3796212E-3F79-11D8-A528-000A95A6F0DC@cluster9.com> <9966.1073316152@sss.pgh.pa.us> <9CC6A792-3FAF-11D8-A528-000A95A6F0DC@cluster9.com> <16230.1073328760@sss.pgh.pa.us> <20540.1073329548@sss.pgh.pa.us> Mime-Version: 1.0 (Apple Message framework v609) Content-Type: text/plain; charset=ISO-8859-1; delsp=yes; format=flowed Message-Id: <4530AF9E-3FB4-11D8-A528-000A95A6F0DC@cluster9.com> Content-Transfer-Encoding: quoted-printable Cc: pgsql-performance@postgresql.org From: David Teran Subject: Re: optimizing Postgres queries Date: Mon, 5 Jan 2004 20:20:49 +0100 To: Tom Lane X-Mailer: Apple Mail (2.609) X-Scanned-By: rockenstein AG X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/48 X-Sequence-Number: 5223 Hi Tom, first of all thanks for your help! I really appreciate your fast=20=20 response and if you ever have a question about WebObjects, just drop me=20= =20 line ;-) >> Seq Scan on "KEY_VALUE_META_DATA" t0 (cost=3D0.00..2671.16 rows=3D931= =20=20 >> width >> =3D1068) (actual time=3D122.669..172.179 rows=3D25 loops=3D1) >> Filter: (("ID_VALUE" =3D 21094) OR ("ID_VALUE" =3D 21103)) > > The problem is evidently that the row estimate is so far off (931 > estimate vs 25 actual). Have you done ANALYZE or VACUUM ANALYZE > on this table recently? If you have, I'd be interested to see the > pg_stats row for ID_VALUE. It might be that you need to increase > the statistics target for this table. > I am absolutely new to PostgreSQL. OK, after VACUUM ANALYZE i get: Index=A0Scan=A0using=A0key_value_meta_data__id_value__fk_index,=A0key_value= _meta=20 _data__id_value__fk_index=A0on=A0"KEY_VALUE_META_DATA"=A0t0=A0=A0(cost=3D0.= 00..19.94=20 =A0rows=3D14=A0width=3D75)=A0(actual=A0time=3D0.615..1.017=A0rows=3D25=A0lo= ops=3D1) =A0=A0Index=A0Cond:=A0(("ID_VALUE"=A0=3D=A021094)=A0OR=A0("ID_VALUE"=A0=3D= =A021103)) Total runtime: 2.565 ms and the second time i invoke this i get Index=A0Scan=A0using=A0key_value_meta_data__id_value__fk_index,=A0key_value= _meta=20 _data__id_value__fk_index=A0on=A0"KEY_VALUE_META_DATA"=A0t0=A0=A0(cost=3D0.= 00..19.94=20 =A0rows=3D14=A0width=3D75)=A0(actual=A0time=3D0.112..0.296=A0rows=3D25=A0lo= ops=3D1) =A0=A0Index=A0Cond:=A0(("ID_VALUE"=A0=3D=A021094)=A0OR=A0("ID_VALUE"=A0=3D= =A021103)) Total runtime: 0.429 ms Much better. So i think i will first read more about this optimization=20= =20 stuff and regular maintenance things. This is something i like very=20=20 much from FrontBase: no need for such things, simply start and run. But=20= =20 other things were not so fine ;-). Is there any hint where to start to understand more about this=20=20 optimization problem? regards David =09=09 From pgsql-performance-owner@postgresql.org Thu Jan 8 22:51:41 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 0D023D1B8C8 for ; Mon, 5 Jan 2004 19:25:34 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 45144-08 for ; Mon, 5 Jan 2004 15:24:43 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 533E2D1B472 for ; Mon, 5 Jan 2004 15:23:46 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i05JNj19020760; Mon, 5 Jan 2004 14:23:45 -0500 (EST) To: David Teran Cc: pgsql-performance@postgresql.org Subject: Re: optimizing Postgres queries In-reply-to: <4530AF9E-3FB4-11D8-A528-000A95A6F0DC@cluster9.com> References: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> <200401051705.55120.shridhar_daithankar@myrealbox.com> <64C8DBBA-3F77-11D8-A528-000A95A6F0DC@cluster9.com> <200401051740.05006.shridhar_daithankar@myrealbox.com> <3796212E-3F79-11D8-A528-000A95A6F0DC@cluster9.com> <9966.1073316152@sss.pgh.pa.us> <9CC6A792-3FAF-11D8-A528-000A95A6F0DC@cluster9.com> <16230.1073328760@sss.pgh.pa.us> <20540.1073329548@sss.pgh.pa.us> <4530AF9E-3FB4-11D8-A528-000A95A6F0DC@cluster9.com> Comments: In-reply-to David Teran message dated "Mon, 05 Jan 2004 20:20:49 +0100" Date: Mon, 05 Jan 2004 14:23:45 -0500 Message-ID: <20759.1073330625@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/90 X-Sequence-Number: 5265 David Teran writes: > Much better. So i think i will first read more about this optimization > stuff and regular maintenance things. See http://www.postgresql.org/docs/7.4/static/maintenance.html > Is there any hint where to start to understand more about this > optimization problem? http://www.postgresql.org/docs/7.4/static/performance-tips.html regards, tom lane From pgsql-performance-owner@postgresql.org Mon Jan 5 15:49:17 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 2DA3FD1B446 for ; Mon, 5 Jan 2004 19:49:16 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 48421-06 for ; Mon, 5 Jan 2004 15:48:25 -0400 (AST) Received: from megazone.bigpanda.com (megazone.bigpanda.com [64.147.171.210]) by svr1.postgresql.org (Postfix) with ESMTP id 8E367D1B44B for ; Mon, 5 Jan 2004 15:48:25 -0400 (AST) Received: by megazone.bigpanda.com (Postfix, from userid 1001) id 8352B3563A; Mon, 5 Jan 2004 11:48:26 -0800 (PST) Received: from localhost (localhost [127.0.0.1]) by megazone.bigpanda.com (Postfix) with ESMTP id 822A235632; Mon, 5 Jan 2004 11:48:26 -0800 (PST) Date: Mon, 5 Jan 2004 11:48:26 -0800 (PST) From: Stephan Szabo To: Vivek Khera Cc: pgsql-performance@postgresql.org Subject: Re: deferred foreign keys In-Reply-To: Message-ID: <20040105114730.U73612@megazone.bigpanda.com> References: <3FF4FA25.7050802@familyhealth.com.au> <20040105183859.GA28448@wolff.to> <20040105105401.C72490@megazone.bigpanda.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/49 X-Sequence-Number: 5224 On Mon, 5 Jan 2004, Vivek Khera wrote: > > On Jan 5, 2004, at 1:57 PM, Stephan Szabo wrote: > > > But, if he's updating the fk table but not the keyed column, it should > > no > > longer be doing the check and grabbing the locks. If he's seeing it > > grab > > the row locks still a full test case would be handy because it'd > > probably > > mean we missed something. > > > > I'm not *sure* it is taking any locks. The transactions appear to be > running lock step (operating on different parts of the same pair of > tables) and I was going to see if deferring the locks made the > difference. It is my feeling now that it will not. However, if there > is a way to detect if locks are being taken, I'll do that. I'd like to > avoid dropping and recreating the foreign keys if I can since it takes > up some bit of time on the table with 20+ million rows. The only way I can think of to see the locks is to do just one of the operations and then manually attempting to select for update the associated pk row. From pgsql-performance-owner@postgresql.org Mon Jan 5 15:53:14 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 6D0F5D1B45F for ; Mon, 5 Jan 2004 19:53:13 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 47404-09 for ; Mon, 5 Jan 2004 15:52:24 -0400 (AST) Received: from mail.hive.nj2.inquent.com (mc.carriermail.com [205.178.180.9]) by svr1.postgresql.org (Postfix) with SMTP id 71CA3D1B44B for ; Mon, 5 Jan 2004 15:52:08 -0400 (AST) Received: (qmail 10697 invoked from network); 5 Jan 2004 19:51:56 -0000 Received: from unknown (HELO ?10.0.2.16?) (216.208.117.7) by 205.178.180.9 with SMTP; 5 Jan 2004 19:51:56 -0000 Subject: Re: Select max(foo) and select count(*) optimization From: Rod Taylor To: siracusa@mindspring.com Cc: Postgresql Performance In-Reply-To: References: Content-Type: text/plain Message-Id: <1073332325.8958.8.camel@jester> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Mon, 05 Jan 2004 14:52:06 -0500 Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/50 X-Sequence-Number: 5225 > Especially with very large tables, hearing the disks grind as Postgres scans > every single row in order to determine the number of rows in a table or the > max value of a column (even a primary key created from a sequence) is pretty > painful. If the implementation is not too horrendous, this is an area where > an orders-of-magnitude performance increase can be had. Actually, it's very painful. For MySQL, they've accepted the concurrancy hit in order to accomplish it -- PostgreSQL would require a more subtle approach. Anyway, with Rules you can force this: ON INSERT UPDATE counter SET tablecount = tablecount + 1; ON DELETE UPDATE counter SET tablecount = tablecount - 1; You need to create a table "counter" with a single row that will keep track of the number of rows in the table. Just remember, you've now serialized all writes to the table, but in your situation it may be worth while. max(foo) optimizations requires an extension to the aggregates system. It will likely happen within a few releases. A work around can be accomplished today through the use of LIMIT and ORDER BY. From pgsql-performance-owner@postgresql.org Thu Jan 8 22:50:41 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5E9EAD1B472 for ; Mon, 5 Jan 2004 19:58:36 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 48699-07 for ; Mon, 5 Jan 2004 15:57:47 -0400 (AST) Received: from ns2.rox.net (ns2.rox.net [212.63.65.2]) by svr1.postgresql.org (Postfix) with ESMTP id 29C77D1B437 for ; Mon, 5 Jan 2004 15:57:47 -0400 (AST) Received: from localhost ([127.0.0.1]) by emma.rox.net with esmtp (Exim 4.22) id 1AdarF-0001dX-Fv; Mon, 05 Jan 2004 20:57:49 +0100 Received: from [195.135.143.205] (helo=[195.135.143.205]) by emma.rox.net with asmtp (Exim 4.22) id 1AdarE-0001c1-GY; Mon, 05 Jan 2004 20:57:48 +0100 In-Reply-To: <20759.1073330625@sss.pgh.pa.us> References: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> <200401051705.55120.shridhar_daithankar@myrealbox.com> <64C8DBBA-3F77-11D8-A528-000A95A6F0DC@cluster9.com> <200401051740.05006.shridhar_daithankar@myrealbox.com> <3796212E-3F79-11D8-A528-000A95A6F0DC@cluster9.com> <9966.1073316152@sss.pgh.pa.us> <9CC6A792-3FAF-11D8-A528-000A95A6F0DC@cluster9.com> <16230.1073328760@sss.pgh.pa.us> <20540.1073329548@sss.pgh.pa.us> <4530AF9E-3FB4-11D8-A528-000A95A6F0DC@cluster9.com> <20759.1073330625@sss.pgh.pa.us> Mime-Version: 1.0 (Apple Message framework v609) Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: <6F2414E8-3FB9-11D8-A528-000A95A6F0DC@cluster9.com> Content-Transfer-Encoding: 7bit Cc: pgsql-performance@postgresql.org From: David Teran Subject: Re: optimizing Postgres queries Date: Mon, 5 Jan 2004 20:57:47 +0100 To: Tom Lane X-Mailer: Apple Mail (2.609) X-Scanned-By: rockenstein AG X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/87 X-Sequence-Number: 5262 ... wow: executing a batch file with about 4250 selects, including lots of joins other things PostgreSQL 7.4 is about 2 times faster than FrontBase 3.6.27. OK, we will start to make larger tests but this is quite interesting already: we did not optimize a lot, just invoked VACUUM ANALYZE and then the selects ;-) Thanks to all who answered to this thread. cheers David From pgsql-performance-owner@postgresql.org Mon Jan 5 16:01:58 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 4903FD1B433 for ; Mon, 5 Jan 2004 20:01:57 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 49271-05 for ; Mon, 5 Jan 2004 16:01:08 -0400 (AST) Received: from cayenne.palmdigitalmedia.com (cayenne.palmdigitalmedia.com [63.110.43.230]) by svr1.postgresql.org (Postfix) with SMTP id 55634D1B444 for ; Mon, 5 Jan 2004 16:01:06 -0400 (AST) Received: (qmail 21093 invoked from network); 5 Jan 2004 20:01:07 -0000 Received: from unknown (HELO ?192.168.124.23?) (63.110.43.226) by 0 with SMTP; 5 Jan 2004 20:01:07 -0000 User-Agent: Microsoft-Entourage/10.1.1.2418 Date: Mon, 05 Jan 2004 15:01:18 -0500 Subject: Re: Select max(foo) and select count(*) optimization From: John Siracusa To: Postgres Performance Message-ID: In-Reply-To: <1073332325.8958.8.camel@jester> Mime-version: 1.0 Reply-To: siracusa@mindspring.com Content-type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/51 X-Sequence-Number: 5226 On 1/5/04 2:52 PM, Rod Taylor wrote: > max(foo) optimizations requires an extension to the aggregates system. > It will likely happen within a few releases. Looking forward to it. > A work around can be accomplished today through the use of LIMIT and ORDER BY. Wowzers, I never imagined that that'd be so much faster. Thanks! :) -John From pgsql-performance-owner@postgresql.org Mon Jan 5 16:15:02 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 3B103D1B444 for ; Mon, 5 Jan 2004 20:15:01 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 49537-08 for ; Mon, 5 Jan 2004 16:14:12 -0400 (AST) Received: from mail.hive.nj2.inquent.com (mc.carriermail.com [205.178.180.9]) by svr1.postgresql.org (Postfix) with SMTP id 58323D1B47F for ; Mon, 5 Jan 2004 16:14:11 -0400 (AST) Received: (qmail 25534 invoked from network); 5 Jan 2004 20:12:48 -0000 Received: from unknown (HELO ?10.0.2.16?) (216.208.117.7) by 205.178.180.9 with SMTP; 5 Jan 2004 20:12:48 -0000 Subject: Re: deferred foreign keys From: Rod Taylor To: Stephan Szabo Cc: Vivek Khera , Postgresql Performance In-Reply-To: <20040105114730.U73612@megazone.bigpanda.com> References: <3FF4FA25.7050802@familyhealth.com.au> <20040105183859.GA28448@wolff.to> <20040105105401.C72490@megazone.bigpanda.com> <20040105114730.U73612@megazone.bigpanda.com> Content-Type: text/plain Message-Id: <1073333664.8958.36.camel@jester> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Mon, 05 Jan 2004 15:14:24 -0500 Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/52 X-Sequence-Number: 5227 On Mon, 2004-01-05 at 14:48, Stephan Szabo wrote: > On Mon, 5 Jan 2004, Vivek Khera wrote: > > > > > On Jan 5, 2004, at 1:57 PM, Stephan Szabo wrote: > > > > > But, if he's updating the fk table but not the keyed column, it should > > > no > > > longer be doing the check and grabbing the locks. If he's seeing it > > > grab > > > the row locks still a full test case would be handy because it'd > > > probably > > > mean we missed something. > > > > > > > I'm not *sure* it is taking any locks. The transactions appear to be > > running lock step (operating on different parts of the same pair of > > tables) and I was going to see if deferring the locks made the > > difference. It is my feeling now that it will not. However, if there > > is a way to detect if locks are being taken, I'll do that. I'd like to > > avoid dropping and recreating the foreign keys if I can since it takes > > up some bit of time on the table with 20+ million rows. > > The only way I can think of to see the locks is to do just one of the > operations and then manually attempting to select for update the > associated pk row. When a locker runs into a row lock held by another transaction, the locker will show a pending lock on the transaction id in pg_locks. From pgsql-performance-owner@postgresql.org Mon Jan 5 16:24:25 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id BBD80D1B4BD for ; Mon, 5 Jan 2004 20:24:24 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 49267-09 for ; Mon, 5 Jan 2004 16:23:34 -0400 (AST) Received: from bob.samurai.com (bob.samurai.com [205.207.28.75]) by svr1.postgresql.org (Postfix) with ESMTP id 876A8D1B495 for ; Mon, 5 Jan 2004 16:23:33 -0400 (AST) Received: from tokyo.samurai.com (d226-89-59.home.cgocable.net [24.226.89.59]) by bob.samurai.com (Postfix) with ESMTP id 0C9C31F94; Mon, 5 Jan 2004 15:23:18 -0500 (EST) To: siracusa@mindspring.com Cc: Postgres Performance Subject: Re: Select max(foo) and select count(*) optimization From: Neil Conway In-Reply-To: (John Siracusa's message of "Mon, 05 Jan 2004 14:03:12 -0500") References: Date: Mon, 05 Jan 2004 15:23:16 -0500 Message-ID: <87k746p2kr.fsf@mailbox.samurai.com> User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/53 X-Sequence-Number: 5228 John Siracusa writes: > 1. The query "select max(foo) from bar" where the column foo has an index. > Aren't indexes ordered? If not, an "ordered index" would be useful in this > situation so that this query, rather than doing a sequential scan of the > whole table, would just "ask the index" for the max value and return nearly > instantly. http://www.postgresql.org/docs/current/static/functions-aggregate.html -Neil From pgsql-performance-owner@postgresql.org Mon Jan 5 16:35:20 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 8FE10D1B4C7 for ; Mon, 5 Jan 2004 20:35:16 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 57507-05 for ; Mon, 5 Jan 2004 16:34:26 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id 8DF00D1B444 for ; Mon, 5 Jan 2004 16:34:26 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id i05KYQU6065368 for ; Mon, 5 Jan 2004 20:34:26 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id i05KStnw062330 for pgsql-performance@postgresql.org; Mon, 5 Jan 2004 20:28:55 GMT From: Christopher Browne X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Select max(foo) and select count(*) optimization Date: Mon, 05 Jan 2004 15:26:15 -0500 Organization: cbbrowne Computing Inc Lines: 50 Message-ID: References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Complaints-To: usenet@news.hub.org X-message-flag: Outlook is rather hackable, isn't it? X-Home-Page: http://www.cbbrowne.com/info/ X-Affero: http://svcs.affero.net/rm.php?r=cbbrowne User-Agent: Gnus/5.1003 (Gnus v5.10.3) XEmacs/21.4 (Reasonable Discussion, linux) Cancel-Lock: sha1:gy8BOild0Jka8eTJbKw2gIJDnBI= To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/55 X-Sequence-Number: 5230 Oops! siracusa@mindspring.com (John Siracusa) was seen spray-painting on a wall: > Speaking of special cases (well, I was on the admin list) there are two > kinds that would really benefit from some attention. > > 1. The query "select max(foo) from bar" where the column foo has an > index. Aren't indexes ordered? If not, an "ordered index" would be > useful in this situation so that this query, rather than doing a > sequential scan of the whole table, would just "ask the index" for > the max value and return nearly instantly. > > 2. The query "select count(*) from bar" Surely the total number of > rows in a table is kept somewhere convenient. If not, it would be > nice if it could be :) Again, rather than doing a sequential scan of > the entire table, this type of query could return instantly. > > I believe MySQL does both of these optimizations (which are probably > a lot easier in that product, given its data storage system). These > were the first areas where I noticed a big performance difference > between MySQL and Postgres. > > Especially with very large tables, hearing the disks grind as > Postgres scans every single row in order to determine the number of > rows in a table or the max value of a column (even a primary key > created from a sequence) is pretty painful. If the implementation > is not too horrendous, this is an area where an orders-of-magnitude > performance increase can be had. These are both VERY frequently asked questions. In the case of question #1, the optimization you suggest could be accomplished via some Small Matter Of Programming. None of the people that have wanted the optimization have, however, offered to actually DO the programming. In the case of #2, the answer is "surely NOT." In MVCC databases, that information CANNOT be stored anywhere convenient because queries requested by transactions started at different points in time must get different answers. I think we need to add these questions and their answers to the FAQ so that the answer can be "See FAQ Item #17" rather than people having to gratuitously explain it over and over and over again. -- (reverse (concatenate 'string "moc.enworbbc" "@" "enworbbc")) http://www.ntlug.org/~cbbrowne/finances.html Rules of the Evil Overlord #127. "Prison guards will have their own cantina featuring a wide variety of tasty treats that will deliver snacks to the guards while on duty. The guards will also be informed that accepting food or drink from any other source will result in execution." From pgsql-performance-owner@postgresql.org Mon Jan 5 16:35:18 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 75B0DD1B4C4 for ; Mon, 5 Jan 2004 20:35:16 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 51665-08 for ; Mon, 5 Jan 2004 16:34:26 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id A0CEDD1B49B for ; Mon, 5 Jan 2004 16:34:26 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id i05KYQU8065368 for ; Mon, 5 Jan 2004 20:34:26 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id i05KStLQ062342 for pgsql-performance@postgresql.org; Mon, 5 Jan 2004 20:28:55 GMT From: Christopher Browne X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Use my (date) index, darn it! Date: Mon, 05 Jan 2004 15:27:43 -0500 Organization: cbbrowne Computing Inc Lines: 14 Message-ID: References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Complaints-To: usenet@news.hub.org X-message-flag: Outlook is rather hackable, isn't it? X-Home-Page: http://www.cbbrowne.com/info/ X-Affero: http://svcs.affero.net/rm.php?r=cbbrowne User-Agent: Gnus/5.1003 (Gnus v5.10.3) XEmacs/21.4 (Reasonable Discussion, linux) Cancel-Lock: sha1:fBD9OABaBFlG3B8g2r62ovVmWqk= To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/54 X-Sequence-Number: 5229 In the last exciting episode, siracusa@mindspring.com (John Siracusa) wrote: > What column(s) should I increase? Do I have to do anything after increasing > the statistics, or do I just wait for the stats collector to do its thing? You have to ANALYZE the table again, to force in new statistics. And if the index in question is on _just_ the date column, then it is probably only that date column where the "SET STATISTICS" needs to be increased. -- let name="cbbrowne" and tld="cbbrowne.com" in name ^ "@" ^ tld;; http://www.ntlug.org/~cbbrowne/sap.html Faith is the quality that enables you to eat blackberry jam on a picnic without looking to see whether the seeds move. -- DeMara Cabrera From pgsql-performance-owner@postgresql.org Mon Jan 5 18:34:05 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 58FAED1B438 for ; Mon, 5 Jan 2004 22:34:04 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 82907-02 for ; Mon, 5 Jan 2004 18:33:16 -0400 (AST) Received: from megazone.bigpanda.com (megazone.bigpanda.com [64.147.171.210]) by svr1.postgresql.org (Postfix) with ESMTP id 3B95AD1B43A for ; Mon, 5 Jan 2004 18:33:15 -0400 (AST) Received: by megazone.bigpanda.com (Postfix, from userid 1001) id 3787D35471; Mon, 5 Jan 2004 14:33:17 -0800 (PST) Received: from localhost (localhost [127.0.0.1]) by megazone.bigpanda.com (Postfix) with ESMTP id 364173532E; Mon, 5 Jan 2004 14:33:17 -0800 (PST) Date: Mon, 5 Jan 2004 14:33:17 -0800 (PST) From: Stephan Szabo To: Rod Taylor Cc: Vivek Khera , Postgresql Performance Subject: Re: deferred foreign keys In-Reply-To: <1073333664.8958.36.camel@jester> Message-ID: <20040105143127.I76870@megazone.bigpanda.com> References: <3FF4FA25.7050802@familyhealth.com.au> <20040105183859.GA28448@wolff.to> <20040105105401.C72490@megazone.bigpanda.com> <20040105114730.U73612@megazone.bigpanda.com> <1073333664.8958.36.camel@jester> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/56 X-Sequence-Number: 5231 On Mon, 5 Jan 2004, Rod Taylor wrote: > On Mon, 2004-01-05 at 14:48, Stephan Szabo wrote: > > On Mon, 5 Jan 2004, Vivek Khera wrote: > > > > > > > > On Jan 5, 2004, at 1:57 PM, Stephan Szabo wrote: > > > > > > > But, if he's updating the fk table but not the keyed column, it should > > > > no > > > > longer be doing the check and grabbing the locks. If he's seeing it > > > > grab > > > > the row locks still a full test case would be handy because it'd > > > > probably > > > > mean we missed something. > > > > > > > > > > I'm not *sure* it is taking any locks. The transactions appear to be > > > running lock step (operating on different parts of the same pair of > > > tables) and I was going to see if deferring the locks made the > > > difference. It is my feeling now that it will not. However, if there > > > is a way to detect if locks are being taken, I'll do that. I'd like to > > > avoid dropping and recreating the foreign keys if I can since it takes > > > up some bit of time on the table with 20+ million rows. > > > > The only way I can think of to see the locks is to do just one of the > > operations and then manually attempting to select for update the > > associated pk row. > > When a locker runs into a row lock held by another transaction, the > locker will show a pending lock on the transaction id in pg_locks. Yeah, but AFAIR that won't let you know if it's blocking on the particular row lock you're expecting. From pgsql-performance-owner@postgresql.org Mon Jan 5 20:22:35 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 797B4D1B4C4 for ; Tue, 6 Jan 2004 00:22:34 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 01245-06 for ; Mon, 5 Jan 2004 20:21:45 -0400 (AST) Received: from monsoon.he.net (monsoon.he.net [64.62.221.2]) by svr1.postgresql.org (Postfix) with SMTP id 53882D1B484 for ; Mon, 5 Jan 2004 20:21:43 -0400 (AST) Received: from [10.244.40.247] ([216.113.168.128]) by monsoon.he.net for ; Mon, 5 Jan 2004 16:21:40 -0800 Subject: Re: Select max(foo) and select count(*) optimization From: Paul Tuckfield To: Christopher Browne Cc: pgsql-performance@postgresql.org In-Reply-To: References: Content-Type: text/plain Organization: Message-Id: <1073348666.19994.3.camel@localhost.localdomain> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.2.2 (1.2.2-4) Date: 05 Jan 2004 16:24:26 -0800 Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/57 X-Sequence-Number: 5232 Not that I'm offering to do the porgramming mind you, :) but . . In the case of select count(*), one optimization is to do a scan of the primary key, not the table itself, if the table has a primary key. In a certain commercial, lesser database, this is called an "index fast full scan". It would be important to scan the index in physical order (sequential physical IO) and not in key order (random physical IO) I'm guessing the payoff as well as real-world-utility of a max(xxx) optimization are much higher than a count(*) optimization tho On Mon, 2004-01-05 at 12:26, Christopher Browne wrote: > Oops! siracusa@mindspring.com (John Siracusa) was seen spray-painting on a wall: > > Speaking of special cases (well, I was on the admin list) there are two > > kinds that would really benefit from some attention. > > > > 1. The query "select max(foo) from bar" where the column foo has an > > index. Aren't indexes ordered? If not, an "ordered index" would be > > useful in this situation so that this query, rather than doing a > > sequential scan of the whole table, would just "ask the index" for > > the max value and return nearly instantly. > > > > 2. The query "select count(*) from bar" Surely the total number of > > rows in a table is kept somewhere convenient. If not, it would be > > nice if it could be :) Again, rather than doing a sequential scan of > > the entire table, this type of query could return instantly. > > > > I believe MySQL does both of these optimizations (which are probably > > a lot easier in that product, given its data storage system). These > > were the first areas where I noticed a big performance difference > > between MySQL and Postgres. > > > > Especially with very large tables, hearing the disks grind as > > Postgres scans every single row in order to determine the number of > > rows in a table or the max value of a column (even a primary key > > created from a sequence) is pretty painful. If the implementation > > is not too horrendous, this is an area where an orders-of-magnitude > > performance increase can be had. > > These are both VERY frequently asked questions. > > In the case of question #1, the optimization you suggest could be > accomplished via some Small Matter Of Programming. None of the people > that have wanted the optimization have, however, offered to actually > DO the programming. > > In the case of #2, the answer is "surely NOT." In MVCC databases, > that information CANNOT be stored anywhere convenient because queries > requested by transactions started at different points in time must get > different answers. > > I think we need to add these questions and their answers to the FAQ so > that the answer can be "See FAQ Item #17" rather than people having to > gratuitously explain it over and over and over again. From pgsql-performance-owner@postgresql.org Mon Jan 5 20:41:16 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C3730D1B44B for ; Tue, 6 Jan 2004 00:41:15 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 05126-02 for ; Mon, 5 Jan 2004 20:40:28 -0400 (AST) Received: from main.gmane.org (main.gmane.org [80.91.224.249]) by svr1.postgresql.org (Postfix) with ESMTP id E02B0D1B45D for ; Mon, 5 Jan 2004 20:40:25 -0400 (AST) Received: from root by main.gmane.org with local (Exim 3.35 #1 (Debian)) id 1AdfGW-0005bM-00 for ; Tue, 06 Jan 2004 01:40:12 +0100 X-Injected-Via-Gmane: http://gmane.org/ To: pgsql-performance@postgresql.org Received: from sea.gmane.org ([80.91.224.252]) by main.gmane.org with esmtp (Exim 3.35 #1 (Debian)) id 1Adf6a-0005X2-00 for ; Tue, 06 Jan 2004 01:29:56 +0100 Received: from news by sea.gmane.org with local (Exim 3.35 #1 (Debian)) id 1Adf6a-0000oh-00 for ; Tue, 06 Jan 2004 01:29:56 +0100 From: Doug McNaught Subject: Re: Select max(foo) and select count(*) optimization Date: Mon, 05 Jan 2004 19:29:56 -0500 Lines: 12 Message-ID: <87oeth537f.fsf@asmodeus.mcnaught.org> References: <1073348666.19994.3.camel@localhost.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Complaints-To: usenet@sea.gmane.org User-Agent: Gnus/5.1002 (Gnus v5.10.2) Emacs/20.7 (gnu/linux) Cancel-Lock: sha1:7WUavIegZvZvtd1zN4AttylaUro= X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/58 X-Sequence-Number: 5233 Paul Tuckfield writes: > In the case of select count(*), one optimization is to do a scan of the > primary key, not the table itself, if the table has a primary key. In a > certain commercial, lesser database, this is called an "index fast full > scan". It would be important to scan the index in physical order > (sequential physical IO) and not in key order (random physical IO) That won't work because you still have to hit the actual tuple to determine visibility. -Doug From pgsql-performance-owner@postgresql.org Mon Jan 5 22:05:19 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 12236D1B436 for ; Tue, 6 Jan 2004 02:05:17 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 15271-07 for ; Mon, 5 Jan 2004 22:04:29 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id 15386D1B4AF for ; Mon, 5 Jan 2004 22:04:27 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id i0624QU6061594 for ; Tue, 6 Jan 2004 02:04:26 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id i061rUi0054163 for pgsql-performance@postgresql.org; Tue, 6 Jan 2004 01:53:30 GMT From: Christopher Browne X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Select max(foo) and select count(*) optimization Date: Mon, 05 Jan 2004 20:46:00 -0500 Organization: cbbrowne Computing Inc Lines: 29 Message-ID: References: <1073348666.19994.3.camel@localhost.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Complaints-To: usenet@news.hub.org X-message-flag: Outlook is rather hackable, isn't it? X-Home-Page: http://www.cbbrowne.com/info/ X-Affero: http://svcs.affero.net/rm.php?r=cbbrowne User-Agent: Gnus/5.1003 (Gnus v5.10.3) XEmacs/21.4 (Reasonable Discussion, linux) Cancel-Lock: sha1:CGrCaeE+coVOy09ohrN8ktqXNRI= To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/59 X-Sequence-Number: 5234 Martha Stewart called it a Good Thing when paul@tuckfield.com (Paul Tuckfield) wrote: > Not that I'm offering to do the porgramming mind you, :) but . . > > In the case of select count(*), one optimization is to do a scan of the > primary key, not the table itself, if the table has a primary key. In a > certain commercial, lesser database, this is called an "index fast full > scan". It would be important to scan the index in physical order > (sequential physical IO) and not in key order (random physical IO) The problem is that this "optimization" does not actually work. The index does not contain transaction visibility information, so you have to go to the pages of tuples in order to determine if any given tuple is visible. > I'm guessing the payoff as well as real-world-utility of a max(xxx) > optimization are much higher than a count(*) optimization tho That's probably so. In many cases, approximations, such as page counts, may be good enough, and pray consider, that ("an approximation") is probably all you were getting from the database systems that had an "optimization" to store the count in a counter. -- let name="cbbrowne" and tld="ntlug.org" in name ^ "@" ^ tld;; http://www3.sympatico.ca/cbbrowne/linuxxian.html "No, you misunderstand. Microsoft asked some hackers how they could make their system secure - the hackers replied "Turn it off.". So they did." -- Anthony Ord From pgsql-performance-owner@postgresql.org Mon Jan 5 22:35:17 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 3EDFAD1B511 for ; Tue, 6 Jan 2004 02:35:16 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 21009-06 for ; Mon, 5 Jan 2004 22:34:29 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id EC6BBD1B49B for ; Mon, 5 Jan 2004 22:34:26 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id i062YQU6078269 for ; Tue, 6 Jan 2004 02:34:26 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id i062VhL0076928 for pgsql-performance@postgresql.org; Tue, 6 Jan 2004 02:31:43 GMT From: Christopher Browne X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Select max(foo) and select count(*) optimization Date: Mon, 05 Jan 2004 21:31:29 -0500 Organization: cbbrowne Computing Inc Lines: 64 Message-ID: References: <1073332325.8958.8.camel@jester> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Complaints-To: usenet@news.hub.org X-message-flag: Outlook is rather hackable, isn't it? X-Home-Page: http://www.cbbrowne.com/info/ X-Affero: http://svcs.affero.net/rm.php?r=cbbrowne User-Agent: Gnus/5.1003 (Gnus v5.10.3) XEmacs/21.4 (Reasonable Discussion, linux) Cancel-Lock: sha1:5NnaKoa0P3JZKv1s3Qp8O/Vhdn0= To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/60 X-Sequence-Number: 5235 pg@rbt.ca (Rod Taylor) wrote: >> Especially with very large tables, hearing the disks grind as Postgres scans >> every single row in order to determine the number of rows in a table or the >> max value of a column (even a primary key created from a sequence) is pretty >> painful. If the implementation is not too horrendous, this is an area where >> an orders-of-magnitude performance increase can be had. > > Actually, it's very painful. For MySQL, they've accepted the concurrancy > hit in order to accomplish it -- PostgreSQL would require a more subtle > approach. > > Anyway, with Rules you can force this: > > ON INSERT UPDATE counter SET tablecount = tablecount + 1; > > ON DELETE UPDATE counter SET tablecount = tablecount - 1; > > You need to create a table "counter" with a single row that will keep > track of the number of rows in the table. Just remember, you've now > serialized all writes to the table, but in your situation it may be > worth while. There's a still more subtle approach that relieves the serialization constraint, at some cost... - You add rules that _insert_ a row each time there is an insert/delete ON INSERT insert into counts(table, value) values ('our_table', 1); ON DELETE insert into counts(table, value) values ('our_table', -1); - The "select count(*) from our_table" is replaced by "select sum(value) from counts where table = 'our_table'" - Periodically, a "compression" process goes through and either: a) Deletes the rows for 'our_table' and replaces them with one row with a conventionally-scanned 'count(*)' value, or b) Computes "select table, sum(value) as value from counts group by table", deletes all the existing rows in counts, and replaces them by the preceding selection, or c) Perhaps does something incremental that's like b), but which only processes parts of the "count" table at once. Process 500 rows, then COMMIT, or something of the sort... Note that this "counts" table can potentially grow _extremely_ large. The "win" comes when it gets compressed, so that instead of scanning through 500K items, it index-scans through 27, the 1 that has the "497000" that was the state of the table at the last compression, and then 26 singletons. A win comes in if an INSERT that adds in 50 rows can lead to inserting ('our_table', 50) into COUNTS, or a delete that eliminates 5000 rows puts in ('our_table', -5000). It's vital to run the "compression" reasonably often (much like VACUUM :-)) in order that the COUNTS summary table stays relatively small. -- let name="cbbrowne" and tld="cbbrowne.com" in String.concat "@" [name;tld];; http://www3.sympatico.ca/cbbrowne/wp.html Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it. -- Brian W. Kernighan From pgsql-performance-owner@postgresql.org Tue Jan 6 01:26:38 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 4D139D1B48A for ; Tue, 6 Jan 2004 05:26:37 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 55905-04 for ; Tue, 6 Jan 2004 01:25:54 -0400 (AST) Received: from candle.pha.pa.us (candle.pha.pa.us [207.106.42.251]) by svr1.postgresql.org (Postfix) with ESMTP id 89CF2D1B471 for ; Tue, 6 Jan 2004 01:25:44 -0400 (AST) Received: (from pgman@localhost) by candle.pha.pa.us (8.11.6/8.11.6) id i065PW915257; Tue, 6 Jan 2004 00:25:32 -0500 (EST) From: Bruce Momjian Message-Id: <200401060525.i065PW915257@candle.pha.pa.us> Subject: Re: optimizing Postgres queries In-Reply-To: <4530AF9E-3FB4-11D8-A528-000A95A6F0DC@cluster9.com> To: David Teran Date: Tue, 6 Jan 2004 00:25:32 -0500 (EST) Cc: Tom Lane , pgsql-performance@postgresql.org X-Mailer: ELM [version 2.4ME+ PL108 (25)] MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/61 X-Sequence-Number: 5236 David Teran wrote: > Index?Scan?using?key_value_meta_data__id_value__fk_index,?key_value_meta > _data__id_value__fk_index?on?"KEY_VALUE_META_DATA"?t0??(cost=0.00..19.94 > ?rows=14?width=75)?(actual?time=0.112..0.296?rows=25?loops=1) > ??Index?Cond:?(("ID_VALUE"?=?21094)?OR?("ID_VALUE"?=?21103)) > Total runtime: 0.429 ms > > Much better. So i think i will first read more about this optimization > stuff and regular maintenance things. This is something i like very > much from FrontBase: no need for such things, simply start and run. But > other things were not so fine ;-). > > Is there any hint where to start to understand more about this > optimization problem? Read the FAQ. There is an item about slow queries and indexes. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073 From pgsql-performance-owner@postgresql.org Tue Jan 6 02:33:35 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 36BA7D1B472 for ; Tue, 6 Jan 2004 06:33:35 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 07027-07 for ; Tue, 6 Jan 2004 02:32:51 -0400 (AST) Received: from smtp.pspl.co.in (unknown [202.54.11.65]) by svr1.postgresql.org (Postfix) with ESMTP id D6888D1B52A for ; Tue, 6 Jan 2004 02:32:43 -0400 (AST) Received: (from root@localhost) by smtp.pspl.co.in (8.12.9/8.12.9) id i066WsdJ025621 for ; Tue, 6 Jan 2004 12:02:54 +0530 Received: from daithan.intranet.pspl.co.in (daithan.intranet.pspl.co.in [192.168.7.161]) (authenticated bits=0) by persistent.co.in (8.12.9/8.12.9) with ESMTP id i066WsJN025608 for ; Tue, 6 Jan 2004 12:02:54 +0530 From: Shridhar Daithankar To: pgsql-performance@postgresql.org Subject: Re: Select max(foo) and select count(*) optimization Date: Tue, 6 Jan 2004 12:01:53 +0530 User-Agent: KMail/1.5.4 References: <1073348666.19994.3.camel@localhost.localdomain> In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200401061201.53978.shridhar_daithankar@persistent.co.in> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/62 X-Sequence-Number: 5237 On Tuesday 06 January 2004 07:16, Christopher Browne wrote: > Martha Stewart called it a Good Thing when paul@tuckfield.com (Paul Tuckfield) wrote: > > Not that I'm offering to do the porgramming mind you, :) but . . > > > > In the case of select count(*), one optimization is to do a scan of the > > primary key, not the table itself, if the table has a primary key. In a > > certain commercial, lesser database, this is called an "index fast full > > scan". It would be important to scan the index in physical order > > (sequential physical IO) and not in key order (random physical IO) > > The problem is that this "optimization" does not actually work. The > index does not contain transaction visibility information, so you have > to go to the pages of tuples in order to determine if any given tuple > is visible. It was rejected as an idea to add transaction visibility information to indexes. The time I proposed, my idea was to vacuum tuples on page level while postgresql pushes buffers out of shared cache. If indexes had visibility information, they could be cleaned out of order than heap tuples. This wouldn't have eliminated vacuum entirely but at least frequently hit data would be clean. But it was rejected because of associated overhead. Just thought worh a mention.. Shridhar From pgsql-performance-owner@postgresql.org Tue Jan 6 02:44:02 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 2943FD1B484 for ; Tue, 6 Jan 2004 06:44:02 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 07249-10 for ; Tue, 6 Jan 2004 02:43:18 -0400 (AST) Received: from smtp.pspl.co.in (unknown [202.54.11.65]) by svr1.postgresql.org (Postfix) with ESMTP id 4CA28D1B483 for ; Tue, 6 Jan 2004 02:43:11 -0400 (AST) Received: (from root@localhost) by smtp.pspl.co.in (8.12.9/8.12.9) id i066hMU6028871 for ; Tue, 6 Jan 2004 12:13:22 +0530 Received: from daithan.intranet.pspl.co.in (daithan.intranet.pspl.co.in [192.168.7.161]) (authenticated bits=0) by persistent.co.in (8.12.9/8.12.9) with ESMTP id i066hLJN028849 for ; Tue, 6 Jan 2004 12:13:22 +0530 From: Shridhar Daithankar To: Postgresql Performance Subject: Re: Select max(foo) and select count(*) optimization Date: Tue, 6 Jan 2004 12:12:21 +0530 User-Agent: KMail/1.5.4 References: <1073332325.8958.8.camel@jester> In-Reply-To: <1073332325.8958.8.camel@jester> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200401061212.21344.shridhar_daithankar@persistent.co.in> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/63 X-Sequence-Number: 5238 On Tuesday 06 January 2004 01:22, Rod Taylor wrote: > Anyway, with Rules you can force this: > > ON INSERT UPDATE counter SET tablecount = tablecount + 1; > > ON DELETE UPDATE counter SET tablecount = tablecount - 1; That would generate lot of dead tuples in counter table. How about select relpages,reltuples from pg_class where relname=; Assuming the stats are recent enough, it would be much faster and accurate.. Shridhar From pgsql-performance-owner@postgresql.org Thu Jan 8 22:51:36 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 55B1CD1B4C7 for ; Tue, 6 Jan 2004 12:05:18 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 62414-02 for ; Tue, 6 Jan 2004 08:04:36 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id 8F76DD1B449 for ; Tue, 6 Jan 2004 08:04:28 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id i06C4WU6092633 for ; Tue, 6 Jan 2004 12:04:32 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id i06BogVQ085444 for pgsql-performance@postgresql.org; Tue, 6 Jan 2004 11:50:42 GMT From: CoL X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Select max(foo) and select count(*) optimization Date: Tue, 06 Jan 2004 12:51:13 +0100 Organization: Hub.Org Networking Services Lines: 13 Message-ID: References: <1073332325.8958.8.camel@jester> <200401061212.21344.shridhar_daithankar@persistent.co.in> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Complaints-To: usenet@news.hub.org User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.6b) Gecko/20031205 Thunderbird/0.4 X-Accept-Language: en-us, en In-Reply-To: <200401061212.21344.shridhar_daithankar@persistent.co.in> To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/88 X-Sequence-Number: 5263 Hi, Shridhar Daithankar wrote: > > select relpages,reltuples from pg_class where relname=; > > Assuming the stats are recent enough, it would be much faster and accurate.. this needs an analyze ; before select from pg_class, cause only after analyze will update pg the pg_class C. From pgsql-performance-owner@postgresql.org Tue Jan 6 08:18:58 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D15DCD1B446 for ; Tue, 6 Jan 2004 12:18:53 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 62220-06 for ; Tue, 6 Jan 2004 08:18:13 -0400 (AST) Received: from druid.net (druid.net [216.126.72.98]) by svr1.postgresql.org (Postfix) with ESMTP id 61406D1B465 for ; Tue, 6 Jan 2004 08:18:04 -0400 (AST) Received: by druid.net (Postfix, from userid 1000) id A26C31A9E; Tue, 6 Jan 2004 07:18:09 -0500 (EST) From: "D'Arcy J.M. Cain" To: Shridhar Daithankar , Postgresql Performance Subject: Re: Select max(foo) and select count(*) optimization Date: Tue, 6 Jan 2004 07:18:08 -0500 User-Agent: KMail/1.5.4 References: <1073332325.8958.8.camel@jester> <200401061212.21344.shridhar_daithankar@persistent.co.in> In-Reply-To: <200401061212.21344.shridhar_daithankar@persistent.co.in> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200401060718.08444.darcy@druid.net> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/64 X-Sequence-Number: 5239 On January 6, 2004 01:42 am, Shridhar Daithankar wrote: > On Tuesday 06 January 2004 01:22, Rod Taylor wrote: > > Anyway, with Rules you can force this: > > > > ON INSERT UPDATE counter SET tablecount = tablecount + 1; > > > > ON DELETE UPDATE counter SET tablecount = tablecount - 1; > > That would generate lot of dead tuples in counter table. How about > > select relpages,reltuples from pg_class where relname=; > > Assuming the stats are recent enough, it would be much faster and > accurate.. Well, I did this: cert=# select relpages,reltuples from pg_class where relname= 'certificate'; relpages | reltuples ----------+------------- 399070 | 2.48587e+07 (1 row) Casting seemed to help: cert=# select relpages,reltuples::bigint from pg_class where relname= 'certificate'; relpages | reltuples ----------+----------- 399070 | 24858736 (1 row) But: cert=# select count(*) from certificate; [*Crunch* *Crunch* *Crunch*] count ---------- 19684668 (1 row) Am I missing something? Max certificate_id is 20569544 btw. -- D'Arcy J.M. Cain | Democracy is three wolves http://www.druid.net/darcy/ | and a sheep voting on +1 416 425 1212 (DoD#0082) (eNTP) | what's for dinner. From pgsql-performance-owner@postgresql.org Tue Jan 6 08:21:57 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 15583D1B465 for ; Tue, 6 Jan 2004 12:21:53 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 60378-07 for ; Tue, 6 Jan 2004 08:21:11 -0400 (AST) Received: from smtp.pspl.co.in (www.pspl.co.in [202.54.11.65]) by svr1.postgresql.org (Postfix) with ESMTP id 90F1DD1B440 for ; Tue, 6 Jan 2004 08:20:58 -0400 (AST) Received: (from root@localhost) by smtp.pspl.co.in (8.12.9/8.12.9) id i06CLEQR019820 for ; Tue, 6 Jan 2004 17:51:14 +0530 Received: from daithan.intranet.pspl.co.in (daithan.intranet.pspl.co.in [192.168.7.161]) (authenticated bits=0) by persistent.co.in (8.12.9/8.12.9) with ESMTP id i06CLEJN019806 for ; Tue, 6 Jan 2004 17:51:14 +0530 From: Shridhar Daithankar To: Postgresql Performance Subject: Re: Select max(foo) and select count(*) optimization Date: Tue, 6 Jan 2004 17:50:09 +0530 User-Agent: KMail/1.5.4 References: <200401061212.21344.shridhar_daithankar@persistent.co.in> <200401060718.08444.darcy@druid.net> In-Reply-To: <200401060718.08444.darcy@druid.net> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200401061750.09530.shridhar_daithankar@persistent.co.in> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/65 X-Sequence-Number: 5240 On Tuesday 06 January 2004 17:48, D'Arcy J.M. Cain wrote: > On January 6, 2004 01:42 am, Shridhar Daithankar wrote: > cert=# select relpages,reltuples::bigint from pg_class where relname= > 'certificate'; > relpages | reltuples > ----------+----------- > 399070 | 24858736 > (1 row) > > But: > > cert=# select count(*) from certificate; > [*Crunch* *Crunch* *Crunch*] > count > ---------- > 19684668 > (1 row) > > Am I missing something? Max certificate_id is 20569544 btw. Do 'vacuum analyze certificate' and try..:-) The numbers from pg_class are estimates updated by vacuum /analyze. Of course you need to run vacuum frequent enough for that statistics to be updated all the time or run autovacuum daemon.. Ran into same problem on my machine till I remembered about vacuum..:-) Shridhar From pgsql-performance-owner@postgresql.org Tue Jan 6 11:41:30 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id EAB1BD1B45D for ; Tue, 6 Jan 2004 15:41:28 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 00842-01 for ; Tue, 6 Jan 2004 11:40:48 -0400 (AST) Received: from apollo.le.ac.uk (apollo.le.ac.uk [143.210.16.125]) by svr1.postgresql.org (Postfix) with ESMTP id CB715D1B47D for ; Tue, 6 Jan 2004 11:40:38 -0400 (AST) Received: from [143.210.8.56] (helo=harrier.le.ac.uk) by apollo.le.ac.uk with esmtp (Exim 4.24) id 1AdtK4-0001mv-Kr for pgsql-performance@postgresql.org; Tue, 06 Jan 2004 15:40:48 +0000 Received: from localhost (cgp@localhost) by harrier.le.ac.uk (SGI-8.12.5/8.12.5/Submit) with ESMTP id i06FemsA10409716 for ; Tue, 6 Jan 2004 15:40:48 GMT X-Authentication-Warning: harrier.le.ac.uk: cgp owned process doing -bs Date: Tue, 6 Jan 2004 15:40:48 +0000 From: Clive Page To: pgsql-performance@postgresql.org Subject: Inefficient SELECT with OFFSET and LIMIT Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/66 X-Sequence-Number: 5241 I have just discovered that if one does a SELECT with a LIMIT and OFFSET values, say SELECT myfunc(mycol) FROM table LIMIT 50 OFFSET 10000 ; Then the whole of the selection expressions, including the function calls, are actuall executed for every record, not just those being selected but also those being skipped, i.e. 10050 in this case. Actually it's even odder, as the number is that plus one, as the next record in sequence is also passed to the function. I discovered this by accident, since I was using a user-defined function in pl/pgsql and included by mistake some debug code using RAISE INFO, so this diagnostic output gave the game away (and all of it came out before any of the results of the selection, which was another surprise). It looks as if OFFSET is implemented just be throwing away the results, until the OFFSET has been reached. It would be nice if OFFSET could be implemented in some more efficient way. -- Clive Page, Dept of Physics & Astronomy, University of Leicester, U.K. From pgsql-performance-owner@postgresql.org Tue Jan 6 12:04:42 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 91114D1B554 for ; Tue, 6 Jan 2004 16:04:41 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 05279-07 for ; Tue, 6 Jan 2004 12:03:52 -0400 (AST) Received: from bramble.mmrd.com (unknown [65.217.53.66]) by svr1.postgresql.org (Postfix) with ESMTP id 0A907D1B8EA for ; Tue, 6 Jan 2004 12:03:51 -0400 (AST) Received: from thorn.mmrd.com (thorn.mmrd.com [172.25.10.100]) by bramble.mmrd.com (8.12.8/8.12.8) with ESMTP id i06FLJcM013644; Tue, 6 Jan 2004 10:21:20 -0500 Received: from gnvex001.mmrd.com (gnvex001.mmrd.com [192.168.3.55]) by thorn.mmrd.com (8.11.6/8.11.6) with ESMTP id i06G3Nl08025; Tue, 6 Jan 2004 11:03:28 -0500 Received: from camel.mmrd.com ([172.25.5.213]) by gnvex001.mmrd.com with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2657.72) id XT879Y02; Tue, 6 Jan 2004 11:03:21 -0500 Subject: Re: Select max(foo) and select count(*) optimization From: Robert Treat To: Shridhar Daithankar Cc: Postgresql Performance In-Reply-To: <200401061750.09530.shridhar_daithankar@persistent.co.in> References: <200401061212.21344.shridhar_daithankar@persistent.co.in> <200401060718.08444.darcy@druid.net> <200401061750.09530.shridhar_daithankar@persistent.co.in> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Ximian Evolution 1.0.8 Date: 06 Jan 2004 11:03:23 -0500 Message-Id: <1073405003.20751.2.camel@camel> Mime-Version: 1.0 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/67 X-Sequence-Number: 5242 On Tue, 2004-01-06 at 07:20, Shridhar Daithankar wrote: > On Tuesday 06 January 2004 17:48, D'Arcy J.M. Cain wrote: > > On January 6, 2004 01:42 am, Shridhar Daithankar wrote: > > cert=# select relpages,reltuples::bigint from pg_class where relname= > > 'certificate'; > > relpages | reltuples > > ----------+----------- > > 399070 | 24858736 > > (1 row) > > > > But: > > > > cert=# select count(*) from certificate; > > [*Crunch* *Crunch* *Crunch*] > > count > > ---------- > > 19684668 > > (1 row) > > > > Am I missing something? Max certificate_id is 20569544 btw. > > Do 'vacuum analyze certificate' and try..:-) > > The numbers from pg_class are estimates updated by vacuum /analyze. Of course > you need to run vacuum frequent enough for that statistics to be updated all > the time or run autovacuum daemon.. > > Ran into same problem on my machine till I remembered about vacuum..:-) > Actually you only need to run analyze to update the statistics. Robert Treat -- Build A Brighter Lamp :: Linux Apache {middleware} PostgreSQL From pgsql-performance-owner@postgresql.org Tue Jan 6 12:34:08 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 02B83D1CA94 for ; Tue, 6 Jan 2004 16:34:07 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 10147-07 for ; Tue, 6 Jan 2004 12:33:18 -0400 (AST) Received: from smtp.pspl.co.in (unknown [202.54.11.65]) by svr1.postgresql.org (Postfix) with ESMTP id 362EBD1CA8E for ; Tue, 6 Jan 2004 12:31:27 -0400 (AST) Received: (from root@localhost) by smtp.pspl.co.in (8.12.9/8.12.9) id i06GVYNo016310 for ; Tue, 6 Jan 2004 22:01:34 +0530 Received: from persistent.co.in (daithan.intranet.pspl.co.in [192.168.7.161]) (authenticated bits=0) by persistent.co.in (8.12.9/8.12.9) with ESMTP id i06GVXJN016283 for ; Tue, 6 Jan 2004 22:01:33 +0530 Message-ID: <3FFAE2D8.6080301@persistent.co.in> Date: Tue, 06 Jan 2004 22:01:20 +0530 From: Shridhar Daithankar Organization: Persistent Systems Pvt. Ltd. User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.6b) Gecko/20031205 Thunderbird/0.4 X-Accept-Language: en-us, en MIME-Version: 1.0 Cc: Postgresql Performance Subject: Re: Select max(foo) and select count(*) optimization References: <200401061212.21344.shridhar_daithankar@persistent.co.in> <200401060718.08444.darcy@druid.net> <200401061750.09530.shridhar_daithankar@persistent.co.in> <1073405003.20751.2.camel@camel> In-Reply-To: <1073405003.20751.2.camel@camel> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/68 X-Sequence-Number: 5243 Robert Treat wrote: > On Tue, 2004-01-06 at 07:20, Shridhar Daithankar wrote: >>The numbers from pg_class are estimates updated by vacuum /analyze. Of course >>you need to run vacuum frequent enough for that statistics to be updated all >>the time or run autovacuum daemon.. >>Ran into same problem on my machine till I remembered about vacuum..:-) > Actually you only need to run analyze to update the statistics. Old habits die hard..:-) shridhar From pgsql-performance-owner@postgresql.org Tue Jan 6 14:13:26 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 73097D1B44B for ; Tue, 6 Jan 2004 18:13:22 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 37697-03 for ; Tue, 6 Jan 2004 14:12:34 -0400 (AST) Received: from smtp.istop.com (dci.doncaster.on.ca [66.11.168.194]) by svr1.postgresql.org (Postfix) with ESMTP id 03EE9D1B48A for ; Tue, 6 Jan 2004 14:12:33 -0400 (AST) Received: from stark.dyndns.tv (gsstark.mtl.istop.com [66.11.160.162]) by smtp.istop.com (Postfix) with ESMTP id B793436BFD; Tue, 6 Jan 2004 13:12:17 -0500 (EST) Received: from localhost ([127.0.0.1] helo=stark.dyndns.tv ident=foobar) by stark.dyndns.tv with smtp (Exim 3.36 #1 (Debian)) id 1Advgf-0000CU-00; Tue, 06 Jan 2004 13:12:17 -0500 To: Clive Page Cc: pgsql-performance@postgresql.org Subject: Re: Inefficient SELECT with OFFSET and LIMIT References: In-Reply-To: From: Greg Stark Organization: The Emacs Conspiracy; member since 1992 Date: 06 Jan 2004 13:12:17 -0500 Message-ID: <878ykl0wvy.fsf@stark.dyndns.tv> Lines: 21 User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/69 X-Sequence-Number: 5244 Clive Page writes: > SELECT myfunc(mycol) FROM table LIMIT 50 OFFSET 10000 ; > It looks as if OFFSET is implemented just be throwing away the results, > until the OFFSET has been reached. > > It would be nice if OFFSET could be implemented in some more efficient > way. You could do something like: select myfunc(mycol) from (select mycol from table limit 50 offset 10000) as x; I think it's not easy for the optimizer to do it because there are lots of cases where it can't. Consider if you had an ORDER BY clause on the myfunc output column for example. Or if myfunc was a set-returning function. -- greg From pgsql-performance-owner@postgresql.org Tue Jan 6 14:37:39 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 4D21DD1B511 for ; Tue, 6 Jan 2004 18:37:38 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 39670-07 for ; Tue, 6 Jan 2004 14:36:50 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id B1E87D1B43A for ; Tue, 6 Jan 2004 14:36:48 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i06Iaj19022306; Tue, 6 Jan 2004 13:36:45 -0500 (EST) To: Greg Stark Cc: Clive Page , pgsql-performance@postgresql.org Subject: Re: Inefficient SELECT with OFFSET and LIMIT In-reply-to: <878ykl0wvy.fsf@stark.dyndns.tv> References: <878ykl0wvy.fsf@stark.dyndns.tv> Comments: In-reply-to Greg Stark message dated "06 Jan 2004 13:12:17 -0500" Date: Tue, 06 Jan 2004 13:36:44 -0500 Message-ID: <22305.1073414204@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/70 X-Sequence-Number: 5245 Greg Stark writes: > Clive Page writes: >> It would be nice if OFFSET could be implemented in some more efficient >> way. > You could do something like: > select myfunc(mycol) from (select mycol from table limit 50 offset 10000) as x; Note that this won't eliminate the major inefficiency, which is having to read 10000+50 rows from the table. But if myfunc() has side-effects or is very expensive to run, it'd probably be worth doing. > I think it's not easy for the optimizer to do it because there are lots of > cases where it can't. I don't actually know of any cases where it could do much of anything to avoid fetching the OFFSET rows. The problems are basically the same as with COUNT(*) optimization: without examining each row, you don't know if it would have been returned or not. We could possibly postpone evaluation of the SELECT output list until after the OFFSET step (thus automating the above hack), but even that only works if there are no set-returning functions in the output list ... regards, tom lane PS: BTW, the one-extra-row effect that Clive noted is gone in 7.4. From pgsql-performance-owner@postgresql.org Tue Jan 6 15:29:21 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 4817AD1B438 for ; Tue, 6 Jan 2004 19:29:19 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 53128-06 for ; Tue, 6 Jan 2004 15:28:30 -0400 (AST) Received: from linda-2.paradise.net.nz (bm-2a.paradise.net.nz [202.0.58.21]) by svr1.postgresql.org (Postfix) with ESMTP id F2FCAD1B4D4 for ; Tue, 6 Jan 2004 15:28:14 -0400 (AST) Received: from smtp-3.paradise.net.nz (smtp-3a.paradise.net.nz [202.0.32.196]) by linda-2.paradise.net.nz (Paradise.net.nz) with ESMTP id <0HR300A9022WT0@linda-2.paradise.net.nz> for pgsql-performance@postgresql.org; Wed, 07 Jan 2004 08:28:09 +1300 (NZDT) Received: from paradise.net.nz (218-101-13-98.paradise.net.nz [218.101.13.98]) by smtp-3.paradise.net.nz (Postfix) with ESMTP id D2943ADF4A; Wed, 07 Jan 2004 08:28:07 +1300 (NZDT) Date: Wed, 07 Jan 2004 08:28:45 +1300 From: Mark Kirkwood Subject: Re: Select max(foo) and select count(*) optimization In-reply-to: <200401060718.08444.darcy@druid.net> To: "D'Arcy J.M. Cain" Cc: Shridhar Daithankar , Postgresql Performance Message-id: <3FFB0C6D.9020704@paradise.net.nz> MIME-version: 1.0 Content-type: text/plain; format=flowed; charset=ISO-8859-1 Content-transfer-encoding: 7bit X-Accept-Language: en-us, en User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.5) Gecko/20031213 References: <1073332325.8958.8.camel@jester> <200401061212.21344.shridhar_daithankar@persistent.co.in> <200401060718.08444.darcy@druid.net> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/71 X-Sequence-Number: 5246 if this situation persists after 'analyze certificate', then you need to: increase the statistics target 'alter table certificate alter column certificate_id set statistics 100' or 'vacuum full certificate' i.e : there are lots of (dead) updated or deleted tuples in the relation, distributed in such a way as to throw off analyze's estimate. regards Mark D'Arcy J.M. Cain wrote: > >Well, I did this: > >cert=# select relpages,reltuples from pg_class where relname= 'certificate'; > relpages | reltuples >----------+------------- > 399070 | 2.48587e+07 >(1 row) > >Casting seemed to help: > >cert=# select relpages,reltuples::bigint from pg_class where relname= >'certificate'; > relpages | reltuples >----------+----------- > 399070 | 24858736 >(1 row) > >But: > >cert=# select count(*) from certificate; >[*Crunch* *Crunch* *Crunch*] > count >---------- > 19684668 >(1 row) > >Am I missing something? Max certificate_id is 20569544 btw. > > > From pgsql-performance-owner@postgresql.org Tue Jan 6 16:50:14 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 981F9D1B45B for ; Tue, 6 Jan 2004 20:50:12 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 72403-03 for ; Tue, 6 Jan 2004 16:49:25 -0400 (AST) Received: from mail.ncsa.uiuc.edu (mail.ncsa.uiuc.edu [141.142.2.28]) by svr1.postgresql.org (Postfix) with ESMTP id F3BF9D1B436 for ; Tue, 6 Jan 2004 16:49:20 -0400 (AST) X-Envelope-From: mshapiro@ncsa.edu X-Envelope-To: Received: from kalika.ncsa.edu (kalika.ncsa.uiuc.edu [141.142.97.63]) by mail.ncsa.uiuc.edu (8.11.6p2/8.11.6) with ESMTP id i06KnHC26086 for ; Tue, 6 Jan 2004 14:49:17 -0600 Message-Id: <5.2.0.9.2.20040106144745.00bcfe30@pop.ncsa.uiuc.edu> X-Sender: mshapiro@pop.ncsa.uiuc.edu (Unverified) X-Mailer: QUALCOMM Windows Eudora Version 5.2.0.9 Date: Tue, 06 Jan 2004 14:49:16 -0600 To: pgsql-performance@postgresql.org From: Michael Shapiro Subject: PgAdmin startup query VERY slow Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="=====================_15580593==_" X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/72 X-Sequence-Number: 5247 --=====================_15580593==_ Content-Type: text/plain; charset="us-ascii"; format=flowed I have reported this on the pgadmin-support mailing list, but Andreas Pflug has asked me to post it here. With a particular database, PgAdmin3 takes a very long time to connect to a database. this is not a general problem with PgAdmin, but only with one database out of many. Other databases do not have the problem. And only with one particular server. The exact same database on a different server does not have the problem. The server in question is running PostgreSQL 7.3.2 on sparc-sun-solaris2.8, compiled by GCC 2.95.2 The other server which has the same database is running Postgres 7.3.4 on i386-redhat-linux-gnu, complied by GCC i386-redhat-linux-gcc 3.2.2. I have attached the query that Andreas says is the one that is run when PgAdmin first connects to a database as well as the output from running the query with explain turned on. Both Andreas and I would be every interested if this group might have any ideas why the query is so slow. NOTE: I have vacuumed the database, but that did not affect the timing at all. NOTE: The startup on the sparc server is 44 seconds, The startup on the linux server is 5 seconds. Andreas writes: I can't see too much from this query plan, it just seems you have 321 triggers an 4750 dependencies which isn't too extraordinary much. But 48 seconds execution time *is* much. Please repost this to pgsql-performance, including the query, backend version, and modified server settings. I'm not deep enough in planner items to analyze this sufficiently. Please let me CCd on this topic so I can see what I should change in pgAdmin3 (if any). --- Michael --- Michael --=====================_15580593==_ Content-Type: text/plain; charset="us-ascii" Content-Disposition: attachment; filename="query.txt" SELECT COUNT(*) FROM (SELECT tgargs from pg_trigger tr LEFT JOIN pg_depend dep ON dep.objid=tr.oid AND deptype = 'i' LEFT JOIN pg_constraint co ON refobjid = co.oid AND contype = 'f' WHERE co.oid IS NULL GROUP BY tgargs HAVING count(1) = 3) AS foo --=====================_15580593==_ Content-Type: text/plain; name="expalin.txt"; x-mac-type="42494E41"; x-mac-creator="74747874" Content-Transfer-Encoding: base64 Content-Disposition: attachment; filename="expalin.txt" UVVFUlkgUExBTgogICB7IEFHRyAKICAgOnN0YXJ0dXBfY29zdCAxODMuMzYg CiAgIDp0b3RhbF9jb3N0IDE4My4zNiAKICAgOnJvd3MgMSAKICAgOndpZHRo IDIyIAogICA6cXB0YXJnZXRsaXN0ICgKICAgICAgeyBUQVJHRVRFTlRSWSAK ICAgICAgOnJlc2RvbSAKICAgICAgICAgeyBSRVNET00gCiAgICAgICAgIDpy ZXNubyAxIAogICAgICAgICA6cmVzdHlwZSAyMCAKICAgICAgICAgOnJlc3R5 cG1vZCAtMSAKICAgICAgICAgOnJlc25hbWUgY291bnQgCiAgICAgICAgIDpy ZXNrZXkgMCAKICAgICAgICAgOnJlc2tleW9wIDAgCiAgICAgICAgIDpyZXNz b3J0Z3JvdXByZWYgMCAKICAgICAgICAgOnJlc2p1bmsgZmFsc2UgCiAgICAg ICAgIH0KICAgICAgIAogICAgICA6ZXhwciAKICAgICAgICAgeyBBR0dSRUcg CiAgICAgICAgIDphZ2dmbm9pZCAyMTQ3IAogICAgICAgICA6YWdndHlwZSAy MCAKICAgICAgICAgOnRhcmdldCAKICAgICAgICAgICAgeyBDT05TVCAKICAg ICAgICAgICAgOmNvbnN0dHlwZSAyMyAKICAgICAgICAgICAgOmNvbnN0bGVu IDQgCiAgICAgICAgICAgIDpjb25zdGJ5dmFsIHRydWUgCiAgICAgICAgICAg IDpjb25zdGlzbnVsbCBmYWxzZSAKICAgICAgICAgICAgOmNvbnN0dmFsdWUg IDQgWyAwIDAgMCAxIF0gCiAgICAgICAgICAgIH0KICAgICAgICAgIAogICAg ICAgICA6YWdnc3RhciB0cnVlIAogICAgICAgICA6YWdnZGlzdGluY3QgZmFs c2UgCiAgICAgICAgIH0KICAgICAgfQogICApCiAgICAKICAgOnFwcXVhbCA8 PiAKICAgOmxlZnR0cmVlIAogICAgICB7IFNVQlFVRVJZU0NBTiAKICAgICAg OnN0YXJ0dXBfY29zdCAxODMuMzQgCiAgICAgIDp0b3RhbF9jb3N0IDE4My4z NiAKICAgICAgOnJvd3MgMSAKICAgICAgOndpZHRoIDIyIAogICAgICA6cXB0 YXJnZXRsaXN0IDw+IAogICAgICA6cXBxdWFsIDw+IAogICAgICA6bGVmdHRy ZWUgPD4gCiAgICAgIDpyaWdodHRyZWUgPD4gCiAgICAgIDpleHRwcm0gKCkK ICAgICAgIAogICAgICA6bG9jcHJtICgpCiAgICAgICAKICAgICAgOmluaXRw bGFuIDw+IAogICAgICA6bnBybSAwICAKICAgICAgOnNjYW5yZWxpZCAxIAog ICAgICA6c3VicGxhbiAKICAgICAgICAgeyBBR0cgCiAgICAgICAgIDpzdGFy dHVwX2Nvc3QgMTgzLjM0IAogICAgICAgICA6dG90YWxfY29zdCAxODMuMzYg CiAgICAgICAgIDpyb3dzIDEgCiAgICAgICAgIDp3aWR0aCAyMiAKICAgICAg ICAgOnFwdGFyZ2V0bGlzdCAoCiAgICAgICAgICAgIHsgVEFSR0VURU5UUlkg CiAgICAgICAgICAgIDpyZXNkb20gCiAgICAgICAgICAgICAgIHsgUkVTRE9N IAogICAgICAgICAgICAgICA6cmVzbm8gMSAKICAgICAgICAgICAgICAgOnJl c3R5cGUgMTcgCiAgICAgICAgICAgICAgIDpyZXN0eXBtb2QgLTEgCiAgICAg ICAgICAgICAgIDpyZXNuYW1lIHRnYXJncyAKICAgICAgICAgICAgICAgOnJl c2tleSAwIAogICAgICAgICAgICAgICA6cmVza2V5b3AgMCAKICAgICAgICAg ICAgICAgOnJlc3NvcnRncm91cHJlZiAxIAogICAgICAgICAgICAgICA6cmVz anVuayBmYWxzZSAKICAgICAgICAgICAgICAgfQogICAgICAgICAgICAgCiAg ICAgICAgICAgIDpleHByIAogICAgICAgICAgICAgICB7IFZBUiAKICAgICAg ICAgICAgICAgOnZhcm5vIDAgCiAgICAgICAgICAgICAgIDp2YXJhdHRubyAx IAogICAgICAgICAgICAgICA6dmFydHlwZSAxNyAKICAgICAgICAgICAgICAg OnZhcnR5cG1vZCAtMSAgCiAgICAgICAgICAgICAgIDp2YXJsZXZlbHN1cCAw IAogICAgICAgICAgICAgICA6dmFybm9vbGQgMSAKICAgICAgICAgICAgICAg OnZhcm9hdHRubyAxMwogICAgICAgICAgICAgICB9CiAgICAgICAgICAgIH0K ICAgICAgICAgKQogICAgICAgICAgCiAgICAgICAgIDpxcHF1YWwgKAogICAg ICAgICAgICB7IEVYUFIgCiAgICAgICAgICAgIDp0eXBlT2lkIDE2ICAKICAg ICAgICAgICAgOm9wVHlwZSBvcCAKICAgICAgICAgICAgOm9wZXIgCiAgICAg ICAgICAgICAgIHsgT1BFUiAKICAgICAgICAgICAgICAgOm9wbm8gNDE2IAog ICAgICAgICAgICAgICA6b3BpZCA0NzQgCiAgICAgICAgICAgICAgIDpvcHJl c3VsdHR5cGUgMTYgCiAgICAgICAgICAgICAgIDpvcHJldHNldCBmYWxzZSAK ICAgICAgICAgICAgICAgfQogICAgICAgICAgICAgCiAgICAgICAgICAgIDph cmdzICgKICAgICAgICAgICAgICAgeyBBR0dSRUcgCiAgICAgICAgICAgICAg IDphZ2dmbm9pZCAyMTQ3IAogICAgICAgICAgICAgICA6YWdndHlwZSAyMCAK ICAgICAgICAgICAgICAgOnRhcmdldCAKICAgICAgICAgICAgICAgICAgeyBD T05TVCAKICAgICAgICAgICAgICAgICAgOmNvbnN0dHlwZSAyMyAKICAgICAg ICAgICAgICAgICAgOmNvbnN0bGVuIDQgCiAgICAgICAgICAgICAgICAgIDpj b25zdGJ5dmFsIHRydWUgCiAgICAgICAgICAgICAgICAgIDpjb25zdGlzbnVs bCBmYWxzZSAKICAgICAgICAgICAgICAgICAgOmNvbnN0dmFsdWUgIDQgWyAw IDAgMCAxIF0gCiAgICAgICAgICAgICAgICAgIH0KICAgICAgICAgICAgICAg IAogICAgICAgICAgICAgICA6YWdnc3RhciBmYWxzZSAKICAgICAgICAgICAg ICAgOmFnZ2Rpc3RpbmN0IGZhbHNlIAogICAgICAgICAgICAgICB9CiAgICAg ICAgICAgICAKICAgICAgICAgICAgICAgeyBDT05TVCAKICAgICAgICAgICAg ICAgOmNvbnN0dHlwZSAyMyAKICAgICAgICAgICAgICAgOmNvbnN0bGVuIDQg CiAgICAgICAgICAgICAgIDpjb25zdGJ5dmFsIHRydWUgCiAgICAgICAgICAg ICAgIDpjb25zdGlzbnVsbCBmYWxzZSAKICAgICAgICAgICAgICAgOmNvbnN0 dmFsdWUgIDQgWyAwIDAgMCAzIF0gCiAgICAgICAgICAgICAgIH0KICAgICAg ICAgICAgKQogICAgICAgICAgICB9CiAgICAgICAgICkKICAgICAgICAgIAog ICAgICAgICA6bGVmdHRyZWUgCiAgICAgICAgICAgIHsgR1JQIAogICAgICAg ICAgICA6c3RhcnR1cF9jb3N0IDE4My4zNCAKICAgICAgICAgICAgOnRvdGFs X2Nvc3QgMTgzLjM1IAogICAgICAgICAgICA6cm93cyAyIAogICAgICAgICAg ICA6d2lkdGggMjIgCiAgICAgICAgICAgIDpxcHRhcmdldGxpc3QgKAogICAg ICAgICAgICAgICB7IFRBUkdFVEVOVFJZIAogICAgICAgICAgICAgICA6cmVz ZG9tIAogICAgICAgICAgICAgICAgICB7IFJFU0RPTSAKICAgICAgICAgICAg ICAgICAgOnJlc25vIDEgCiAgICAgICAgICAgICAgICAgIDpyZXN0eXBlIDE3 IAogICAgICAgICAgICAgICAgICA6cmVzdHlwbW9kIC0xIAogICAgICAgICAg ICAgICAgICA6cmVzbmFtZSA8PiAKICAgICAgICAgICAgICAgICAgOnJlc2tl eSAwIAogICAgICAgICAgICAgICAgICA6cmVza2V5b3AgMCAKICAgICAgICAg ICAgICAgICAgOnJlc3NvcnRncm91cHJlZiAwIAogICAgICAgICAgICAgICAg ICA6cmVzanVuayBmYWxzZSAKICAgICAgICAgICAgICAgICAgfQogICAgICAg ICAgICAgICAgCiAgICAgICAgICAgICAgIDpleHByIAogICAgICAgICAgICAg ICAgICB7IFZBUiAKICAgICAgICAgICAgICAgICAgOnZhcm5vIDAgCiAgICAg ICAgICAgICAgICAgIDp2YXJhdHRubyAxIAogICAgICAgICAgICAgICAgICA6 dmFydHlwZSAxNyAKICAgICAgICAgICAgICAgICAgOnZhcnR5cG1vZCAtMSAg CiAgICAgICAgICAgICAgICAgIDp2YXJsZXZlbHN1cCAwIAogICAgICAgICAg ICAgICAgICA6dmFybm9vbGQgMSAKICAgICAgICAgICAgICAgICAgOnZhcm9h dHRubyAxMwogICAgICAgICAgICAgICAgICB9CiAgICAgICAgICAgICAgIH0K ICAgICAgICAgICAgKQogICAgICAgICAgICAgCiAgICAgICAgICAgIDpxcHF1 YWwgPD4gCiAgICAgICAgICAgIDpsZWZ0dHJlZSAKICAgICAgICAgICAgICAg eyBTT1JUIAogICAgICAgICAgICAgICA6c3RhcnR1cF9jb3N0IDE4My4zNCAK ICAgICAgICAgICAgICAgOnRvdGFsX2Nvc3QgMTgzLjM1IAogICAgICAgICAg ICAgICA6cm93cyAyIAogICAgICAgICAgICAgICA6d2lkdGggMjIgCiAgICAg ICAgICAgICAgIDpxcHRhcmdldGxpc3QgKAogICAgICAgICAgICAgICAgICB7 IFRBUkdFVEVOVFJZIAogICAgICAgICAgICAgICAgICA6cmVzZG9tIAogICAg ICAgICAgICAgICAgICAgICB7IFJFU0RPTSAKICAgICAgICAgICAgICAgICAg ICAgOnJlc25vIDEgCiAgICAgICAgICAgICAgICAgICAgIDpyZXN0eXBlIDE3 IAogICAgICAgICAgICAgICAgICAgICA6cmVzdHlwbW9kIC0xIAogICAgICAg ICAgICAgICAgICAgICA6cmVzbmFtZSA8PiAKICAgICAgICAgICAgICAgICAg ICAgOnJlc2tleSAxIAogICAgICAgICAgICAgICAgICAgICA6cmVza2V5b3Ag MTk1NyAKICAgICAgICAgICAgICAgICAgICAgOnJlc3NvcnRncm91cHJlZiAw IAogICAgICAgICAgICAgICAgICAgICA6cmVzanVuayBmYWxzZSAKICAgICAg ICAgICAgICAgICAgICAgfQogICAgICAgICAgICAgICAgICAgCiAgICAgICAg ICAgICAgICAgIDpleHByIAogICAgICAgICAgICAgICAgICAgICB7IFZBUiAK ICAgICAgICAgICAgICAgICAgICAgOnZhcm5vIDEgCiAgICAgICAgICAgICAg ICAgICAgIDp2YXJhdHRubyAxMyAKICAgICAgICAgICAgICAgICAgICAgOnZh cnR5cGUgMTcgCiAgICAgICAgICAgICAgICAgICAgIDp2YXJ0eXBtb2QgLTEg IAogICAgICAgICAgICAgICAgICAgICA6dmFybGV2ZWxzdXAgMCAKICAgICAg ICAgICAgICAgICAgICAgOnZhcm5vb2xkIDEgCiAgICAgICAgICAgICAgICAg ICAgIDp2YXJvYXR0bm8gMTMKICAgICAgICAgICAgICAgICAgICAgfQogICAg ICAgICAgICAgICAgICB9CiAgICAgICAgICAgICAgICkKICAgICAgICAgICAg ICAgIAogICAgICAgICAgICAgICA6cXBxdWFsIDw+IAogICAgICAgICAgICAg ICA6bGVmdHRyZWUgCiAgICAgICAgICAgICAgICAgIHsgSEFTSEpPSU4gCiAg ICAgICAgICAgICAgICAgIDpzdGFydHVwX2Nvc3QgMC4wMCAKICAgICAgICAg ICAgICAgICAgOnRvdGFsX2Nvc3QgMTgzLjMzIAogICAgICAgICAgICAgICAg ICA6cm93cyAyIAogICAgICAgICAgICAgICAgICA6d2lkdGggMjIgCiAgICAg ICAgICAgICAgICAgIDpxcHRhcmdldGxpc3QgKAogICAgICAgICAgICAgICAg ICAgICB7IFRBUkdFVEVOVFJZIAogICAgICAgICAgICAgICAgICAgICA6cmVz ZG9tIAogICAgICAgICAgICAgICAgICAgICAgICB7IFJFU0RPTSAKICAgICAg ICAgICAgICAgICAgICAgICAgOnJlc25vIDEgCiAgICAgICAgICAgICAgICAg ICAgICAgIDpyZXN0eXBlIDE3IAogICAgICAgICAgICAgICAgICAgICAgICA6 cmVzdHlwbW9kIC0xIAogICAgICAgICAgICAgICAgICAgICAgICA6cmVzbmFt ZSA8PiAKICAgICAgICAgICAgICAgICAgICAgICAgOnJlc2tleSAwIAogICAg ICAgICAgICAgICAgICAgICAgICA6cmVza2V5b3AgMCAKICAgICAgICAgICAg ICAgICAgICAgICAgOnJlc3NvcnRncm91cHJlZiAwIAogICAgICAgICAgICAg ICAgICAgICAgICA6cmVzanVuayBmYWxzZSAKICAgICAgICAgICAgICAgICAg ICAgICAgfQogICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAg ICAgICAgIDpleHByIAogICAgICAgICAgICAgICAgICAgICAgICB7IFZBUiAK ICAgICAgICAgICAgICAgICAgICAgICAgOnZhcm5vIDY1MDAxIAogICAgICAg ICAgICAgICAgICAgICAgICA6dmFyYXR0bm8gMSAKICAgICAgICAgICAgICAg ICAgICAgICAgOnZhcnR5cGUgMTcgCiAgICAgICAgICAgICAgICAgICAgICAg IDp2YXJ0eXBtb2QgLTEgIAogICAgICAgICAgICAgICAgICAgICAgICA6dmFy bGV2ZWxzdXAgMCAKICAgICAgICAgICAgICAgICAgICAgICAgOnZhcm5vb2xk IDEgCiAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJvYXR0bm8gMTMKICAg ICAgICAgICAgICAgICAgICAgICAgfQogICAgICAgICAgICAgICAgICAgICB9 CiAgICAgICAgICAgICAgICAgICkKICAgICAgICAgICAgICAgICAgIAogICAg ICAgICAgICAgICAgICA6cXBxdWFsICgKICAgICAgICAgICAgICAgICAgICAg eyBOVUxMVEVTVCAKICAgICAgICAgICAgICAgICAgICAgOmFyZyAKICAgICAg ICAgICAgICAgICAgICAgICAgeyBWQVIgCiAgICAgICAgICAgICAgICAgICAg ICAgIDp2YXJubyA2NTAwMCAKICAgICAgICAgICAgICAgICAgICAgICAgOnZh cmF0dG5vIDEgCiAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJ0eXBlIDI2 IAogICAgICAgICAgICAgICAgICAgICAgICA6dmFydHlwbW9kIC0xICAKICAg ICAgICAgICAgICAgICAgICAgICAgOnZhcmxldmVsc3VwIDAgCiAgICAgICAg ICAgICAgICAgICAgICAgIDp2YXJub29sZCA0IAogICAgICAgICAgICAgICAg ICAgICAgICA6dmFyb2F0dG5vIC0yCiAgICAgICAgICAgICAgICAgICAgICAg IH0KICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAg ICA6bnVsbHRlc3R0eXBlIDAgCiAgICAgICAgICAgICAgICAgICAgIH0KICAg ICAgICAgICAgICAgICAgKQogICAgICAgICAgICAgICAgICAgCiAgICAgICAg ICAgICAgICAgIDpsZWZ0dHJlZSAKICAgICAgICAgICAgICAgICAgICAgeyBO RVNUTE9PUCAKICAgICAgICAgICAgICAgICAgICAgOnN0YXJ0dXBfY29zdCAw LjAwIAogICAgICAgICAgICAgICAgICAgICA6dG90YWxfY29zdCAxODMuMzIg CiAgICAgICAgICAgICAgICAgICAgIDpyb3dzIDIgCiAgICAgICAgICAgICAg ICAgICAgIDp3aWR0aCAxNyAKICAgICAgICAgICAgICAgICAgICAgOnFwdGFy Z2V0bGlzdCAoCiAgICAgICAgICAgICAgICAgICAgICAgIHsgVEFSR0VURU5U UlkgCiAgICAgICAgICAgICAgICAgICAgICAgIDpyZXNkb20gCiAgICAgICAg ICAgICAgICAgICAgICAgICAgIHsgUkVTRE9NIAogICAgICAgICAgICAgICAg ICAgICAgICAgICA6cmVzbm8gMSAKICAgICAgICAgICAgICAgICAgICAgICAg ICAgOnJlc3R5cGUgMTcgCiAgICAgICAgICAgICAgICAgICAgICAgICAgIDpy ZXN0eXBtb2QgLTEgCiAgICAgICAgICAgICAgICAgICAgICAgICAgIDpyZXNu YW1lIDw+IAogICAgICAgICAgICAgICAgICAgICAgICAgICA6cmVza2V5IDAg CiAgICAgICAgICAgICAgICAgICAgICAgICAgIDpyZXNrZXlvcCAwIAogICAg ICAgICAgICAgICAgICAgICAgICAgICA6cmVzc29ydGdyb3VwcmVmIDAgCiAg ICAgICAgICAgICAgICAgICAgICAgICAgIDpyZXNqdW5rIGZhbHNlIAogICAg ICAgICAgICAgICAgICAgICAgICAgICB9CiAgICAgICAgICAgICAgICAgICAg ICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgOmV4cHIgCiAgICAgICAg ICAgICAgICAgICAgICAgICAgIHsgVkFSIAogICAgICAgICAgICAgICAgICAg ICAgICAgICA6dmFybm8gNjUwMDEgCiAgICAgICAgICAgICAgICAgICAgICAg ICAgIDp2YXJhdHRubyAxIAogICAgICAgICAgICAgICAgICAgICAgICAgICA6 dmFydHlwZSAxNyAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZhcnR5 cG1vZCAtMSAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJsZXZl bHN1cCAwIAogICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFybm9vbGQg MSAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZhcm9hdHRubyAxMwog ICAgICAgICAgICAgICAgICAgICAgICAgICB9CiAgICAgICAgICAgICAgICAg ICAgICAgIH0KICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAg ICAgICAgICAgICB7IFRBUkdFVEVOVFJZIAogICAgICAgICAgICAgICAgICAg ICAgICA6cmVzZG9tIAogICAgICAgICAgICAgICAgICAgICAgICAgICB7IFJF U0RPTSAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnJlc25vIDIgCiAg ICAgICAgICAgICAgICAgICAgICAgICAgIDpyZXN0eXBlIDI2IAogICAgICAg ICAgICAgICAgICAgICAgICAgICA6cmVzdHlwbW9kIC0xIAogICAgICAgICAg ICAgICAgICAgICAgICAgICA6cmVzbmFtZSA8PiAKICAgICAgICAgICAgICAg ICAgICAgICAgICAgOnJlc2tleSAwIAogICAgICAgICAgICAgICAgICAgICAg ICAgICA6cmVza2V5b3AgMCAKICAgICAgICAgICAgICAgICAgICAgICAgICAg OnJlc3NvcnRncm91cHJlZiAwIAogICAgICAgICAgICAgICAgICAgICAgICAg ICA6cmVzanVuayBmYWxzZSAKICAgICAgICAgICAgICAgICAgICAgICAgICAg fQogICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAg ICAgICAgIDpleHByIAogICAgICAgICAgICAgICAgICAgICAgICAgICB7IFZB UiAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZhcm5vIDY1MDAxIAog ICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFyYXR0bm8gMiAKICAgICAg ICAgICAgICAgICAgICAgICAgICAgOnZhcnR5cGUgMjYgCiAgICAgICAgICAg ICAgICAgICAgICAgICAgIDp2YXJ0eXBtb2QgLTEgIAogICAgICAgICAgICAg ICAgICAgICAgICAgICA6dmFybGV2ZWxzdXAgMCAKICAgICAgICAgICAgICAg ICAgICAgICAgICAgOnZhcm5vb2xkIDEgCiAgICAgICAgICAgICAgICAgICAg ICAgICAgIDp2YXJvYXR0bm8gLTIKICAgICAgICAgICAgICAgICAgICAgICAg ICAgfQogICAgICAgICAgICAgICAgICAgICAgICB9CiAgICAgICAgICAgICAg ICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgeyBUQVJHRVRFTlRS WSAKICAgICAgICAgICAgICAgICAgICAgICAgOnJlc2RvbSAKICAgICAgICAg ICAgICAgICAgICAgICAgICAgeyBSRVNET00gCiAgICAgICAgICAgICAgICAg ICAgICAgICAgIDpyZXNubyAzIAogICAgICAgICAgICAgICAgICAgICAgICAg ICA6cmVzdHlwZSAyNiAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnJl c3R5cG1vZCAtMSAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnJlc25h bWUgPD4gCiAgICAgICAgICAgICAgICAgICAgICAgICAgIDpyZXNrZXkgMCAK ICAgICAgICAgICAgICAgICAgICAgICAgICAgOnJlc2tleW9wIDAgCiAgICAg ICAgICAgICAgICAgICAgICAgICAgIDpyZXNzb3J0Z3JvdXByZWYgMCAKICAg ICAgICAgICAgICAgICAgICAgICAgICAgOnJlc2p1bmsgZmFsc2UgCiAgICAg ICAgICAgICAgICAgICAgICAgICAgIH0KICAgICAgICAgICAgICAgICAgICAg ICAgIAogICAgICAgICAgICAgICAgICAgICAgICA6ZXhwciAKICAgICAgICAg ICAgICAgICAgICAgICAgICAgeyBWQVIgCiAgICAgICAgICAgICAgICAgICAg ICAgICAgIDp2YXJubyA2NTAwMCAKICAgICAgICAgICAgICAgICAgICAgICAg ICAgOnZhcmF0dG5vIDEgCiAgICAgICAgICAgICAgICAgICAgICAgICAgIDp2 YXJ0eXBlIDI2IAogICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFydHlw bW9kIC0xICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZhcmxldmVs c3VwIDAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJub29sZCAy IAogICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFyb2F0dG5vIDIKICAg ICAgICAgICAgICAgICAgICAgICAgICAgfQogICAgICAgICAgICAgICAgICAg ICAgICB9CiAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAg ICAgICAgICAgeyBUQVJHRVRFTlRSWSAKICAgICAgICAgICAgICAgICAgICAg ICAgOnJlc2RvbSAKICAgICAgICAgICAgICAgICAgICAgICAgICAgeyBSRVNE T00gCiAgICAgICAgICAgICAgICAgICAgICAgICAgIDpyZXNubyA0IAogICAg ICAgICAgICAgICAgICAgICAgICAgICA6cmVzdHlwZSAxOCAKICAgICAgICAg ICAgICAgICAgICAgICAgICAgOnJlc3R5cG1vZCAtMSAKICAgICAgICAgICAg ICAgICAgICAgICAgICAgOnJlc25hbWUgPD4gCiAgICAgICAgICAgICAgICAg ICAgICAgICAgIDpyZXNrZXkgMCAKICAgICAgICAgICAgICAgICAgICAgICAg ICAgOnJlc2tleW9wIDAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgIDpy ZXNzb3J0Z3JvdXByZWYgMCAKICAgICAgICAgICAgICAgICAgICAgICAgICAg OnJlc2p1bmsgZmFsc2UgCiAgICAgICAgICAgICAgICAgICAgICAgICAgIH0K ICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAg ICAgICA6ZXhwciAKICAgICAgICAgICAgICAgICAgICAgICAgICAgeyBWQVIg CiAgICAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJubyA2NTAwMCAKICAg ICAgICAgICAgICAgICAgICAgICAgICAgOnZhcmF0dG5vIDIgCiAgICAgICAg ICAgICAgICAgICAgICAgICAgIDp2YXJ0eXBlIDE4IAogICAgICAgICAgICAg ICAgICAgICAgICAgICA6dmFydHlwbW9kIC0xICAKICAgICAgICAgICAgICAg ICAgICAgICAgICAgOnZhcmxldmVsc3VwIDAgCiAgICAgICAgICAgICAgICAg ICAgICAgICAgIDp2YXJub29sZCAyIAogICAgICAgICAgICAgICAgICAgICAg ICAgICA6dmFyb2F0dG5vIDcKICAgICAgICAgICAgICAgICAgICAgICAgICAg fQogICAgICAgICAgICAgICAgICAgICAgICB9CiAgICAgICAgICAgICAgICAg ICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgeyBUQVJHRVRFTlRSWSAK ICAgICAgICAgICAgICAgICAgICAgICAgOnJlc2RvbSAKICAgICAgICAgICAg ICAgICAgICAgICAgICAgeyBSRVNET00gCiAgICAgICAgICAgICAgICAgICAg ICAgICAgIDpyZXNubyA1IAogICAgICAgICAgICAgICAgICAgICAgICAgICA6 cmVzdHlwZSAyNiAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnJlc3R5 cG1vZCAtMSAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnJlc25hbWUg PD4gCiAgICAgICAgICAgICAgICAgICAgICAgICAgIDpyZXNrZXkgMCAKICAg ICAgICAgICAgICAgICAgICAgICAgICAgOnJlc2tleW9wIDAgCiAgICAgICAg ICAgICAgICAgICAgICAgICAgIDpyZXNzb3J0Z3JvdXByZWYgMCAKICAgICAg ICAgICAgICAgICAgICAgICAgICAgOnJlc2p1bmsgZmFsc2UgCiAgICAgICAg ICAgICAgICAgICAgICAgICAgIH0KICAgICAgICAgICAgICAgICAgICAgICAg IAogICAgICAgICAgICAgICAgICAgICAgICA6ZXhwciAKICAgICAgICAgICAg ICAgICAgICAgICAgICAgeyBWQVIgCiAgICAgICAgICAgICAgICAgICAgICAg ICAgIDp2YXJubyA2NTAwMCAKICAgICAgICAgICAgICAgICAgICAgICAgICAg OnZhcmF0dG5vIDMgCiAgICAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJ0 eXBlIDI2IAogICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFydHlwbW9k IC0xICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZhcmxldmVsc3Vw IDAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJub29sZCAyIAog ICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFyb2F0dG5vIDUKICAgICAg ICAgICAgICAgICAgICAgICAgICAgfQogICAgICAgICAgICAgICAgICAgICAg ICB9CiAgICAgICAgICAgICAgICAgICAgICkKICAgICAgICAgICAgICAgICAg ICAgIAogICAgICAgICAgICAgICAgICAgICA6cXBxdWFsIDw+IAogICAgICAg ICAgICAgICAgICAgICA6bGVmdHRyZWUgCiAgICAgICAgICAgICAgICAgICAg ICAgIHsgU0VRU0NBTiAKICAgICAgICAgICAgICAgICAgICAgICAgOnN0YXJ0 dXBfY29zdCAwLjAwIAogICAgICAgICAgICAgICAgICAgICAgICA6dG90YWxf Y29zdCAxLjAyIAogICAgICAgICAgICAgICAgICAgICAgICA6cm93cyAyIAog ICAgICAgICAgICAgICAgICAgICAgICA6d2lkdGggOCAKICAgICAgICAgICAg ICAgICAgICAgICAgOnFwdGFyZ2V0bGlzdCAoCiAgICAgICAgICAgICAgICAg ICAgICAgICAgIHsgVEFSR0VURU5UUlkgCiAgICAgICAgICAgICAgICAgICAg ICAgICAgIDpyZXNkb20gCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAg IHsgUkVTRE9NIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICA6cmVz bm8gMSAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgOnJlc3R5cGUg MTcgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIDpyZXN0eXBtb2Qg LTEgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIDpyZXNuYW1lIDw+ IAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICA6cmVza2V5IDAgCiAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgIDpyZXNrZXlvcCAwIAogICAg ICAgICAgICAgICAgICAgICAgICAgICAgICA6cmVzc29ydGdyb3VwcmVmIDAg CiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIDpyZXNqdW5rIGZhbHNl IAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICB9CiAgICAgICAgICAg ICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAg OmV4cHIgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIHsgVkFSIAog ICAgICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFybm8gMSAKICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgOnZhcmF0dG5vIDEzIAogICAgICAg ICAgICAgICAgICAgICAgICAgICAgICA6dmFydHlwZSAxNyAKICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgOnZhcnR5cG1vZCAtMSAgCiAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgIDp2YXJsZXZlbHN1cCAwIAogICAgICAg ICAgICAgICAgICAgICAgICAgICAgICA6dmFybm9vbGQgMSAKICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgOnZhcm9hdHRubyAxMwogICAgICAgICAg ICAgICAgICAgICAgICAgICAgICB9CiAgICAgICAgICAgICAgICAgICAgICAg ICAgIH0KICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAg ICAgICAgICAgICAgICB7IFRBUkdFVEVOVFJZIAogICAgICAgICAgICAgICAg ICAgICAgICAgICA6cmVzZG9tIAogICAgICAgICAgICAgICAgICAgICAgICAg ICAgICB7IFJFU0RPTSAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg OnJlc25vIDIgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIDpyZXN0 eXBlIDI2IAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICA6cmVzdHlw bW9kIC0xIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICA6cmVzbmFt ZSA8PiAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgOnJlc2tleSAw IAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICA6cmVza2V5b3AgMCAK ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgOnJlc3NvcnRncm91cHJl ZiAwIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICA6cmVzanVuayBm YWxzZSAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgfQogICAgICAg ICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAg ICAgIDpleHByIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICB7IFZB UiAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZhcm5vIDEgCiAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJhdHRubyAtMiAKICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZhcnR5cGUgMjYgCiAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJ0eXBtb2QgLTEgIAogICAg ICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFybGV2ZWxzdXAgMCAKICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZhcm5vb2xkIDEgCiAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJvYXR0bm8gLTIKICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgfQogICAgICAgICAgICAgICAgICAg ICAgICAgICB9CiAgICAgICAgICAgICAgICAgICAgICAgICkKICAgICAgICAg ICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICA6cXBx dWFsIDw+IAogICAgICAgICAgICAgICAgICAgICAgICA6bGVmdHRyZWUgPD4g CiAgICAgICAgICAgICAgICAgICAgICAgIDpyaWdodHRyZWUgPD4gCiAgICAg ICAgICAgICAgICAgICAgICAgIDpleHRwcm0gKCkKICAgICAgICAgICAgICAg ICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICA6bG9jcHJtICgp CiAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAg ICAgICAgOmluaXRwbGFuIDw+IAogICAgICAgICAgICAgICAgICAgICAgICA6 bnBybSAwICAKICAgICAgICAgICAgICAgICAgICAgICAgOnNjYW5yZWxpZCAx IAogICAgICAgICAgICAgICAgICAgICAgICB9CiAgICAgICAgICAgICAgICAg ICAgICAKICAgICAgICAgICAgICAgICAgICAgOnJpZ2h0dHJlZSAKICAgICAg ICAgICAgICAgICAgICAgICAgeyBTRVFTQ0FOIAogICAgICAgICAgICAgICAg ICAgICAgICA6c3RhcnR1cF9jb3N0IDAuMDAgCiAgICAgICAgICAgICAgICAg ICAgICAgIDp0b3RhbF9jb3N0IDQ5LjA2IAogICAgICAgICAgICAgICAgICAg ICAgICA6cm93cyAyODA2IAogICAgICAgICAgICAgICAgICAgICAgICA6d2lk dGggOSAKICAgICAgICAgICAgICAgICAgICAgICAgOnFwdGFyZ2V0bGlzdCAo CiAgICAgICAgICAgICAgICAgICAgICAgICAgIHsgVEFSR0VURU5UUlkgCiAg ICAgICAgICAgICAgICAgICAgICAgICAgIDpyZXNkb20gCiAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgIHsgUkVTRE9NIAogICAgICAgICAgICAgICAg ICAgICAgICAgICAgICA6cmVzbm8gMSAKICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgOnJlc3R5cGUgMjYgCiAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgIDpyZXN0eXBtb2QgLTEgCiAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgIDpyZXNuYW1lIDw+IAogICAgICAgICAgICAgICAgICAgICAgICAg ICAgICA6cmVza2V5IDAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAg IDpyZXNrZXlvcCAwIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICA6 cmVzc29ydGdyb3VwcmVmIDAgCiAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgIDpyZXNqdW5rIGZhbHNlIAogICAgICAgICAgICAgICAgICAgICAgICAg ICAgICB9CiAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAg ICAgICAgICAgICAgICAgICAgOmV4cHIgCiAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgIHsgVkFSIAogICAgICAgICAgICAgICAgICAgICAgICAgICAg ICA6dmFybm8gMiAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZh cmF0dG5vIDIgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJ0 eXBlIDI2IAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFydHlw bW9kIC0xICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZhcmxl dmVsc3VwIDAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJu b29sZCAyIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFyb2F0 dG5vIDIKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgfQogICAgICAg ICAgICAgICAgICAgICAgICAgICB9CiAgICAgICAgICAgICAgICAgICAgICAg ICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgeyBUQVJHRVRFTlRSWSAK ICAgICAgICAgICAgICAgICAgICAgICAgICAgOnJlc2RvbSAKICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgeyBSRVNET00gCiAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgIDpyZXNubyAyIAogICAgICAgICAgICAgICAgICAg ICAgICAgICAgICA6cmVzdHlwZSAxOCAKICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgOnJlc3R5cG1vZCAtMSAKICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgOnJlc25hbWUgPD4gCiAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgIDpyZXNrZXkgMCAKICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgOnJlc2tleW9wIDAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAg IDpyZXNzb3J0Z3JvdXByZWYgMCAKICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgOnJlc2p1bmsgZmFsc2UgCiAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgIH0KICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAg ICAgICAgICAgICAgICAgICAgICA6ZXhwciAKICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgeyBWQVIgCiAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgIDp2YXJubyAyIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICA6 dmFyYXR0bm8gNyAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZh cnR5cGUgMTggCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJ0 eXBtb2QgLTEgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFy bGV2ZWxzdXAgMCAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZh cm5vb2xkIDIgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJv YXR0bm8gNwogICAgICAgICAgICAgICAgICAgICAgICAgICAgICB9CiAgICAg ICAgICAgICAgICAgICAgICAgICAgIH0KICAgICAgICAgICAgICAgICAgICAg ICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICB7IFRBUkdFVEVOVFJZ IAogICAgICAgICAgICAgICAgICAgICAgICAgICA6cmVzZG9tIAogICAgICAg ICAgICAgICAgICAgICAgICAgICAgICB7IFJFU0RPTSAKICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgOnJlc25vIDMgCiAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgIDpyZXN0eXBlIDI2IAogICAgICAgICAgICAgICAgICAg ICAgICAgICAgICA6cmVzdHlwbW9kIC0xIAogICAgICAgICAgICAgICAgICAg ICAgICAgICAgICA6cmVzbmFtZSA8PiAKICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgOnJlc2tleSAwIAogICAgICAgICAgICAgICAgICAgICAgICAg ICAgICA6cmVza2V5b3AgMCAKICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgOnJlc3NvcnRncm91cHJlZiAwIAogICAgICAgICAgICAgICAgICAgICAg ICAgICAgICA6cmVzanVuayBmYWxzZSAKICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgfQogICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAg ICAgICAgICAgICAgICAgICAgICAgIDpleHByIAogICAgICAgICAgICAgICAg ICAgICAgICAgICAgICB7IFZBUiAKICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgOnZhcm5vIDIgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAg IDp2YXJhdHRubyA1IAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICA6 dmFydHlwZSAyNiAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZh cnR5cG1vZCAtMSAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIDp2 YXJsZXZlbHN1cCAwIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICA6 dmFybm9vbGQgMiAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZh cm9hdHRubyA1CiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIH0KICAg ICAgICAgICAgICAgICAgICAgICAgICAgfQogICAgICAgICAgICAgICAgICAg ICAgICApCiAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAg ICAgICAgICAgICAgOnFwcXVhbCA8PiAKICAgICAgICAgICAgICAgICAgICAg ICAgOmxlZnR0cmVlIDw+IAogICAgICAgICAgICAgICAgICAgICAgICA6cmln aHR0cmVlIDw+IAogICAgICAgICAgICAgICAgICAgICAgICA6ZXh0cHJtICgp CiAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAg ICAgICAgOmxvY3BybSAoKQogICAgICAgICAgICAgICAgICAgICAgICAgCiAg ICAgICAgICAgICAgICAgICAgICAgIDppbml0cGxhbiA8PiAKICAgICAgICAg ICAgICAgICAgICAgICAgOm5wcm0gMCAgCiAgICAgICAgICAgICAgICAgICAg ICAgIDpzY2FucmVsaWQgMiAKICAgICAgICAgICAgICAgICAgICAgICAgfQog ICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgIDpl eHRwcm0gKCkKICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAg ICAgICAgICA6bG9jcHJtICgpCiAgICAgICAgICAgICAgICAgICAgICAKICAg ICAgICAgICAgICAgICAgICAgOmluaXRwbGFuIDw+IAogICAgICAgICAgICAg ICAgICAgICA6bnBybSAwICAKICAgICAgICAgICAgICAgICAgICAgOmpvaW50 eXBlIDEgCiAgICAgICAgICAgICAgICAgICAgIDpqb2lucXVhbCAoCiAgICAg ICAgICAgICAgICAgICAgICAgIHsgRVhQUiAKICAgICAgICAgICAgICAgICAg ICAgICAgOnR5cGVPaWQgMTYgIAogICAgICAgICAgICAgICAgICAgICAgICA6 b3BUeXBlIG9wIAogICAgICAgICAgICAgICAgICAgICAgICA6b3BlciAKICAg ICAgICAgICAgICAgICAgICAgICAgICAgeyBPUEVSIAogICAgICAgICAgICAg ICAgICAgICAgICAgICA6b3BubyA2MDcgCiAgICAgICAgICAgICAgICAgICAg ICAgICAgIDpvcGlkIDE4NCAKICAgICAgICAgICAgICAgICAgICAgICAgICAg Om9wcmVzdWx0dHlwZSAxNiAKICAgICAgICAgICAgICAgICAgICAgICAgICAg Om9wcmV0c2V0IGZhbHNlIAogICAgICAgICAgICAgICAgICAgICAgICAgICB9 CiAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAg ICAgICAgOmFyZ3MgKAogICAgICAgICAgICAgICAgICAgICAgICAgICB7IFZB UiAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZhcm5vIDY1MDAwIAog ICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFyYXR0bm8gMSAKICAgICAg ICAgICAgICAgICAgICAgICAgICAgOnZhcnR5cGUgMjYgCiAgICAgICAgICAg ICAgICAgICAgICAgICAgIDp2YXJ0eXBtb2QgLTEgIAogICAgICAgICAgICAg ICAgICAgICAgICAgICA6dmFybGV2ZWxzdXAgMCAKICAgICAgICAgICAgICAg ICAgICAgICAgICAgOnZhcm5vb2xkIDIgCiAgICAgICAgICAgICAgICAgICAg ICAgICAgIDp2YXJvYXR0bm8gMgogICAgICAgICAgICAgICAgICAgICAgICAg ICB9CiAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAg ICAgICAgICAgICAgeyBWQVIgCiAgICAgICAgICAgICAgICAgICAgICAgICAg IDp2YXJubyA2NTAwMSAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZh cmF0dG5vIDIgCiAgICAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJ0eXBl IDI2IAogICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFydHlwbW9kIC0x ICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZhcmxldmVsc3VwIDAg CiAgICAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJub29sZCAxIAogICAg ICAgICAgICAgICAgICAgICAgICAgICA6dmFyb2F0dG5vIC0yCiAgICAgICAg ICAgICAgICAgICAgICAgICAgIH0KICAgICAgICAgICAgICAgICAgICAgICAg KQogICAgICAgICAgICAgICAgICAgICAgICB9CiAgICAgICAgICAgICAgICAg ICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgeyBFWFBSIAogICAgICAg ICAgICAgICAgICAgICAgICA6dHlwZU9pZCAxNiAgCiAgICAgICAgICAgICAg ICAgICAgICAgIDpvcFR5cGUgb3AgCiAgICAgICAgICAgICAgICAgICAgICAg IDpvcGVyIAogICAgICAgICAgICAgICAgICAgICAgICAgICB7IE9QRVIgCiAg ICAgICAgICAgICAgICAgICAgICAgICAgIDpvcG5vIDkyIAogICAgICAgICAg ICAgICAgICAgICAgICAgICA6b3BpZCA2MSAKICAgICAgICAgICAgICAgICAg ICAgICAgICAgOm9wcmVzdWx0dHlwZSAxNiAKICAgICAgICAgICAgICAgICAg ICAgICAgICAgOm9wcmV0c2V0IGZhbHNlIAogICAgICAgICAgICAgICAgICAg ICAgICAgICB9CiAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAg ICAgICAgICAgICAgICAgOmFyZ3MgKAogICAgICAgICAgICAgICAgICAgICAg ICAgICB7IFZBUiAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZhcm5v IDY1MDAwIAogICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFyYXR0bm8g MiAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZhcnR5cGUgMTggCiAg ICAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJ0eXBtb2QgLTEgIAogICAg ICAgICAgICAgICAgICAgICAgICAgICA6dmFybGV2ZWxzdXAgMCAKICAgICAg ICAgICAgICAgICAgICAgICAgICAgOnZhcm5vb2xkIDIgCiAgICAgICAgICAg ICAgICAgICAgICAgICAgIDp2YXJvYXR0bm8gNwogICAgICAgICAgICAgICAg ICAgICAgICAgICB9CiAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAg ICAgICAgICAgICAgICAgICAgICAgeyBDT05TVCAKICAgICAgICAgICAgICAg ICAgICAgICAgICAgOmNvbnN0dHlwZSAxOCAKICAgICAgICAgICAgICAgICAg ICAgICAgICAgOmNvbnN0bGVuIDEgCiAgICAgICAgICAgICAgICAgICAgICAg ICAgIDpjb25zdGJ5dmFsIHRydWUgCiAgICAgICAgICAgICAgICAgICAgICAg ICAgIDpjb25zdGlzbnVsbCBmYWxzZSAKICAgICAgICAgICAgICAgICAgICAg ICAgICAgOmNvbnN0dmFsdWUgIDEgWyAwIDAgMCAxMDUgXSAKICAgICAgICAg ICAgICAgICAgICAgICAgICAgfQogICAgICAgICAgICAgICAgICAgICAgICAp CiAgICAgICAgICAgICAgICAgICAgICAgIH0KICAgICAgICAgICAgICAgICAg ICAgKQogICAgICAgICAgICAgICAgICAgICB9CiAgICAgICAgICAgICAgICAg ICAKICAgICAgICAgICAgICAgICAgOnJpZ2h0dHJlZSAKICAgICAgICAgICAg ICAgICAgICAgeyBIQVNIIAogICAgICAgICAgICAgICAgICAgICA6c3RhcnR1 cF9jb3N0IDAuMDAgCiAgICAgICAgICAgICAgICAgICAgIDp0b3RhbF9jb3N0 IDAuMDAgCiAgICAgICAgICAgICAgICAgICAgIDpyb3dzIDEgCiAgICAgICAg ICAgICAgICAgICAgIDp3aWR0aCA1IAogICAgICAgICAgICAgICAgICAgICA6 cXB0YXJnZXRsaXN0ICgKICAgICAgICAgICAgICAgICAgICAgICAgeyBUQVJH RVRFTlRSWSAKICAgICAgICAgICAgICAgICAgICAgICAgOnJlc2RvbSAKICAg ICAgICAgICAgICAgICAgICAgICAgICAgeyBSRVNET00gCiAgICAgICAgICAg ICAgICAgICAgICAgICAgIDpyZXNubyAxIAogICAgICAgICAgICAgICAgICAg ICAgICAgICA6cmVzdHlwZSAyNiAKICAgICAgICAgICAgICAgICAgICAgICAg ICAgOnJlc3R5cG1vZCAtMSAKICAgICAgICAgICAgICAgICAgICAgICAgICAg OnJlc25hbWUgPD4gCiAgICAgICAgICAgICAgICAgICAgICAgICAgIDpyZXNr ZXkgMCAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnJlc2tleW9wIDAg CiAgICAgICAgICAgICAgICAgICAgICAgICAgIDpyZXNzb3J0Z3JvdXByZWYg MCAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnJlc2p1bmsgZmFsc2Ug CiAgICAgICAgICAgICAgICAgICAgICAgICAgIH0KICAgICAgICAgICAgICAg ICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICA6ZXhwciAKICAg ICAgICAgICAgICAgICAgICAgICAgICAgeyBWQVIgCiAgICAgICAgICAgICAg ICAgICAgICAgICAgIDp2YXJubyA0IAogICAgICAgICAgICAgICAgICAgICAg ICAgICA6dmFyYXR0bm8gLTIgCiAgICAgICAgICAgICAgICAgICAgICAgICAg IDp2YXJ0eXBlIDI2IAogICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFy dHlwbW9kIC0xICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgOnZhcmxl dmVsc3VwIDAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJub29s ZCA0IAogICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFyb2F0dG5vIC0y CiAgICAgICAgICAgICAgICAgICAgICAgICAgIH0KICAgICAgICAgICAgICAg ICAgICAgICAgfQogICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAg ICAgICAgICAgICAgIHsgVEFSR0VURU5UUlkgCiAgICAgICAgICAgICAgICAg ICAgICAgIDpyZXNkb20gCiAgICAgICAgICAgICAgICAgICAgICAgICAgIHsg UkVTRE9NIAogICAgICAgICAgICAgICAgICAgICAgICAgICA6cmVzbm8gMiAK ICAgICAgICAgICAgICAgICAgICAgICAgICAgOnJlc3R5cGUgMTggCiAgICAg ICAgICAgICAgICAgICAgICAgICAgIDpyZXN0eXBtb2QgLTEgCiAgICAgICAg ICAgICAgICAgICAgICAgICAgIDpyZXNuYW1lIDw+IAogICAgICAgICAgICAg ICAgICAgICAgICAgICA6cmVza2V5IDAgCiAgICAgICAgICAgICAgICAgICAg ICAgICAgIDpyZXNrZXlvcCAwIAogICAgICAgICAgICAgICAgICAgICAgICAg ICA6cmVzc29ydGdyb3VwcmVmIDAgCiAgICAgICAgICAgICAgICAgICAgICAg ICAgIDpyZXNqdW5rIGZhbHNlIAogICAgICAgICAgICAgICAgICAgICAgICAg ICB9CiAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAg ICAgICAgICAgOmV4cHIgCiAgICAgICAgICAgICAgICAgICAgICAgICAgIHsg VkFSIAogICAgICAgICAgICAgICAgICAgICAgICAgICA6dmFybm8gNCAKICAg ICAgICAgICAgICAgICAgICAgICAgICAgOnZhcmF0dG5vIDMgCiAgICAgICAg ICAgICAgICAgICAgICAgICAgIDp2YXJ0eXBlIDE4IAogICAgICAgICAgICAg ICAgICAgICAgICAgICA6dmFydHlwbW9kIC0xICAKICAgICAgICAgICAgICAg ICAgICAgICAgICAgOnZhcmxldmVsc3VwIDAgCiAgICAgICAgICAgICAgICAg ICAgICAgICAgIDp2YXJub29sZCA0IAogICAgICAgICAgICAgICAgICAgICAg ICAgICA6dmFyb2F0dG5vIDMKICAgICAgICAgICAgICAgICAgICAgICAgICAg fQogICAgICAgICAgICAgICAgICAgICAgICB9CiAgICAgICAgICAgICAgICAg ICAgICkKICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAg ICAgICA6cXBxdWFsIDw+IAogICAgICAgICAgICAgICAgICAgICA6bGVmdHRy ZWUgCiAgICAgICAgICAgICAgICAgICAgICAgIHsgU0VRU0NBTiAKICAgICAg ICAgICAgICAgICAgICAgICAgOnN0YXJ0dXBfY29zdCAwLjAwIAogICAgICAg ICAgICAgICAgICAgICAgICA6dG90YWxfY29zdCAwLjAwIAogICAgICAgICAg ICAgICAgICAgICAgICA6cm93cyAxIAogICAgICAgICAgICAgICAgICAgICAg ICA6d2lkdGggNSAKICAgICAgICAgICAgICAgICAgICAgICAgOnFwdGFyZ2V0 bGlzdCAoCiAgICAgICAgICAgICAgICAgICAgICAgICAgIHsgVEFSR0VURU5U UlkgCiAgICAgICAgICAgICAgICAgICAgICAgICAgIDpyZXNkb20gCiAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgIHsgUkVTRE9NIAogICAgICAgICAg ICAgICAgICAgICAgICAgICAgICA6cmVzbm8gMSAKICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgOnJlc3R5cGUgMjYgCiAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgIDpyZXN0eXBtb2QgLTEgCiAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgIDpyZXNuYW1lIDw+IAogICAgICAgICAgICAgICAgICAg ICAgICAgICAgICA6cmVza2V5IDAgCiAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgIDpyZXNrZXlvcCAwIAogICAgICAgICAgICAgICAgICAgICAgICAg ICAgICA6cmVzc29ydGdyb3VwcmVmIDAgCiAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgIDpyZXNqdW5rIGZhbHNlIAogICAgICAgICAgICAgICAgICAg ICAgICAgICAgICB9CiAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAg ICAgICAgICAgICAgICAgICAgICAgICAgOmV4cHIgCiAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgIHsgVkFSIAogICAgICAgICAgICAgICAgICAgICAg ICAgICAgICA6dmFybm8gNCAKICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgOnZhcmF0dG5vIC0yIAogICAgICAgICAgICAgICAgICAgICAgICAgICAg ICA6dmFydHlwZSAyNiAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg OnZhcnR5cG1vZCAtMSAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAg IDp2YXJsZXZlbHN1cCAwIAogICAgICAgICAgICAgICAgICAgICAgICAgICAg ICA6dmFybm9vbGQgNCAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg OnZhcm9hdHRubyAtMgogICAgICAgICAgICAgICAgICAgICAgICAgICAgICB9 CiAgICAgICAgICAgICAgICAgICAgICAgICAgIH0KICAgICAgICAgICAgICAg ICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICB7IFRBUkdF VEVOVFJZIAogICAgICAgICAgICAgICAgICAgICAgICAgICA6cmVzZG9tIAog ICAgICAgICAgICAgICAgICAgICAgICAgICAgICB7IFJFU0RPTSAKICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgOnJlc25vIDIgCiAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgIDpyZXN0eXBlIDE4IAogICAgICAgICAgICAg ICAgICAgICAgICAgICAgICA6cmVzdHlwbW9kIC0xIAogICAgICAgICAgICAg ICAgICAgICAgICAgICAgICA6cmVzbmFtZSA8PiAKICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgOnJlc2tleSAwIAogICAgICAgICAgICAgICAgICAg ICAgICAgICAgICA6cmVza2V5b3AgMCAKICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgOnJlc3NvcnRncm91cHJlZiAwIAogICAgICAgICAgICAgICAg ICAgICAgICAgICAgICA6cmVzanVuayBmYWxzZSAKICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgfQogICAgICAgICAgICAgICAgICAgICAgICAgICAg CiAgICAgICAgICAgICAgICAgICAgICAgICAgIDpleHByIAogICAgICAgICAg ICAgICAgICAgICAgICAgICAgICB7IFZBUiAKICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgOnZhcm5vIDQgCiAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgIDp2YXJhdHRubyAzIAogICAgICAgICAgICAgICAgICAgICAgICAg ICAgICA6dmFydHlwZSAxOCAKICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgOnZhcnR5cG1vZCAtMSAgCiAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgIDp2YXJsZXZlbHN1cCAwIAogICAgICAgICAgICAgICAgICAgICAgICAg ICAgICA6dmFybm9vbGQgNCAKICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgOnZhcm9hdHRubyAzCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAg IH0KICAgICAgICAgICAgICAgICAgICAgICAgICAgfQogICAgICAgICAgICAg ICAgICAgICAgICApCiAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAg ICAgICAgICAgICAgICAgICAgOnFwcXVhbCA8PiAKICAgICAgICAgICAgICAg ICAgICAgICAgOmxlZnR0cmVlIDw+IAogICAgICAgICAgICAgICAgICAgICAg ICA6cmlnaHR0cmVlIDw+IAogICAgICAgICAgICAgICAgICAgICAgICA6ZXh0 cHJtICgpCiAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAg ICAgICAgICAgICAgOmxvY3BybSAoKQogICAgICAgICAgICAgICAgICAgICAg ICAgCiAgICAgICAgICAgICAgICAgICAgICAgIDppbml0cGxhbiA8PiAKICAg ICAgICAgICAgICAgICAgICAgICAgOm5wcm0gMCAgCiAgICAgICAgICAgICAg ICAgICAgICAgIDpzY2FucmVsaWQgNCAKICAgICAgICAgICAgICAgICAgICAg ICAgfQogICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAg ICAgIDpyaWdodHRyZWUgPD4gCiAgICAgICAgICAgICAgICAgICAgIDpleHRw cm0gKCkKICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAg ICAgICA6bG9jcHJtICgpCiAgICAgICAgICAgICAgICAgICAgICAKICAgICAg ICAgICAgICAgICAgICAgOmluaXRwbGFuIDw+IAogICAgICAgICAgICAgICAg ICAgICA6bnBybSAwICAKICAgICAgICAgICAgICAgICAgICAgOmhhc2hrZXkg CiAgICAgICAgICAgICAgICAgICAgICAgIHsgVkFSIAogICAgICAgICAgICAg ICAgICAgICAgICA6dmFybm8gNjUwMDAgCiAgICAgICAgICAgICAgICAgICAg ICAgIDp2YXJhdHRubyAxIAogICAgICAgICAgICAgICAgICAgICAgICA6dmFy dHlwZSAyNiAKICAgICAgICAgICAgICAgICAgICAgICAgOnZhcnR5cG1vZCAt MSAgCiAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJsZXZlbHN1cCAwIAog ICAgICAgICAgICAgICAgICAgICAgICA6dmFybm9vbGQgNCAKICAgICAgICAg ICAgICAgICAgICAgICAgOnZhcm9hdHRubyAtMgogICAgICAgICAgICAgICAg ICAgICAgICB9CiAgICAgICAgICAgICAgICAgICAgIH0KICAgICAgICAgICAg ICAgICAgIAogICAgICAgICAgICAgICAgICA6ZXh0cHJtICgpCiAgICAgICAg ICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgOmxvY3BybSAoKQogICAg ICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgIDppbml0cGxhbiA8 PiAKICAgICAgICAgICAgICAgICAgOm5wcm0gMCAgCiAgICAgICAgICAgICAg ICAgIDpqb2ludHlwZSAxIAogICAgICAgICAgICAgICAgICA6am9pbnF1YWwg KAogICAgICAgICAgICAgICAgICAgICB7IEVYUFIgCiAgICAgICAgICAgICAg ICAgICAgIDp0eXBlT2lkIDE2ICAKICAgICAgICAgICAgICAgICAgICAgOm9w VHlwZSBvcCAKICAgICAgICAgICAgICAgICAgICAgOm9wZXIgCiAgICAgICAg ICAgICAgICAgICAgICAgIHsgT1BFUiAKICAgICAgICAgICAgICAgICAgICAg ICAgOm9wbm8gOTIgCiAgICAgICAgICAgICAgICAgICAgICAgIDpvcGlkIDYx IAogICAgICAgICAgICAgICAgICAgICAgICA6b3ByZXN1bHR0eXBlIDE2IAog ICAgICAgICAgICAgICAgICAgICAgICA6b3ByZXRzZXQgZmFsc2UgCiAgICAg ICAgICAgICAgICAgICAgICAgIH0KICAgICAgICAgICAgICAgICAgICAgIAog ICAgICAgICAgICAgICAgICAgICA6YXJncyAoCiAgICAgICAgICAgICAgICAg ICAgICAgIHsgVkFSIAogICAgICAgICAgICAgICAgICAgICAgICA6dmFybm8g NjUwMDAgCiAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJhdHRubyAyIAog ICAgICAgICAgICAgICAgICAgICAgICA6dmFydHlwZSAxOCAKICAgICAgICAg ICAgICAgICAgICAgICAgOnZhcnR5cG1vZCAtMSAgCiAgICAgICAgICAgICAg ICAgICAgICAgIDp2YXJsZXZlbHN1cCAwIAogICAgICAgICAgICAgICAgICAg ICAgICA6dmFybm9vbGQgNCAKICAgICAgICAgICAgICAgICAgICAgICAgOnZh cm9hdHRubyAzCiAgICAgICAgICAgICAgICAgICAgICAgIH0KICAgICAgICAg ICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICB7IENPTlNU IAogICAgICAgICAgICAgICAgICAgICAgICA6Y29uc3R0eXBlIDE4IAogICAg ICAgICAgICAgICAgICAgICAgICA6Y29uc3RsZW4gMSAKICAgICAgICAgICAg ICAgICAgICAgICAgOmNvbnN0Ynl2YWwgdHJ1ZSAKICAgICAgICAgICAgICAg ICAgICAgICAgOmNvbnN0aXNudWxsIGZhbHNlIAogICAgICAgICAgICAgICAg ICAgICAgICA6Y29uc3R2YWx1ZSAgMSBbIDAgMCAwIDEwMiBdIAogICAgICAg ICAgICAgICAgICAgICAgICB9CiAgICAgICAgICAgICAgICAgICAgICkKICAg ICAgICAgICAgICAgICAgICAgfQogICAgICAgICAgICAgICAgICApCiAgICAg ICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgOmhhc2hjbGF1c2Vz ICgKICAgICAgICAgICAgICAgICAgICAgeyBFWFBSIAogICAgICAgICAgICAg ICAgICAgICA6dHlwZU9pZCAxNiAgCiAgICAgICAgICAgICAgICAgICAgIDpv cFR5cGUgb3AgCiAgICAgICAgICAgICAgICAgICAgIDpvcGVyIAogICAgICAg ICAgICAgICAgICAgICAgICB7IE9QRVIgCiAgICAgICAgICAgICAgICAgICAg ICAgIDpvcG5vIDYwNyAKICAgICAgICAgICAgICAgICAgICAgICAgOm9waWQg MTg0IAogICAgICAgICAgICAgICAgICAgICAgICA6b3ByZXN1bHR0eXBlIDE2 IAogICAgICAgICAgICAgICAgICAgICAgICA6b3ByZXRzZXQgZmFsc2UgCiAg ICAgICAgICAgICAgICAgICAgICAgIH0KICAgICAgICAgICAgICAgICAgICAg IAogICAgICAgICAgICAgICAgICAgICA6YXJncyAoCiAgICAgICAgICAgICAg ICAgICAgICAgIHsgVkFSIAogICAgICAgICAgICAgICAgICAgICAgICA6dmFy bm8gNjUwMDEgCiAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJhdHRubyA1 IAogICAgICAgICAgICAgICAgICAgICAgICA6dmFydHlwZSAyNiAKICAgICAg ICAgICAgICAgICAgICAgICAgOnZhcnR5cG1vZCAtMSAgCiAgICAgICAgICAg ICAgICAgICAgICAgIDp2YXJsZXZlbHN1cCAwIAogICAgICAgICAgICAgICAg ICAgICAgICA6dmFybm9vbGQgMiAKICAgICAgICAgICAgICAgICAgICAgICAg OnZhcm9hdHRubyA1CiAgICAgICAgICAgICAgICAgICAgICAgIH0KICAgICAg ICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICB7IFZB UiAKICAgICAgICAgICAgICAgICAgICAgICAgOnZhcm5vIDY1MDAwIAogICAg ICAgICAgICAgICAgICAgICAgICA6dmFyYXR0bm8gMSAKICAgICAgICAgICAg ICAgICAgICAgICAgOnZhcnR5cGUgMjYgCiAgICAgICAgICAgICAgICAgICAg ICAgIDp2YXJ0eXBtb2QgLTEgIAogICAgICAgICAgICAgICAgICAgICAgICA6 dmFybGV2ZWxzdXAgMCAKICAgICAgICAgICAgICAgICAgICAgICAgOnZhcm5v b2xkIDQgCiAgICAgICAgICAgICAgICAgICAgICAgIDp2YXJvYXR0bm8gLTIK ICAgICAgICAgICAgICAgICAgICAgICAgfQogICAgICAgICAgICAgICAgICAg ICApCiAgICAgICAgICAgICAgICAgICAgIH0KICAgICAgICAgICAgICAgICAg KQogICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgIDpoYXNo am9pbm9wIDAgCiAgICAgICAgICAgICAgICAgIH0KICAgICAgICAgICAgICAg IAogICAgICAgICAgICAgICA6cmlnaHR0cmVlIDw+IAogICAgICAgICAgICAg ICA6ZXh0cHJtICgpCiAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAg OmxvY3BybSAoKQogICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgIDpp bml0cGxhbiA8PiAKICAgICAgICAgICAgICAgOm5wcm0gMCAgCiAgICAgICAg ICAgICAgIDprZXljb3VudCAxIAogICAgICAgICAgICAgICB9CiAgICAgICAg ICAgICAKICAgICAgICAgICAgOnJpZ2h0dHJlZSA8PiAKICAgICAgICAgICAg OmV4dHBybSAoKQogICAgICAgICAgICAgCiAgICAgICAgICAgIDpsb2Nwcm0g KCkKICAgICAgICAgICAgIAogICAgICAgICAgICA6aW5pdHBsYW4gPD4gCiAg ICAgICAgICAgIDpucHJtIDAgIAogICAgICAgICAgICA6bnVtQ29scyAxIAog ICAgICAgICAgICA6dHVwbGVQZXJHcm91cCB0cnVlIAogICAgICAgICAgICB9 CiAgICAgICAgICAKICAgICAgICAgOnJpZ2h0dHJlZSA8PiAKICAgICAgICAg OmV4dHBybSAoKQogICAgICAgICAgCiAgICAgICAgIDpsb2Nwcm0gKCkKICAg ICAgICAgIAogICAgICAgICA6aW5pdHBsYW4gPD4gCiAgICAgICAgIDpucHJt IDAgCiAgICAgICAgIH0KICAgICAgfQogICAgCiAgIDpyaWdodHRyZWUgPD4g CiAgIDpleHRwcm0gKCkKICAgIAogICA6bG9jcHJtICgpCiAgICAKICAgOmlu aXRwbGFuIDw+IAogICA6bnBybSAwIAogICB9CgpBZ2dyZWdhdGUgIChjb3N0 PTE4My4zNi4uMTgzLjM2IHJvd3M9MSB3aWR0aD0yMikgKGFjdHVhbCB0aW1l PTQzNTg5LjczLi40MzU4OS43MyByb3dzPTEgbG9vcHM9MSkKICAtPiAgU3Vi cXVlcnkgU2NhbiBmb28gIChjb3N0PTE4My4zNC4uMTgzLjM2IHJvd3M9MSB3 aWR0aD0yMikgKGFjdHVhbCB0aW1lPTQzNTg5LjcxLi40MzU4OS43MSByb3dz PTAgbG9vcHM9MSkKICAgICAgICAtPiAgQWdncmVnYXRlICAoY29zdD0xODMu MzQuLjE4My4zNiByb3dzPTEgd2lkdGg9MjIpIChhY3R1YWwgdGltZT00MzU4 OS43MC4uNDM1ODkuNzAgcm93cz0wIGxvb3BzPTEpCiAgICAgICAgICAgICAg RmlsdGVyOiAoY291bnQoMSkgPSAzKQogICAgICAgICAgICAgIC0+ICBHcm91 cCAgKGNvc3Q9MTgzLjM0Li4xODMuMzUgcm93cz0yIHdpZHRoPTIyKSAoYWN0 dWFsIHRpbWU9NDM1ODkuMzIuLjQzNTg5LjUzIHJvd3M9MTUgbG9vcHM9MSkK ICAgICAgICAgICAgICAgICAgICAtPiAgU29ydCAgKGNvc3Q9MTgzLjM0Li4x ODMuMzUgcm93cz0yIHdpZHRoPTIyKSAoYWN0dWFsIHRpbWU9NDM1ODkuMzEu LjQzNTg5LjMyIHJvd3M9MTUgbG9vcHM9MSkKICAgICAgICAgICAgICAgICAg ICAgICAgICBTb3J0IEtleTogdHIudGdhcmdzCiAgICAgICAgICAgICAgICAg ICAgICAgICAgLT4gIEhhc2ggSm9pbiAgKGNvc3Q9MC4wMC4uMTgzLjMzIHJv d3M9MiB3aWR0aD0yMikgKGFjdHVhbCB0aW1lPTE0My41My4uNDM1ODguMjIg cm93cz0xNSBsb29wcz0xKQogICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgIEhhc2ggQ29uZDogKCJvdXRlciIucmVmb2JqaWQgPSAiaW5uZXIiLm9p ZCkKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICBKb2luIEZpbHRl cjogKCJpbm5lciIuY29udHlwZSA9ICdmJzo6ImNoYXIiKQogICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgIEZpbHRlcjogKCJpbm5lciIub2lkIElT IE5VTEwpCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgLT4gIE5l c3RlZCBMb29wICAoY29zdD0wLjAwLi4xODMuMzIgcm93cz0yIHdpZHRoPTE3 KSAoYWN0dWFsIHRpbWU9MTM4LjE3Li40MzU3My4yNCByb3dzPTMyMSBsb29w cz0xKQogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIEpv aW4gRmlsdGVyOiAoKCJpbm5lciIub2JqaWQgPSAib3V0ZXIiLm9pZCkgQU5E ICgiaW5uZXIiLmRlcHR5cGUgPSAnaSc6OiJjaGFyIikpCiAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgLT4gIFNlcSBTY2FuIG9uIHBn X3RyaWdnZXIgdHIgIChjb3N0PTAuMDAuLjEuMDIgcm93cz0yIHdpZHRoPTgp IChhY3R1YWwgdGltZT0wLjExLi4xMC4yMSByb3dzPTMyMSBsb29wcz0xKQog ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIC0+ICBTZXEg U2NhbiBvbiBwZ19kZXBlbmQgZGVwICAoY29zdD0wLjAwLi40OS4wNiByb3dz PTI4MDYgd2lkdGg9OSkgKGFjdHVhbCB0aW1lPTAuMDMuLjg3Ljc4IHJvd3M9 NDU3MCBsb29wcz0zMjEpCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgLT4gIEhhc2ggIChjb3N0PTAuMDAuLjAuMDAgcm93cz0xIHdpZHRoPTUp IChhY3R1YWwgdGltZT00Ljg5Li40Ljg5IHJvd3M9MCBsb29wcz0xKQogICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIC0+ICBTZXEgU2Nh biBvbiBwZ19jb25zdHJhaW50IGNvICAoY29zdD0wLjAwLi4wLjAwIHJvd3M9 MSB3aWR0aD01KSAoYWN0dWFsIHRpbWU9MC4xMy4uMy40MSByb3dzPTE4MyBs b29wcz0xKQpUb3RhbCBydW50aW1lOiA0MzU5My42MiBtc2VjCg== --=====================_15580593==_-- From pgsql-performance-owner@postgresql.org Tue Jan 6 17:30:22 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 34D36D1B4AC for ; Tue, 6 Jan 2004 21:30:21 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 81950-05 for ; Tue, 6 Jan 2004 17:29:34 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id CB2EBD1C9CE for ; Tue, 6 Jan 2004 17:25:07 -0400 (AST) Received: from [63.195.55.98] (HELO temoku) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 4212837; Tue, 06 Jan 2004 13:26:08 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Reply-To: josh@agliodbs.com Organization: Aglio Database Solutions To: Michael Shapiro , pgsql-performance@postgresql.org Subject: Re: PgAdmin startup query VERY slow Date: Tue, 6 Jan 2004 13:01:43 -0800 User-Agent: KMail/1.4.3 References: <5.2.0.9.2.20040106144745.00bcfe30@pop.ncsa.uiuc.edu> In-Reply-To: <5.2.0.9.2.20040106144745.00bcfe30@pop.ncsa.uiuc.edu> MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Message-Id: <200401061301.43954.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/73 X-Sequence-Number: 5248 Michael, > With a particular database, PgAdmin3 takes a very long time to connect to= a=20 > database. this is not a general problem with PgAdmin, but only with one= =20 > database out of many. Other databases do not have the problem. And only= =20 > with one particular server. The exact same database on a different server= =20 > does not have the problem. Have you run VACUUM ANALYZE *as the superuser* on the faulty server recentl= y?=20=20 >From the look of the explain, PG is grossly underestimating the number of= =20 items in the pg_trigger and pg_depend tables, and thus choosing an=20 inappropriate nested loop execution. --=20 -Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Tue Jan 6 17:42:13 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 55CAED1B433 for ; Tue, 6 Jan 2004 21:42:12 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 82664-08 for ; Tue, 6 Jan 2004 17:41:25 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id D80BBD1B4D4 for ; Tue, 6 Jan 2004 17:40:54 -0400 (AST) Received: from [63.195.55.98] (HELO temoku) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 4212926; Tue, 06 Jan 2004 13:41:55 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Reply-To: josh@agliodbs.com Organization: Aglio Database Solutions To: Michael Shapiro , pgsql-performance@postgresql.org Subject: Re: PgAdmin startup query VERY slow Date: Tue, 6 Jan 2004 13:17:30 -0800 User-Agent: KMail/1.4.3 References: <5.2.0.9.2.20040106144745.00bcfe30@pop.ncsa.uiuc.edu> <5.2.0.9.2.20040106152859.02a17ea8@pop.ncsa.uiuc.edu> In-Reply-To: <5.2.0.9.2.20040106152859.02a17ea8@pop.ncsa.uiuc.edu> MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Message-Id: <200401061317.30877.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/75 X-Sequence-Number: 5250 Mark, > That seemed to fix it. What does VACUUM ANALYZE do that VACUUM FULL does= =20 > not? What causes a database to need vacuuming? See the Online Docs: http://www.postgresql.org/docs/current/static/maintenance.html Incidentally, just ANALYZE would probably have fixed your problem. Please= do=20 suggest to the PGAdmin team that they add a FAQ item about this. --=20 -Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Tue Jan 6 17:33:21 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 8B229D1B465 for ; Tue, 6 Jan 2004 21:33:19 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 81670-08 for ; Tue, 6 Jan 2004 17:32:31 -0400 (AST) Received: from mail.ncsa.uiuc.edu (mail.ncsa.uiuc.edu [141.142.2.28]) by svr1.postgresql.org (Postfix) with ESMTP id 83C26D1C957 for ; Tue, 6 Jan 2004 17:31:28 -0400 (AST) X-Envelope-From: mshapiro@ncsa.edu X-Envelope-To: Received: from kalika.ncsa.edu (kalika.ncsa.uiuc.edu [141.142.97.63]) by mail.ncsa.uiuc.edu (8.11.6p2/8.11.6) with ESMTP id i06LVUC29113 for ; Tue, 6 Jan 2004 15:31:30 -0600 Message-Id: <5.2.0.9.2.20040106152859.02a17ea8@pop.ncsa.uiuc.edu> X-Sender: mshapiro@pop.ncsa.uiuc.edu (Unverified) X-Mailer: QUALCOMM Windows Eudora Version 5.2.0.9 Date: Tue, 06 Jan 2004 15:31:29 -0600 To: pgsql-performance@postgresql.org From: Michael Shapiro Subject: Re: PgAdmin startup query VERY slow In-Reply-To: <200401061301.43954.josh@agliodbs.com> References: <5.2.0.9.2.20040106144745.00bcfe30@pop.ncsa.uiuc.edu> <5.2.0.9.2.20040106144745.00bcfe30@pop.ncsa.uiuc.edu> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/74 X-Sequence-Number: 5249 That seemed to fix it. What does VACUUM ANALYZE do that VACUUM FULL does not? What causes a database to need vacuuming? At 01:01 PM 1/6/2004 -0800, Josh Berkus wrote: >Michael, > > > With a particular database, PgAdmin3 takes a very long time to connect > to a > > database. this is not a general problem with PgAdmin, but only with one > > database out of many. Other databases do not have the problem. And only > > with one particular server. The exact same database on a different server > > does not have the problem. > >Have you run VACUUM ANALYZE *as the superuser* on the faulty server >recently? > >From the look of the explain, PG is grossly underestimating the number of >items in the pg_trigger and pg_depend tables, and thus choosing an >inappropriate nested loop execution. > >-- >-Josh Berkus > Aglio Database Solutions > San Francisco --- Michael From pgsql-performance-owner@postgresql.org Tue Jan 6 18:57:51 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 88D22D1B46E for ; Tue, 6 Jan 2004 22:57:49 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 93225-06 for ; Tue, 6 Jan 2004 18:57:01 -0400 (AST) Received: from druid.net (druid.net [216.126.72.98]) by svr1.postgresql.org (Postfix) with ESMTP id 142B1D1B433 for ; Tue, 6 Jan 2004 18:56:59 -0400 (AST) Received: by druid.net (Postfix, from userid 1000) id 661BE1A92; Tue, 6 Jan 2004 17:57:07 -0500 (EST) From: "D'Arcy J.M. Cain" To: Shridhar Daithankar , Postgresql Performance Subject: Re: Select max(foo) and select count(*) optimization Date: Tue, 6 Jan 2004 17:57:05 -0500 User-Agent: KMail/1.5.4 References: <200401060718.08444.darcy@druid.net> <200401061750.09530.shridhar_daithankar@persistent.co.in> In-Reply-To: <200401061750.09530.shridhar_daithankar@persistent.co.in> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200401061757.05278.darcy@druid.net> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/76 X-Sequence-Number: 5251 On January 6, 2004 07:20 am, Shridhar Daithankar wrote: > On Tuesday 06 January 2004 17:48, D'Arcy J.M. Cain wrote: > > On January 6, 2004 01:42 am, Shridhar Daithankar wrote: > > cert=# select relpages,reltuples::bigint from pg_class where relname= > > 'certificate'; > > relpages | reltuples > > ----------+----------- > > 399070 | 24858736 > > (1 row) > > > > But: > > > > cert=# select count(*) from certificate; > > [*Crunch* *Crunch* *Crunch*] > > count > > ---------- > > 19684668 > > (1 row) > > > > Am I missing something? Max certificate_id is 20569544 btw. > > Do 'vacuum analyze certificate' and try..:-) Kind of invalidates the part about being accurate then, don't it? Besides, I vacuum that table every day (*) and we have reorganized the schema so that we never update it except in exceptional cases. I would be less surprised if the result was less than the real count since we only insert into that table. In any case, if I have to vacuum a 20,000,000 row table to get an accurate count then I may as well run count(*) on it. (*): Actually I only analyze but I understand that that should be sufficient. -- D'Arcy J.M. Cain | Democracy is three wolves http://www.druid.net/darcy/ | and a sheep voting on +1 416 425 1212 (DoD#0082) (eNTP) | what's for dinner. From pgsql-performance-owner@postgresql.org Tue Jan 6 19:13:28 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5ED51D1B49B for ; Tue, 6 Jan 2004 23:13:27 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 92307-10 for ; Tue, 6 Jan 2004 19:12:39 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id A502DD1B47D for ; Tue, 6 Jan 2004 19:12:36 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i06NCZ19001688; Tue, 6 Jan 2004 18:12:35 -0500 (EST) To: josh@agliodbs.com Cc: Michael Shapiro , pgsql-performance@postgresql.org Subject: Re: PgAdmin startup query VERY slow In-reply-to: <200401061317.30877.josh@agliodbs.com> References: <5.2.0.9.2.20040106144745.00bcfe30@pop.ncsa.uiuc.edu> <5.2.0.9.2.20040106152859.02a17ea8@pop.ncsa.uiuc.edu> <200401061317.30877.josh@agliodbs.com> Comments: In-reply-to Josh Berkus message dated "Tue, 06 Jan 2004 13:17:30 -0800" Date: Tue, 06 Jan 2004 18:12:35 -0500 Message-ID: <1687.1073430755@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/77 X-Sequence-Number: 5252 Josh Berkus writes: > Incidentally, just ANALYZE would probably have fixed your problem. ... or just VACUUM; that would have updated the row count which is all that was really needed here. The main point is that you do have to do that as superuser, since the same commands issued as a non-superuser won't touch the system tables (or any table you do not own). regards, tom lane From pgsql-performance-owner@postgresql.org Tue Jan 6 19:20:50 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id B8113D1B4C2 for ; Tue, 6 Jan 2004 23:20:49 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 02715-03 for ; Tue, 6 Jan 2004 19:20:03 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 56807D1B438 for ; Tue, 6 Jan 2004 19:20:00 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i06NJt19002699; Tue, 6 Jan 2004 18:19:56 -0500 (EST) To: "D'Arcy J.M. Cain" Cc: Shridhar Daithankar , Postgresql Performance Subject: Re: Select max(foo) and select count(*) optimization In-reply-to: <200401061757.05278.darcy@druid.net> References: <200401060718.08444.darcy@druid.net> <200401061750.09530.shridhar_daithankar@persistent.co.in> <200401061757.05278.darcy@druid.net> Comments: In-reply-to "D'Arcy J.M. Cain" message dated "Tue, 06 Jan 2004 17:57:05 -0500" Date: Tue, 06 Jan 2004 18:19:55 -0500 Message-ID: <2698.1073431195@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/78 X-Sequence-Number: 5253 "D'Arcy J.M. Cain" writes: > In any case, if I have to vacuum a 20,000,000 row table to get an accurate > count then I may as well run count(*) on it. > (*): Actually I only analyze but I understand that that should be sufficient. ANALYZE without VACUUM will deliver a not-very-accurate estimate, since it only looks at a sample of the table's pages and doesn't grovel through every one. Any of the VACUUM variants, on the other hand, will set pg_class.reltuples reasonably accurately (as the number of rows actually seen and left undeleted by the VACUUM pass). There are pathological cases where ANALYZE's estimate of the overall row count can be horribly bad --- mainly, when the early pages of the table are empty or nearly so, but there are well-filled pages out near the end. I have a TODO item to try to make ANALYZE less prone to getting fooled that way... regards, tom lane From pgsql-general-owner@postgresql.org Wed Jan 7 13:18:26 2004 X-Original-To: pgsql-general-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5EB3CD1E1C7; Wed, 7 Jan 2004 17:18:22 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 45900-09; Wed, 7 Jan 2004 13:17:39 -0400 (AST) Received: from rwcrmhc11.comcast.net (rwcrmhc11.comcast.net [204.127.198.35]) by svr1.postgresql.org (Postfix) with ESMTP id C5A3FD1E1FE; Wed, 7 Jan 2004 13:15:03 -0400 (AST) Received: from lorenso.com (c-24-1-26-144.client.comcast.net[24.1.26.144]) by comcast.net (rwcrmhc11) with ESMTP id <2004010716573001300797rke>; Wed, 7 Jan 2004 16:57:30 +0000 Message-ID: <3FFC3A82.4000401@lorenso.com> Date: Wed, 07 Jan 2004 10:57:38 -0600 From: "D. Dante Lorenso" User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.6b) Gecko/20031205 Thunderbird/0.4 X-Accept-Language: en-us, en MIME-Version: 1.0 To: "pgsql-general@postgresql.org" , pgsql-performance@postgresql.org Subject: Find original number of rows before applied LIMIT/OFFSET? Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/268 X-Sequence-Number: 55296 I need to know that original number of rows that WOULD have been returned by a SELECT statement if the LIMIT / OFFSET where not present in the statement. Is there a way to get this data from PG ? SELECT ... ; ----> returns 100,000 rows but, SELECT ... LIMIT x OFFSET y; ----> returns at most x rows In order to build a list pager on a web site, I want to select 'pages' of a result set at a time. However, I need to know the original select result set size because I still have to draw the 'page numbers' to display what pages are available. I've done this TWO ways in the past: 1) TWO queries. The first query will perform a SELECT COUNT(*) ...; and the second query performs the actualy SELECT ... LIMIT x OFFSET y; 2) Using PHP row seek and only selecting the number of rows I need. Here is an example of method number 2 in PHP: //---------------------------------------------------------------------- function query_assoc_paged ($sql, $limit=0, $offset=0) { $this->num_rows = false; // open a result set for this query... $result = $this->query($sql); if (! $result) return (false); // save the number of rows we are working with $this->num_rows = @pg_num_rows($result); // moves the internal row pointer of the result to point to our // desired offset. The next call to pg_fetch_assoc() would return // that row. if (! empty($offset)) { if (! @pg_result_seek($result, $offset)) { return (array()); }; } // gather the results together in an array of arrays... $data = array(); while (($row = pg_fetch_assoc($result)) !== false) { $data[] = $row; // After reading N rows from this result set, free our memory // and return the rows we fetched... if (! empty($limit) && count($data) >= $limit) { pg_free_result($result); return ($data); } } pg_free_result($result); return($data); } //---------------------------------------------------------------------- In this approach, I am 'emulating' the LIMIT / OFFSET features in PostgreSQL by just seeking forward in the result set (offset) and only fetching the number of rows that match my needs (LIMIT). QUESTION: Is this the best way to do this, or is there a more efficient way to get at the data I want? Is there a variable set in PG that tells me the original number of rows in the query? Something like: SELECT ORIG_RESULT_SIZE, ... ... LIMIT x OFFSET y; Or can I run another select right afterwards...like: SELECT ... ... LIMIT x OFFSET y; SELECT unfiltered_size_of_last_query(); Any thoughts? Sure, the PHP function I'm using above 'works', but is it the most efficient? I hope I'm not actually pulling all 100,000 records across the wire when I only intend to show 10 at a time. See what I'm getting at? TIA, Dante --------- D. Dante Lorenso dante@lorenso.com From pgsql-performance-owner@postgresql.org Wed Jan 7 13:15:38 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1BAC8D1B491 for ; Wed, 7 Jan 2004 17:15:34 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 45804-06 for ; Wed, 7 Jan 2004 13:14:51 -0400 (AST) Received: from expasy-ng.isb-sib.ch (expasy-f.unige.ch [192.33.215.142]) by svr1.postgresql.org (Postfix) with ESMTP id 58975D1B4B3 for ; Wed, 7 Jan 2004 13:12:28 -0400 (AST) Received: from caliente (router.isb-sib.ch [192.33.215.254]) (authenticated bits=0) by expasy-ng.isb-sib.ch (8.12.10/8.12.8) with ESMTP id i07H80is030931 for ; Wed, 7 Jan 2004 18:08:04 +0100 Message-ID: <001101c3d540$d3ea8090$c300000a@caliente> From: "Eric Jain" To: "pgsql-performance" Subject: Index creation Date: Wed, 7 Jan 2004 18:08:06 +0100 Organization: Swiss Institute of Bioinformatics MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1158 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 X-sib-MailScanner-Information: Please contact the ISP for more information X-sib-MailScanner: Found to be clean X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=0.0 tagged_above=0.0 required=5.0 tests= X-Spam-Level: X-Archive-Number: 200401/79 X-Sequence-Number: 5254 Any tips for speeding up index creation? I need to bulk load a large table with 100M rows and several indexes, some of which span two columns. By dropping all indexes prior to issuing the 'copy from' command, the operation completes 10x as fast (1.5h vs 15h). Unfortunately, recreating a single index takes nearly as long as loading all of the data into the table; this more or less eliminates the time gained by dropping the index in the first place. Also, there doesn't seem to be a simple way to disable/recreate all indexes for a specific table short of explicitely dropping and later recreating each index? -- Eric Jain From pgsql-performance-owner@postgresql.org Wed Jan 7 13:27:57 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 24A76D1E20A for ; Wed, 7 Jan 2004 17:21:21 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 48514-02 for ; Wed, 7 Jan 2004 13:20:38 -0400 (AST) Received: from jefftrout.com (h00a0cc4084e5.ne.client2.attbi.com [24.128.241.68]) by svr1.postgresql.org (Postfix) with SMTP id 51AFCD1E1EA for ; Wed, 7 Jan 2004 13:19:42 -0400 (AST) Received: (qmail 42047 invoked from network); 7 Jan 2004 17:19:51 -0000 Received: from localhost (HELO squeegit) (threshar@127.0.0.1) by localhost with SMTP; 7 Jan 2004 17:19:51 -0000 Date: Wed, 7 Jan 2004 12:20:15 -0500 From: Jeff To: "Eric Jain" Cc: pgsql-performance@postgresql.org Subject: Re: Index creation Message-Id: <20040107122015.1466a8c4.threshar@torgo.978.org> In-Reply-To: <001101c3d540$d3ea8090$c300000a@caliente> References: <001101c3d540$d3ea8090$c300000a@caliente> X-Mailer: Sylpheed version 0.9.7 (GTK+ 1.2.10; i686-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/81 X-Sequence-Number: 5256 On Wed, 7 Jan 2004 18:08:06 +0100 "Eric Jain" wrote: > Any tips for speeding up index creation? > > I need to bulk load a large table with 100M rows and several indexes, > some of which span two columns. > > By dropping all indexes prior to issuing the 'copy from' command, the > operation completes 10x as fast (1.5h vs 15h). > > Unfortunately, recreating a single index takes nearly as long as > loading all of the data into the table; this more or less eliminates > the time gained by dropping the index in the first place. > > Also, there doesn't seem to be a simple way to disable/recreate all > indexes for a specific table short of explicitely dropping and later > recreating each index? Before creating your index bump up your sort_mem high. set sort_mem = 64000 create index foo on baz(a, b); BIG increases. [This also helps on FK creation] -- Jeff Trout http://www.jefftrout.com/ http://www.stuarthamm.net/ From pgsql-general-owner@postgresql.org Wed Jan 7 15:42:23 2004 X-Original-To: pgsql-general-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C0E3DD1CA7D; Wed, 7 Jan 2004 19:42:19 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 82634-08; Wed, 7 Jan 2004 15:41:36 -0400 (AST) Received: from smtp.istop.com (dci.doncaster.on.ca [66.11.168.194]) by svr1.postgresql.org (Postfix) with ESMTP id 0A560D1B4AC; Wed, 7 Jan 2004 15:40:58 -0400 (AST) Received: from stark.xeocode.com (gsstark.mtl.istop.com [66.11.160.162]) by smtp.istop.com (Postfix) with ESMTP id 80E1E36C88; Wed, 7 Jan 2004 14:41:04 -0500 (EST) Received: from localhost ([127.0.0.1] helo=stark.dyndns.tv ident=foobar) by stark.xeocode.com with smtp (Exim 3.36 #1 (Debian)) id 1AeJY8-0005gj-00; Wed, 07 Jan 2004 14:41:04 -0500 To: "D. Dante Lorenso" Cc: "pgsql-general@postgresql.org" , pgsql-performance@postgresql.org Subject: Re: [PERFORM] Find original number of rows before applied LIMIT/OFFSET? References: <3FFC3A82.4000401@lorenso.com> In-Reply-To: <3FFC3A82.4000401@lorenso.com> From: Greg Stark Organization: The Emacs Conspiracy; member since 1992 Date: 07 Jan 2004 14:41:04 -0500 Message-ID: <873carwnqn.fsf@stark.dyndns.tv> Lines: 22 User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/280 X-Sequence-Number: 55308 "D. Dante Lorenso" writes: > Any thoughts? Sure, the PHP function I'm using above 'works', but is it > the most efficient? I hope I'm not actually pulling all 100,000 records > across the wire when I only intend to show 10 at a time. See what I'm > getting at? I tend to do it using a separate select count(*). My thinking is that the count(*) query can be simplified and exclude things like the ORDER BY clause and any select list entries that require extra work. It can often even exclude whole joins. By doing a separate query I can do that extra work only for the rows that i actually need for display. Hopefully using an index to pull up those rows. And do the count(*) in the most efficient way possible, probably a sequential scan with no joins for foreign keys etc. But I suspect the two methods both work out to suck about equally. -- greg From pgsql-performance-owner@postgresql.org Wed Jan 7 15:47:21 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 8E897D1B481 for ; Wed, 7 Jan 2004 19:47:19 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 85709-03 for ; Wed, 7 Jan 2004 15:46:36 -0400 (AST) Received: from smtp.istop.com (dci.doncaster.on.ca [66.11.168.194]) by svr1.postgresql.org (Postfix) with ESMTP id 81972D1B4D3 for ; Wed, 7 Jan 2004 15:46:27 -0400 (AST) Received: from stark.xeocode.com (gsstark.mtl.istop.com [66.11.160.162]) by smtp.istop.com (Postfix) with ESMTP id 18E2C36D05; Wed, 7 Jan 2004 14:45:25 -0500 (EST) Received: from localhost ([127.0.0.1] helo=stark.dyndns.tv ident=foobar) by stark.xeocode.com with smtp (Exim 3.36 #1 (Debian)) id 1AeJcK-0005i7-00; Wed, 07 Jan 2004 14:45:25 -0500 To: pgsql-performance@postgresql.org Subject: RAID array stripe sizes From: Greg Stark Organization: The Emacs Conspiracy; member since 1992 Date: 07 Jan 2004 14:45:24 -0500 Message-ID: <87wu83v8yz.fsf@stark.dyndns.tv> Lines: 7 User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/83 X-Sequence-Number: 5258 Does anyone have any data to support arguing for a particular stripe size in RAID-0? Do large stripe sizes allow drives to stream data more efficiently or defeat read-ahead? -- greg From pgsql-performance-owner@postgresql.org Wed Jan 7 16:10:31 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5779BD1B450 for ; Wed, 7 Jan 2004 20:10:29 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 91809-01 for ; Wed, 7 Jan 2004 16:09:47 -0400 (AST) Received: from mail1.ihs.com (mail1.ihs.com [170.207.70.222]) by svr1.postgresql.org (Postfix) with ESMTP id 4FD6CD1C957 for ; Wed, 7 Jan 2004 16:09:35 -0400 (AST) Received: from css120.ihs.com (css120.ihs.com [170.207.105.120]) by mail1.ihs.com (8.12.10/8.12.10) with ESMTP id i07K8DCs007136; Wed, 7 Jan 2004 13:08:14 -0700 (MST) Date: Wed, 7 Jan 2004 12:51:30 -0700 (MST) From: "scott.marlowe" To: Eric Jain Cc: pgsql-performance Subject: Re: Index creation In-Reply-To: <001101c3d540$d3ea8090$c300000a@caliente> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-IHS-MailScanner: Found to be clean X-IHS-MailScanner-SpamCheck: X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/84 X-Sequence-Number: 5259 On Wed, 7 Jan 2004, Eric Jain wrote: > Any tips for speeding up index creation? > > I need to bulk load a large table with 100M rows and several indexes, > some of which span two columns. > > By dropping all indexes prior to issuing the 'copy from' command, the > operation completes 10x as fast (1.5h vs 15h). > > Unfortunately, recreating a single index takes nearly as long as loading > all of the data into the table; this more or less eliminates the time > gained by dropping the index in the first place. > > Also, there doesn't seem to be a simple way to disable/recreate all > indexes for a specific table short of explicitely dropping and later > recreating each index? Note that you can issue the following command to see all the index definitions for a table: select * from pg_indexes where tablename='sometable'; And store those elsewhere to be reused when you need to recreate the index. From pgsql-performance-owner@postgresql.org Wed Jan 7 19:06:51 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C8C27D1B472 for ; Wed, 7 Jan 2004 23:06:50 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 14758-10 for ; Wed, 7 Jan 2004 19:06:08 -0400 (AST) Received: from mail.libertyrms.com (unknown [209.167.124.227]) by svr1.postgresql.org (Postfix) with ESMTP id 47896D1B4D3 for ; Wed, 7 Jan 2004 19:06:00 -0400 (AST) Received: from [10.1.2.146] (helo=dba3.int.libertyrms.info) by mail.libertyrms.com with esmtp (Exim 3.22 #3 (Debian)) id 1AeI4K-0005vM-00 for ; Wed, 07 Jan 2004 13:06:12 -0500 Received: by dba3.int.libertyrms.info (Postfix, from userid 1019) id 7596F1390B; Wed, 7 Jan 2004 18:06:08 -0500 (EST) Date: Wed, 7 Jan 2004 18:06:08 -0500 From: Andrew Sullivan To: pgsql-performance@postgresql.org Subject: failures on machines using jfs Message-ID: <20040107230608.GN12531@libertyrms.info> Mail-Followup-To: pgsql-performance@postgresql.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/85 X-Sequence-Number: 5260 Hi all, Chris Browne (one of my colleagues here) has posted some tests in the past indicating that jfs may be the fastest filesystem for Postgres use on Linux. We have lately had a couple of cases where machines either locked up, slowed down to the point of complete unusability, or died completely while using jfs. We are _not_ sure that jfs is in fact the culprit. In one case, a kernel panic appeared to be referring to the jfs kernel module, but I can't be sure as I lost the output immediately thereafter. Yesterday, we had a problem of data corruption on a failed jfs volume. None of this is to say that jfs is in fact to blame, nor even that, if it is, it does not have something to do with the age of our installations, &c. (these are all RH 8). In fact, I suspect hardware in both cases. But I thought I'd mention it just in case other people are seeing strange behaviour, on the principle of "better safe than sorry." A -- ---- Andrew Sullivan 204-4141 Yonge Street Afilias Canada Toronto, Ontario Canada M2P 2A8 +1 416 646 3304 x110 From pgsql-performance-owner@postgresql.org Thu Jan 8 22:51:37 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 6BCE4D1B45D for ; Thu, 8 Jan 2004 05:53:12 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 78638-04 for ; Thu, 8 Jan 2004 01:52:24 -0400 (AST) Received: from mail01.objectmastery.net (mail01.objectmastery.net [203.34.143.230]) by svr1.postgresql.org (Postfix) with ESMTP id 2D838D1B44C for ; Thu, 8 Jan 2004 01:52:22 -0400 (AST) Received: from objectmastery.com (rev.objectmastery.net [203.34.143.18]) by mail01.objectmastery.net (Postfix) with ESMTP id 0951F3FA5 for ; Thu, 8 Jan 2004 16:52:06 +1100 (EST) Message-ID: <3FFCF005.70108@objectmastery.com> Date: Thu, 08 Jan 2004 16:52:05 +1100 From: Bradley Tate User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030624 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: Slow query problem X-Enigmail-Version: 0.76.6.0 X-Enigmail-Supports: pgp-inline, pgp-mime Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/89 X-Sequence-Number: 5264 Hi, We've set up a little test box (1GHz Athlon, 40G IDE drive, 256M RAM, Redhat 9) to do some basic comparisons between postgresql and firebird 1.0.3 and 1.5rc8. Mostly the results are comparable, with one significant exception. QUERY select invheadref, invprodref, sum(units) from invtran group by invheadref, invprodref RESULTS pg 7.3.4 - 5.5 min pg 7.4.0 - 10 min fb 1.0.3 - 64 sec fb 1.5 - 44 sec * The invtran table has about 2.5 million records, invheadref and invprodref are both char(10) and indexed. * shared_buffers = 12000 and sort_mem = 8192 are the only changes I've made to postgresql.conf, with relevant changes to shmall and shmmax. This is an explain analyse plan from postgresql 7.4: QUERY PLAN ------------------------------------------------------------------------------------------------------------------------------------ GroupAggregate (cost=572484.23..601701.15 rows=1614140 width=39) (actual time=500091.171..554203.189 rows=147621 loops=1) -> Sort (cost=572484.23..578779.62 rows=2518157 width=39) (actual time=500090.939..527500.940 rows=2521530 loops=1) Sort Key: invheadref, invprodref -> Seq Scan on invtran (cost=0.00..112014.57 rows=2518157 width=39) (actual time=16.002..25516.917 rows=2521530 loops=1) Total runtime: 554826.827 ms (5 rows) Am I correct in interpreting that most time was spent doing the sorting? Explain confuses the heck out of me and any help on how I could make this run faster would be gratefully received. Cheers, Bradley. From pgsql-performance-owner@postgresql.org Thu Jan 8 15:17:05 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 67404D1B439 for ; Thu, 8 Jan 2004 19:17:04 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 38006-03 for ; Thu, 8 Jan 2004 15:16:15 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id 22479D1B43D for ; Thu, 8 Jan 2004 15:16:14 -0400 (AST) Received: from [63.195.55.98] (HELO temoku) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 4222151; Thu, 08 Jan 2004 11:17:12 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Reply-To: josh@agliodbs.com Organization: Aglio Database Solutions To: Andrew Sullivan , pgsql-performance@postgresql.org Subject: Re: failures on machines using jfs Date: Thu, 8 Jan 2004 10:52:40 -0800 User-Agent: KMail/1.4.3 References: <20040107230608.GN12531@libertyrms.info> In-Reply-To: <20040107230608.GN12531@libertyrms.info> MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Message-Id: <200401081052.40807.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/86 X-Sequence-Number: 5261 Andrew, > None of this is to say that jfs is in fact to blame, nor even that, > if it is, it does not have something to do with the age of our > installations, &c. (these are all RH 8). In fact, I suspect hardware > in both cases. But I thought I'd mention it just in case other > people are seeing strange behaviour, on the principle of "better > safe than sorry." Always useful. Actually, I just fielded on IRC a report of poor I/O=20 utilization with XFS during checkpointing. Not sure if the problem is XF= S=20 or PostgreSQL, but the fact that XFS (alone among filesystems) does its own= =20 cache management instead of using the kernel cache makes me suspicious. --=20 -Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Sun Jan 11 15:16:38 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5FE96D1B450 for ; Fri, 9 Jan 2004 03:23:51 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 28314-06 for ; Thu, 8 Jan 2004 23:23:05 -0400 (AST) Received: from mail.hive.nj2.inquent.com (mc.carriermail.com [205.178.180.9]) by svr1.postgresql.org (Postfix) with SMTP id 082E7D1B482 for ; Thu, 8 Jan 2004 23:17:04 -0400 (AST) Received: (qmail 16135 invoked from network); 9 Jan 2004 03:17:34 -0000 Received: from unknown (HELO ?192.168.1.199?) (134.22.68.14) by 205.178.180.9 with SMTP; 9 Jan 2004 03:17:34 -0000 Subject: Re: optimizing Postgres queries From: Rod Taylor To: David Teran Cc: Tom Lane , Postgresql Performance In-Reply-To: <6F2414E8-3FB9-11D8-A528-000A95A6F0DC@cluster9.com> References: <4B0023AC-3F72-11D8-B3EB-000A95A6F0DC@cluster9.com> <200401051705.55120.shridhar_daithankar@myrealbox.com> <64C8DBBA-3F77-11D8-A528-000A95A6F0DC@cluster9.com> <200401051740.05006.shridhar_daithankar@myrealbox.com> <3796212E-3F79-11D8-A528-000A95A6F0DC@cluster9.com> <9966.1073316152@sss.pgh.pa.us> <9CC6A792-3FAF-11D8-A528-000A95A6F0DC@cluster9.com> <16230.1073328760@sss.pgh.pa.us> <20540.1073329548@sss.pgh.pa.us> <4530AF9E-3FB4-11D8-A528-000A95A6F0DC@cluster9.com> <20759.1073330625@sss.pgh.pa.us> <6F2414E8-3FB9-11D8-A528-000A95A6F0DC@cluster9.com> Content-Type: text/plain Message-Id: <1073618208.322.18.camel@jester> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Thu, 08 Jan 2004 22:16:49 -0500 Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/109 X-Sequence-Number: 5284 On Mon, 2004-01-05 at 14:57, David Teran wrote: > ... wow: > > executing a batch file with about 4250 selects, including lots of joins > other things PostgreSQL 7.4 is about 2 times faster than FrontBase > 3.6.27. OK, we will start to make larger tests but this is quite > interesting already: we did not optimize a lot, just invoked VACUUM > ANALYZE and then the selects ;-) > > Thanks to all who answered to this thread. I presume that batch file was executed linearly -- no parallelism? You're actually testing one of PostgreSQL's shortcomings. PostgreSQL (in my experience) does much better in such comparisons with a parallel load -- multiple connections executing varied work (short selects, complex selects, inserts, updates, deletes). Anyway, just a tip that you will want to test your actual load. If you do batch work with a single thread, what you have is fine. But if you have a website with tens or hundreds of simultaneous connections then your non-parallel testing will not reflect that work load. From pgsql-performance-owner@postgresql.org Thu Jan 8 23:29:47 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id CB31CD1B44C for ; Fri, 9 Jan 2004 03:29:45 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 27922-09 for ; Thu, 8 Jan 2004 23:28:59 -0400 (AST) Received: from pomeray.duluoz.net (12-241-192-70.client.attbi.com [12.241.192.70]) by svr1.postgresql.org (Postfix) with ESMTP id 89B5FD1B4D4 for ; Thu, 8 Jan 2004 23:27:17 -0400 (AST) Received: (qmail 6232 invoked by uid 1000); 9 Jan 2004 03:27:20 -0000 Date: Thu, 8 Jan 2004 19:27:16 -0800 From: Mike Glover To: Bradley Tate Cc: pgsql-performance@postgresql.org Subject: Re: Slow query problem Message-Id: <20040108192716.26c4b369.mpg4@duluoz.net> In-Reply-To: <3FFCF005.70108@objectmastery.com> References: <3FFCF005.70108@objectmastery.com> X-Mailer: Sylpheed version 0.9.0claws (GTK+ 1.2.10; i686-pc-linux-gnu) Mime-Version: 1.0 Content-Type: multipart/signed; protocol="application/pgp-signature"; micalg="pgp-sha1"; boundary="=.U(0Zldy2q1+SH7" X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/91 X-Sequence-Number: 5266 --=.U(0Zldy2q1+SH7 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit On Thu, 08 Jan 2004 16:52:05 +1100 Bradley Tate wrote: > Am I correct in interpreting that most time was spent doing the > sorting? looks so. your table is about 70MB total size, and its getting loaded completely into memory (you have 12000 * 8k = 96M available). 26s to load 70MB from disk seems reasonable. The rest of the time is used for sorting. > Explain confuses the heck out of me and any help on how I could make > this run faster would be gratefully received. > You should bump sort_mem as high as you can stand. with only 8MB sort memory available, you're swapping intermediate sort pages to disk -- a lot. Try the query with sort_mem set to 75MB (to do the entire sort in memory). -mike > Cheers, > > Bradley. > > > ---------------------------(end of > broadcast)--------------------------- TIP 8: explain analyze is your > friend -- Mike Glover Key ID BFD19F2C --=.U(0Zldy2q1+SH7 Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.3 (GNU/Linux) iD8DBQE//h+YZrNpxr/RnywRAqFgAJ940IieUUxhaZesAzC4Rv933yVCDACdH3DU 4ylBJp0SmxnCdnuks1v54Jg= =VDUZ -----END PGP SIGNATURE----- --=.U(0Zldy2q1+SH7-- From pgsql-performance-owner@postgresql.org Fri Jan 9 00:19:34 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id F0F43D1B443 for ; Fri, 9 Jan 2004 04:19:32 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 31524-09 for ; Fri, 9 Jan 2004 00:18:44 -0400 (AST) Received: from slgan.com (c-67-163-28-202.client.comcast.net [67.163.28.202]) by svr1.postgresql.org (Postfix) with ESMTP id E85A6D1B468 for ; Fri, 9 Jan 2004 00:18:42 -0400 (AST) Received: from [192.168.10.154] ([192.168.10.154]) by slgan.com (8.10.2/8.10.2) with ESMTP id i095KCL04308 for ; Thu, 8 Jan 2004 23:20:12 -0600 Mime-Version: 1.0 X-Sender: slgan@192.168.10.10 Message-Id: Date: Thu, 8 Jan 2004 22:19:53 -0600 To: pgsql-performance@postgresql.org From: Seum-Lim Gan Subject: PostgreSQL vs. Oracle disk space usage Content-Type: text/plain; charset="us-ascii" ; format="flowed" X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/92 X-Sequence-Number: 5267 Hi, I searched through the archive and could not find any conclusive discussion of results on this. Has anyone compared the disk space usage between PostgreSQL and Oracle ? I am interested in knowing for the same tuple (i.e same dictionary), the disk usage between the two. Thanks. Gan -- +--------------------------------------------------------+ | Seum-Lim GAN email : slgan@lucent.com | | Lucent Technologies | | 2000 N. Naperville Road, 6B-403F tel : (630)-713-6665 | | Naperville, IL 60566, USA. fax : (630)-713-7272 | | web : http://inuweb.ih.lucent.com/~slgan | +--------------------------------------------------------+ From pgsql-performance-owner@postgresql.org Fri Jan 9 00:24:00 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A7815D1B49E for ; Fri, 9 Jan 2004 04:23:59 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 38596-01 for ; Fri, 9 Jan 2004 00:23:09 -0400 (AST) Received: from wolff.to (wolff.to [66.93.249.74]) by svr1.postgresql.org (Postfix) with SMTP id EB9E6D1B44F for ; Fri, 9 Jan 2004 00:23:08 -0400 (AST) Received: (qmail 7165 invoked by uid 500); 9 Jan 2004 04:23:38 -0000 Date: Thu, 8 Jan 2004 22:23:38 -0600 From: Bruno Wolff III To: Mike Glover Cc: Bradley Tate , pgsql-performance@postgresql.org Subject: Re: Slow query problem Message-ID: <20040109042338.GD31586@wolff.to> Mail-Followup-To: Mike Glover , Bradley Tate , pgsql-performance@postgresql.org References: <3FFCF005.70108@objectmastery.com> <20040108192716.26c4b369.mpg4@duluoz.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20040108192716.26c4b369.mpg4@duluoz.net> User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/93 X-Sequence-Number: 5268 On Thu, Jan 08, 2004 at 19:27:16 -0800, Mike Glover wrote: > > You should bump sort_mem as high as you can stand. with only 8MB sort > memory available, you're swapping intermediate sort pages to disk -- > a lot. Try the query with sort_mem set to 75MB (to do the entire sort in > memory). Postgres also might be able to switch to a hash aggregate instead of using a sort if sortmem is made large enough to hold the results for all of the (estimated) groups. From pgsql-performance-owner@postgresql.org Fri Jan 9 01:13:21 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 4E7CED1B551 for ; Fri, 9 Jan 2004 05:13:20 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 45965-02 for ; Fri, 9 Jan 2004 01:12:30 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 171B1D1B4D4 for ; Fri, 9 Jan 2004 01:12:20 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i095C719020724; Fri, 9 Jan 2004 00:12:08 -0500 (EST) To: Mike Glover Cc: Bradley Tate , pgsql-performance@postgresql.org Subject: Re: Slow query problem In-reply-to: <20040108192716.26c4b369.mpg4@duluoz.net> References: <3FFCF005.70108@objectmastery.com> <20040108192716.26c4b369.mpg4@duluoz.net> Comments: In-reply-to Mike Glover message dated "Thu, 08 Jan 2004 19:27:16 -0800" Date: Fri, 09 Jan 2004 00:12:07 -0500 Message-ID: <20723.1073625127@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/94 X-Sequence-Number: 5269 Mike Glover writes: > You should bump sort_mem as high as you can stand. with only 8MB sort > memory available, you're swapping intermediate sort pages to disk -- > a lot. Try the query with sort_mem set to 75MB (to do the entire sort in > memory). 7.4 will probably flip over to a hash-based aggregation method, and not sort at all, once you make sort_mem large enough that it thinks the hash table will fit in sort_mem. regards, tom lane From pgsql-performance-owner@postgresql.org Fri Jan 9 03:30:47 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 88677D1B479 for ; Fri, 9 Jan 2004 07:30:46 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 60758-02 for ; Fri, 9 Jan 2004 03:29:57 -0400 (AST) Received: from zigo.dhs.org (as2-4-3.an.g.bonet.se [194.236.34.191]) by svr1.postgresql.org (Postfix) with ESMTP id 67161D1B457 for ; Fri, 9 Jan 2004 03:29:56 -0400 (AST) Received: from zigo.zigo.dhs.org (zigo.zigo.dhs.org [192.168.0.1]) by zigo.dhs.org (Postfix) with ESMTP id 0F90E8E0D; Fri, 9 Jan 2004 08:29:57 +0100 (CET) Date: Fri, 9 Jan 2004 08:29:57 +0100 (CET) From: =?ISO-8859-1?Q?Dennis_Bj=F6rklund?= To: Bradley Tate Cc: pgsql-performance@postgresql.org Subject: Re: Slow query problem In-Reply-To: <3FFCF005.70108@objectmastery.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=ISO-8859-1 Content-Transfer-Encoding: 8BIT X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/95 X-Sequence-Number: 5270 On Thu, 8 Jan 2004, Bradley Tate wrote: > We've set up a little test box (1GHz Athlon, 40G IDE drive, 256M RAM, > Redhat 9) to do some basic comparisons between postgresql and firebird > 1.0.3 and 1.5rc8. Mostly the results are comparable, with one > significant exception. > > QUERY > select invheadref, invprodref, sum(units) > from invtran > group by invheadref, invprodref > > RESULTS > pg 7.3.4 - 5.5 min > pg 7.4.0 - 10 min > fb 1.0.3 - 64 sec > fb 1.5 - 44 sec > > * The invtran table has about 2.5 million records, invheadref and > invprodref are both char(10) and indexed. For the above query, shouldn't you have one index for both columns (invheadref, invprodref). Then it should not need to sort at all to do the grouping and it should all be fast. -- /Dennis Bj�rklund From pgsql-performance-owner@postgresql.org Fri Jan 9 04:55:40 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D3D01D1B473 for ; Fri, 9 Jan 2004 08:55:38 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 72653-09 for ; Fri, 9 Jan 2004 04:54:50 -0400 (AST) Received: from anchor-post-32.mail.demon.net (anchor-post-32.mail.demon.net [194.217.242.90]) by svr1.postgresql.org (Postfix) with ESMTP id A1AEED1B477 for ; Fri, 9 Jan 2004 04:54:48 -0400 (AST) Received: from mwynhau.demon.co.uk ([193.237.186.96] helo=mainbox.archonet.com) by anchor-post-32.mail.demon.net with esmtp (Exim 3.35 #1) id 1AesPp-000MeW-0W; Fri, 09 Jan 2004 08:54:49 +0000 Received: by mainbox.archonet.com (Postfix, from userid 529) id ADF3517942; Fri, 9 Jan 2004 08:54:48 +0000 (GMT) Received: from client17.archonet.com (client17.archonet.com [192.168.1.17]) by mainbox.archonet.com (Postfix) with ESMTP id 12008178D5; Fri, 9 Jan 2004 08:54:47 +0000 (GMT) From: Richard Huxton To: Dennis =?iso-8859-1?q?Bj=F6rklund?= , Bradley Tate Subject: Re: Slow query problem Date: Fri, 9 Jan 2004 08:54:46 +0000 User-Agent: KMail/1.5 Cc: pgsql-performance@postgresql.org References: In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Message-Id: <200401090854.46480.dev@archonet.com> X-Bogosity: No, tests=bogofilter, spamicity=0.000000, version=0.15.3 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/96 X-Sequence-Number: 5271 On Friday 09 January 2004 07:29, Dennis Bj=F6rklund wrote: > On Thu, 8 Jan 2004, Bradley Tate wrote: > > > > select invheadref, invprodref, sum(units) > > from invtran > > group by invheadref, invprodref > For the above query, shouldn't you have one index for both columns > (invheadref, invprodref). Then it should not need to sort at all to do the > grouping and it should all be fast. Not sure if that would make a difference here, since the whole table is bei= ng=20 read.=20 --=20 Richard Huxton Archonet Ltd From pgsql-performance-owner@postgresql.org Fri Jan 9 04:57:59 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 417A6D1B450 for ; Fri, 9 Jan 2004 08:57:58 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 75877-03 for ; Fri, 9 Jan 2004 04:57:10 -0400 (AST) Received: from zigo.dhs.org (as2-4-3.an.g.bonet.se [194.236.34.191]) by svr1.postgresql.org (Postfix) with ESMTP id 12C94D1B457 for ; Fri, 9 Jan 2004 04:57:08 -0400 (AST) Received: from zigo.zigo.dhs.org (zigo.zigo.dhs.org [192.168.0.1]) by zigo.dhs.org (Postfix) with ESMTP id 526FA8E0D; Fri, 9 Jan 2004 09:57:09 +0100 (CET) Date: Fri, 9 Jan 2004 09:57:09 +0100 (CET) From: =?ISO-8859-1?Q?Dennis_Bj=F6rklund?= To: Richard Huxton Cc: Bradley Tate , Subject: Re: Slow query problem In-Reply-To: <200401090854.46480.dev@archonet.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=ISO-8859-1 Content-Transfer-Encoding: 8BIT X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/97 X-Sequence-Number: 5272 On Fri, 9 Jan 2004, Richard Huxton wrote: > > > select invheadref, invprodref, sum(units) > > > from invtran > > > group by invheadref, invprodref > > > For the above query, shouldn't you have one index for both columns > > (invheadref, invprodref). Then it should not need to sort at all to do the > > grouping and it should all be fast. > > Not sure if that would make a difference here, since the whole table is being > read. The goal was to avoid the sorting which should not be needed with that index (I hope). So I still think that it would help in this case. -- /Dennis Bj�rklund From pgsql-performance-owner@postgresql.org Fri Jan 9 05:19:59 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A6A2FD1B454 for ; Fri, 9 Jan 2004 09:19:57 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 76501-06 for ; Fri, 9 Jan 2004 05:19:09 -0400 (AST) Received: from anchor-post-34.mail.demon.net (anchor-post-34.mail.demon.net [194.217.242.92]) by svr1.postgresql.org (Postfix) with ESMTP id 202EAD1B466 for ; Fri, 9 Jan 2004 05:19:07 -0400 (AST) Received: from mwynhau.demon.co.uk ([193.237.186.96] helo=mainbox.archonet.com) by anchor-post-34.mail.demon.net with esmtp (Exim 3.35 #1) id 1AesnM-0002bV-0Y; Fri, 09 Jan 2004 09:19:08 +0000 Received: by mainbox.archonet.com (Postfix, from userid 529) id 3579217972; Fri, 9 Jan 2004 09:19:07 +0000 (GMT) Received: from client17.archonet.com (client17.archonet.com [192.168.1.17]) by mainbox.archonet.com (Postfix) with ESMTP id 4F6EA17969; Fri, 9 Jan 2004 09:19:05 +0000 (GMT) From: Richard Huxton To: Dennis =?iso-8859-1?q?Bj=F6rklund?= Subject: Re: Slow query problem Date: Fri, 9 Jan 2004 09:19:04 +0000 User-Agent: KMail/1.5 Cc: Bradley Tate , References: In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Message-Id: <200401090919.04718.dev@archonet.com> X-Bogosity: No, tests=bogofilter, spamicity=0.000000, version=0.15.3 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/98 X-Sequence-Number: 5273 On Friday 09 January 2004 08:57, Dennis Bj=F6rklund wrote: > On Fri, 9 Jan 2004, Richard Huxton wrote: > > > > select invheadref, invprodref, sum(units) > > > > from invtran > > > > group by invheadref, invprodref > > > > > > For the above query, shouldn't you have one index for both columns > > > (invheadref, invprodref). Then it should not need to sort at all to do > > > the grouping and it should all be fast. > > > > Not sure if that would make a difference here, since the whole table is > > being read. > > The goal was to avoid the sorting which should not be needed with that > index (I hope). So I still think that it would help in this case. Sorry - not being clear. I can see how it _might_ help, but will the planne= r=20 take into account the fact that even though: index-cost > seqscan-cost that (index-cost + no-sorting) < (seqscan-cost + sort-cost) assuming of course, that the costs turn out that way. --=20 Richard Huxton Archonet Ltd From pgsql-performance-owner@postgresql.org Sun Jan 11 15:20:55 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 3A502D1B495 for ; Fri, 9 Jan 2004 14:13:36 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 43217-01 for ; Fri, 9 Jan 2004 10:12:49 -0400 (AST) Received: from localhost.net (cust.89.121.adsl.cistron.nl [195.64.89.121]) by svr1.postgresql.org (Postfix) with ESMTP id 9E599D1B537 for ; Fri, 9 Jan 2004 10:12:45 -0400 (AST) Received: from trust-factory.com (localhost [127.0.0.1]) by localhost.net (Postfix) with ESMTP id 9B70083C80 for ; Fri, 9 Jan 2004 15:12:43 +0100 (CET) Message-ID: <3FFEB6DA.8000905@trust-factory.com> Date: Fri, 09 Jan 2004 15:12:42 +0100 From: Richard van den Berg Organization: Trust Factory User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5) Gecko/20031007 X-Accept-Language: en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: Explain not accurate Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/110 X-Sequence-Number: 5285 Hi there, I am quite new to postgresql, and love the explain feature. It enables us to predict which SQL queries needs to be optimized before we see any problems. However, I've run into an issue where explain tells us a the costs of a quiry are tremendous (105849017586), but the query actually runs quite fast. Even "explain analyze" shows these costs. This makes me wonder: can the estimates explain shows be dead wrong? I can explain in more detail (including the query and output of explain) if needed. I'm using 7.4 on Solaris 8. Sincerely, -- Richard van den Berg, CISSP Trust Factory B.V. | http://www.trust-factory.com/ Bazarstraat 44a | Phone: +31 70 3620684 NL-2518AK The Hague | Fax : +31 70 3603009 The Netherlands | Visit us at Lotusphere 2004 http://www.trust-factory.com/lotusphere From pgsql-performance-owner@postgresql.org Fri Jan 9 12:26:07 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 08DB2D1B4BD for ; Fri, 9 Jan 2004 16:26:04 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 04103-09 for ; Fri, 9 Jan 2004 12:25:33 -0400 (AST) Received: from mail012.syd.optusnet.com.au (mail012.syd.optusnet.com.au [211.29.132.66]) by svr1.postgresql.org (Postfix) with ESMTP id 4C3ECD1DB76 for ; Fri, 9 Jan 2004 10:53:50 -0400 (AST) Received: from objectmastery.com (c211-28-197-254.eburwd1.vic.optusnet.com.au [211.28.197.254]) by mail012.syd.optusnet.com.au (8.11.6p2/8.11.6) with ESMTP id i09Erfr07771; Sat, 10 Jan 2004 01:53:42 +1100 Message-ID: <3FFEC075.2090001@objectmastery.com> Date: Sat, 10 Jan 2004 01:53:41 +1100 From: Bradley Tate User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030703 X-Accept-Language: en-us, en MIME-Version: 1.0 To: =?ISO-8859-1?Q?Dennis_Bj=F6rklund?= Cc: Richard Huxton , pgsql-performance@postgresql.org Subject: Re: Slow query problem References: In-Reply-To: X-Enigmail-Version: 0.76.7.0 X-Enigmail-Supports: pgp-inline, pgp-mime Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 8bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/99 X-Sequence-Number: 5274 Dennis Bj�rklund wrote: >On Fri, 9 Jan 2004, Richard Huxton wrote: > > > >>>>select invheadref, invprodref, sum(units) >>>>from invtran >>>>group by invheadref, invprodref >>>> >>>> >>>For the above query, shouldn't you have one index for both columns >>>(invheadref, invprodref). Then it should not need to sort at all to do the >>>grouping and it should all be fast. >>> >>> >>Not sure if that would make a difference here, since the whole table is being >>read. >> >> > >The goal was to avoid the sorting which should not be needed with that >index (I hope). So I still think that it would help in this case. > > > Thanks for the advice. I tried creating a compound index along with clustering the invtran table on it, adding another 512MB RAM, increasing shared_buffers to 60000 and increasing sort_mem to 100MB, playing with effective cache size in postgresql.conf. This cut the execution time down to 4 minutes, which was helpful but still way behind firebird. There was still an awful lot of disk activity while it was happening which seems to imply lots of sorting going on (?) Invtran is a big table but it is clustered and static i.e. no updates, select statements only. Mostly my performance problems are with sorts - group by, order by. I was hoping for better results than I've been getting so far. Thanks. p.s. Can someone confirm whether this should work from pgadmin3? i.e. will the size of the sort_mem be changed for the duration of the query or session? set sort_mem to 100000; select ....etc....; From pgsql-performance-owner@postgresql.org Fri Jan 9 12:42:41 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 803BED1B478 for ; Fri, 9 Jan 2004 16:42:39 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 30156-10 for ; Fri, 9 Jan 2004 12:42:08 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 0625BD1DC76 for ; Fri, 9 Jan 2004 11:07:11 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i09F7919023793; Fri, 9 Jan 2004 10:07:09 -0500 (EST) To: Richard Huxton Cc: Dennis =?iso-8859-1?q?Bj=F6rklund?= , Bradley Tate , pgsql-performance@postgresql.org Subject: Re: Slow query problem In-reply-to: <200401090919.04718.dev@archonet.com> References: <200401090919.04718.dev@archonet.com> Comments: In-reply-to Richard Huxton message dated "Fri, 09 Jan 2004 09:19:04 +0000" Date: Fri, 09 Jan 2004 10:07:09 -0500 Message-ID: <23792.1073660829@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/100 X-Sequence-Number: 5275 Richard Huxton writes: >> The goal was to avoid the sorting which should not be needed with that >> index (I hope). So I still think that it would help in this case. > Sorry - not being clear. I can see how it _might_ help, but will the planner > take into account the fact that even though: > index-cost > seqscan-cost > that > (index-cost + no-sorting) < (seqscan-cost + sort-cost) Yes, it would. > assuming of course, that the costs turn out that way. That I'm less sure about. A sort frequently looks cheaper than a full indexscan, unless the table is pretty well clustered on that index, or you knock random_page_cost way down. With no stats at all, CVS tip has these preferences: regression=# create table fooey (f1 int, f2 int, unique(f1,f2)); NOTICE: CREATE TABLE / UNIQUE will create implicit index "fooey_f1_key" for table "fooey" CREATE TABLE regression=# explain select * from fooey group by f1,f2; QUERY PLAN --------------------------------------------------------------- HashAggregate (cost=25.00..25.00 rows=1000 width=8) -> Seq Scan on fooey (cost=0.00..20.00 rows=1000 width=8) (2 rows) regression=# set enable_hashagg TO 0; SET regression=# explain select * from fooey group by f1,f2; QUERY PLAN ------------------------------------------------------------------------------------ Group (cost=0.00..57.00 rows=1000 width=8) -> Index Scan using fooey_f1_key on fooey (cost=0.00..52.00 rows=1000 width=8) (2 rows) regression=# set enable_indexscan TO 0; SET regression=# explain select * from fooey group by f1,f2; QUERY PLAN --------------------------------------------------------------------- Group (cost=69.83..77.33 rows=1000 width=8) -> Sort (cost=69.83..72.33 rows=1000 width=8) Sort Key: f1, f2 -> Seq Scan on fooey (cost=0.00..20.00 rows=1000 width=8) (4 rows) but remember this is for a relatively small (estimated size of) table. regards, tom lane From pgsql-performance-owner@postgresql.org Fri Jan 9 13:16:09 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 16871D1BB71 for ; Fri, 9 Jan 2004 17:16:06 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 38574-08 for ; Fri, 9 Jan 2004 13:15:34 -0400 (AST) Received: from megazone.bigpanda.com (megazone.bigpanda.com [64.147.171.210]) by svr1.postgresql.org (Postfix) with ESMTP id AB644D1DDA8 for ; Fri, 9 Jan 2004 11:38:12 -0400 (AST) Received: by megazone.bigpanda.com (Postfix, from userid 1001) id E36E335566; Fri, 9 Jan 2004 07:38:01 -0800 (PST) Received: from localhost (localhost [127.0.0.1]) by megazone.bigpanda.com (Postfix) with ESMTP id E1DB23539B; Fri, 9 Jan 2004 07:38:01 -0800 (PST) Date: Fri, 9 Jan 2004 07:38:01 -0800 (PST) From: Stephan Szabo To: Richard Huxton Cc: Dennis =?iso-8859-1?q?Bj=F6rklund?= , Bradley Tate , pgsql-performance@postgresql.org Subject: Re: Slow query problem In-Reply-To: <200401090919.04718.dev@archonet.com> Message-ID: <20040109073400.D69748@megazone.bigpanda.com> References: <200401090919.04718.dev@archonet.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=X-UNKNOWN Content-Transfer-Encoding: QUOTED-PRINTABLE X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/101 X-Sequence-Number: 5276 On Fri, 9 Jan 2004, Richard Huxton wrote: > On Friday 09 January 2004 08:57, Dennis Bj=F6rklund wrote: > > On Fri, 9 Jan 2004, Richard Huxton wrote: > > > > > select invheadref, invprodref, sum(units) > > > > > from invtran > > > > > group by invheadref, invprodref > > > > > > > > For the above query, shouldn't you have one index for both columns > > > > (invheadref, invprodref). Then it should not need to sort at all to= do > > > > the grouping and it should all be fast. > > > > > > Not sure if that would make a difference here, since the whole table = is > > > being read. > > > > The goal was to avoid the sorting which should not be needed with that > > index (I hope). So I still think that it would help in this case. > > Sorry - not being clear. I can see how it _might_ help, but will the plan= ner > take into account the fact that even though: > index-cost > seqscan-cost > that > (index-cost + no-sorting) < (seqscan-cost + sort-cost) > assuming of course, that the costs turn out that way. AFAICS, yes it does take that effect into account (as best it can with the estimates). From pgsql-performance-owner@postgresql.org Fri Jan 9 23:30:02 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5CAD6D1B4D3 for ; Sat, 10 Jan 2004 03:29:45 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 30346-08 for ; Fri, 9 Jan 2004 23:29:18 -0400 (AST) Received: from srvr3.iniquinet.com (srvr2.iniquinet.com [64.240.87.12]) by svr1.postgresql.org (Postfix) with SMTP id AA72CD1B46B for ; Fri, 9 Jan 2004 23:29:13 -0400 (AST) Received: (qmail 8682 invoked by uid 104); 10 Jan 2004 03:22:07 -0000 Received: from Robert_Creager@LogicalChaos.org by srvr3.iniquinet.com by uid 101 with qmail-scanner-1.15 (clamscan: 0.54. spamassassin: 2.55. Clear:SA:0(-7.8/6.0):. Processed in 34.927623 secs); 10 Jan 2004 03:22:07 -0000 Received: from vsat-148-64-8-86.c119.t7.mrt.starband.net (HELO chaos.mshome.net) (148.64.8.86) by srvr3.iniquinet.com with SMTP; 10 Jan 2004 03:21:32 -0000 Received: from localhost.localdomain (thunder.mshome.net [192.168.0.250]) by chaos.mshome.net (Postfix) with ESMTP id C701A172966; Fri, 9 Jan 2004 20:28:24 -0700 (MST) Received: from logicalchaos.org (thunder.mshome.net [192.168.0.250]) by chaos.mshome.net (Postfix) with SMTP id A5C51172966; Fri, 9 Jan 2004 20:28:22 -0700 (MST) Date: Fri, 9 Jan 2004 20:28:16 -0700 From: Robert Creager To: Andrew Sullivan Cc: pgsql-performance@postgresql.org Subject: Re: failures on machines using jfs Message-Id: <20040109202816.38072cf8.Robert_Creager@LogicalChaos.org> In-Reply-To: <20040107230608.GN12531@libertyrms.info> References: <20040107230608.GN12531@libertyrms.info> Organization: Starlight Vision, LLC. X-Mailer: Sylpheed version 0.9.5claws (GTK+ 1.2.10; i586-mandrake-linux-gnu) Mime-Version: 1.0 Content-Type: multipart/signed; protocol="application/pgp-signature"; micalg="pgp-sha1"; boundary="Multipart_Fri__9_Jan_2004_20_28_16_-0700_A.V)I+=.HNNonbC." X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/102 X-Sequence-Number: 5277 --Multipart_Fri__9_Jan_2004_20_28_16_-0700_A.V)I+=.HNNonbC. Content-Type: text/plain; charset=US-ASCII Content-Disposition: inline Content-Transfer-Encoding: 7bit When grilled further on (Wed, 7 Jan 2004 18:06:08 -0500), Andrew Sullivan confessed: > > We have lately had a couple of cases where machines either locked up, > slowed down to the point of complete unusability, or died completely > while using jfs. We are _not_ sure that jfs is in fact the culprit. > In one case, a kernel panic appeared to be referring to the jfs > kernel module, but I can't be sure as I lost the output immediately > thereafter. Yesterday, we had a problem of data corruption on a > failed jfs volume. > > None of this is to say that jfs is in fact to blame, nor even that, > if it is, it does not have something to do with the age of our > installations, &c. (these are all RH 8). In fact, I suspect hardware > in both cases. But I thought I'd mention it just in case other > people are seeing strange behaviour, on the principle of "better > safe than sorry." > Interestingly enough, I'm using JFS on a new scsi disk with Mandrake 9.1 and was having similar problems. I was generating heavy disk usage through database and astronomical data reductions. My machine (dual AMD) would suddenly hang. No new jobs would run, just increase the load, until I reboot the machine. I solved my problems by creating a 128Mb ram disk (using EXT2) for the temp data produced my reduction runs. I believe JFS was to blame, not hardware, but you never know... Cheers, Rob -- 20:22:27 up 12 days, 10:13, 4 users, load average: 2.00, 2.01, 2.03 --Multipart_Fri__9_Jan_2004_20_28_16_-0700_A.V)I+=.HNNonbC. Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.2 (GNU/Linux) iEYEARECAAYFAj//cVYACgkQLQ/DKuwDYzkAygCffjpZt3eIcb1eTlUCZU3rAbCF XBAAnimY+oRBTwhe7aXkQWxTIC2W53bP =EdCh -----END PGP SIGNATURE----- --Multipart_Fri__9_Jan_2004_20_28_16_-0700_A.V)I+=.HNNonbC.-- From pgsql-performance-owner@postgresql.org Sat Jan 10 22:36:04 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id EB0F3D1D56A for ; Sun, 11 Jan 2004 02:36:03 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 85381-03 for ; Sat, 10 Jan 2004 22:35:35 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id 2DCC1D1D643 for ; Sat, 10 Jan 2004 22:35:32 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id i0B2ZVU6099942 for ; Sun, 11 Jan 2004 02:35:31 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id i0B2A1gH065715 for pgsql-performance@postgresql.org; Sun, 11 Jan 2004 02:10:01 GMT From: Christopher Browne X-Newsgroups: comp.databases.postgresql.performance Subject: Re: failures on machines using jfs Date: Sat, 10 Jan 2004 21:08:50 -0500 Organization: Hub.Org Networking Services Lines: 56 Message-ID: <60k73zck3x.fsf@dev6.int.libertyrms.info> References: <20040107230608.GN12531@libertyrms.info> <20040109202816.38072cf8.Robert_Creager@LogicalChaos.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Complaints-To: usenet@news.hub.org User-Agent: Gnus/5.1004 (Gnus v5.10.4) XEmacs/21.4 (Reasonable Discussion, linux) Cancel-Lock: sha1:WSkb524vYimex3rYkNGK2HYPAYI= To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/103 X-Sequence-Number: 5278 Robert_Creager@LogicalChaos.org (Robert Creager) writes: > When grilled further on (Wed, 7 Jan 2004 18:06:08 -0500), > Andrew Sullivan confessed: > >> We have lately had a couple of cases where machines either locked >> up, slowed down to the point of complete unusability, or died >> completely while using jfs. We are _not_ sure that jfs is in fact >> the culprit. In one case, a kernel panic appeared to be referring >> to the jfs kernel module, but I can't be sure as I lost the output >> immediately thereafter. Yesterday, we had a problem of data >> corruption on a failed jfs volume. >> >> None of this is to say that jfs is in fact to blame, nor even that, >> if it is, it does not have something to do with the age of our >> installations, &c. (these are all RH 8). In fact, I suspect >> hardware in both cases. But I thought I'd mention it just in case >> other people are seeing strange behaviour, on the principle of >> "better safe than sorry." > > Interestingly enough, I'm using JFS on a new scsi disk with Mandrake > 9.1 and was having similar problems. I was generating heavy disk > usage through database and astronomical data reductions. My machine > (dual AMD) would suddenly hang. No new jobs would run, just > increase the load, until I reboot the machine. > > I solved my problems by creating a 128Mb ram disk (using EXT2) for > the temp data produced my reduction runs. > > I believe JFS was to blame, not hardware, but you never know... Interesting. The set of concurrent factors that came together to appear when this happened "consistently" were thus: 1. Heavy DB updates taking place on JFS filesystems; 2. SMP (we suspected Xeon hyperthreading as a possible factor, but shut it off and still saw the same problem...) 3. The third factor that appeared a catalyst was copying, via scp, a file > 2GB in size onto the system. The third piece was a particularly interesting aspect; the file would get copied over successfully, and the scp process would hang (to the point of "kill -9" being unable to touch it) immediately thereafter. At that point, processes on the system that were accessing files on the hung-up filesystem were locked, also unkillable by "kill 9." That's certainly consistent with JFS being at the root of the problem, whether it was the cause or not... -- let name="cbbrowne" and tld="libertyrms.info" in String.concat "@" [name;tld];; Christopher Browne (416) 646 3304 x124 (land) From pgsql-performance-owner@postgresql.org Sun Jan 11 11:54:17 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 565B6D1D56B for ; Sun, 11 Jan 2004 15:54:16 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 62380-07 for ; Sun, 11 Jan 2004 11:53:48 -0400 (AST) Received: from sccrmhc13.comcast.net (sccrmhc13.comcast.net [204.127.202.64]) by svr1.postgresql.org (Postfix) with ESMTP id 66C05D1D561 for ; Sun, 11 Jan 2004 11:53:39 -0400 (AST) Received: from lorenso.com (c-24-1-26-144.client.comcast.net[24.1.26.144]) by comcast.net (sccrmhc13) with ESMTP id <2004011115533801600nikgte>; Sun, 11 Jan 2004 15:53:38 +0000 Message-ID: <40017182.1090801@lorenso.com> Date: Sun, 11 Jan 2004 09:53:38 -0600 From: "D. Dante Lorenso" User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.6b) Gecko/20031205 Thunderbird/0.4 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: Postgresql on Quad CPU machine Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/104 X-Sequence-Number: 5279 I'm running PostgreSQL 7.4 on a quad Xeon attached to a beefy disk array. However, I am begining to wonder if this is a waste of CPU power. I think I read somewhere that PostgreSQL is NOT multi-threaded. But, will it be able to take advantage of multiple CPUs? Will I have to run separate postmaster instances to get the advantage? I'm not running a high load on that machine yet, So I can't tell if the load is being balanced across the CPUs. I expect that as some of the newly launched sites grow it will require more resources but maybe some of you could share your results of this type of deployment setup. Dante From pgsql-performance-owner@postgresql.org Sun Jan 11 12:08:09 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id F1DACD1B519 for ; Sun, 11 Jan 2004 16:08:07 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 64273-07 for ; Sun, 11 Jan 2004 12:07:36 -0400 (AST) Received: from main.gmane.org (main.gmane.org [80.91.224.249]) by svr1.postgresql.org (Postfix) with ESMTP id 98E20D1D552 for ; Sun, 11 Jan 2004 12:07:35 -0400 (AST) Received: from list by main.gmane.org with local (Exim 3.35 #1 (Debian)) id 1Afi7b-0000XA-00 for ; Sun, 11 Jan 2004 17:07:27 +0100 X-Injected-Via-Gmane: http://gmane.org/ To: pgsql-performance@postgresql.org Received: from sea.gmane.org ([80.91.224.252]) by main.gmane.org with esmtp (Exim 3.35 #1 (Debian)) id 1Afi7a-0000X2-00 for ; Sun, 11 Jan 2004 17:07:26 +0100 Received: from news by sea.gmane.org with local (Exim 3.35 #1 (Debian)) id 1Afi7a-0005cd-00 for ; Sun, 11 Jan 2004 17:07:26 +0100 From: Doug McNaught Subject: Re: Postgresql on Quad CPU machine Date: Sun, 11 Jan 2004 11:07:25 -0500 Lines: 15 Message-ID: <87wu7y5v0i.fsf@asmodeus.mcnaught.org> References: <40017182.1090801@lorenso.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Complaints-To: usenet@sea.gmane.org User-Agent: Gnus/5.1002 (Gnus v5.10.2) Emacs/20.7 (gnu/linux) Cancel-Lock: sha1:9ZwWNdwZdg0oHwf0jkdAnM3l0gY= X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/105 X-Sequence-Number: 5280 "D. Dante Lorenso" writes: > I'm running PostgreSQL 7.4 on a quad Xeon attached to a > beefy disk array. However, I am begining to wonder if this is > a waste of CPU power. > > I think I read somewhere that PostgreSQL is NOT multi-threaded. > But, will it be able to take advantage of multiple CPUs? Will > I have to run separate postmaster instances to get the advantage? PG uses a separate backend process for each connection, so if you have multiple simultaneous connections they will use different CPUs. Single queries will not be split across CPUs. -Doug From pgsql-performance-owner@postgresql.org Sun Jan 11 12:18:11 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1178ED1B498 for ; Sun, 11 Jan 2004 16:18:10 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 65207-08 for ; Sun, 11 Jan 2004 12:17:39 -0400 (AST) Received: from COLSWEEPER.cranel.com (newmail.cranel.com [12.32.71.147]) by svr1.postgresql.org (Postfix) with ESMTP id 9F683D1B436 for ; Sun, 11 Jan 2004 12:17:38 -0400 (AST) Received: from colmail01.cranel.com (colmail01.cranel.com) by COLSWEEPER.cranel.com (Content Technologies SMTPRS 4.3.12) with ESMTP id ; Sun, 11 Jan 2004 11:16:28 -0500 Received: by colmail01.cranel.com with Internet Mail Service (5.5.2657.72) id ; Sun, 11 Jan 2004 11:21:15 -0500 Message-ID: <387C22290D3FD71195D300508BF7DB5238AE92@colmail01.cranel.com> From: "Spiegelberg, Greg" To: 'Christopher Browne ' , "'pgsql-performance@postgresql.org '" Subject: Re: failures on machines using jfs Date: Sun, 11 Jan 2004 11:21:14 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2657.72) Content-Type: text/plain; charset="iso-8859-1" X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/106 X-Sequence-Number: 5281 It would seem we're experiencing somthing similiar with our scratch volume (JFS mounted with noatime). It is still much faster than our experiments with ext2, ext3, and reiserfs but occasionally during large loads it will hiccup for a couple seconds but no crashes yet. I'm reluctant to switch back to any other file system because the data import took a little over 1.5 hours but now takes just under 20 minutes and we haven't crashed yet. For future reference: RedHat 7.3 w/2.4.18-18.7smp PostgreSQL 7.3.3 from source jfsutils 1.0.17-1 Dual PIII Intel 1.4GHz & 2GB ECC Internal disk: 2xU160 SCSI, mirrored, location of our JFS file system External disk Qlogic 2310 attached to FC-SW @2Gbps with ext3 on those LUNs Greg -----Original Message----- From: Christopher Browne To: pgsql-performance@postgresql.org Sent: 1/10/04 9:08 PM Subject: Re: [PERFORM] failures on machines using jfs Robert_Creager@LogicalChaos.org (Robert Creager) writes: > When grilled further on (Wed, 7 Jan 2004 18:06:08 -0500), > Andrew Sullivan confessed: > >> We have lately had a couple of cases where machines either locked >> up, slowed down to the point of complete unusability, or died >> completely while using jfs. We are _not_ sure that jfs is in fact >> the culprit. In one case, a kernel panic appeared to be referring >> to the jfs kernel module, but I can't be sure as I lost the output >> immediately thereafter. Yesterday, we had a problem of data >> corruption on a failed jfs volume. >> >> None of this is to say that jfs is in fact to blame, nor even that, >> if it is, it does not have something to do with the age of our >> installations, &c. (these are all RH 8). In fact, I suspect >> hardware in both cases. But I thought I'd mention it just in case >> other people are seeing strange behaviour, on the principle of >> "better safe than sorry." > > Interestingly enough, I'm using JFS on a new scsi disk with Mandrake > 9.1 and was having similar problems. I was generating heavy disk > usage through database and astronomical data reductions. My machine > (dual AMD) would suddenly hang. No new jobs would run, just > increase the load, until I reboot the machine. > > I solved my problems by creating a 128Mb ram disk (using EXT2) for > the temp data produced my reduction runs. > > I believe JFS was to blame, not hardware, but you never know... Interesting. The set of concurrent factors that came together to appear when this happened "consistently" were thus: 1. Heavy DB updates taking place on JFS filesystems; 2. SMP (we suspected Xeon hyperthreading as a possible factor, but shut it off and still saw the same problem...) 3. The third factor that appeared a catalyst was copying, via scp, a file > 2GB in size onto the system. The third piece was a particularly interesting aspect; the file would get copied over successfully, and the scp process would hang (to the point of "kill -9" being unable to touch it) immediately thereafter. At that point, processes on the system that were accessing files on the hung-up filesystem were locked, also unkillable by "kill 9." That's certainly consistent with JFS being at the root of the problem, whether it was the cause or not... -- let name="cbbrowne" and tld="libertyrms.info" in String.concat "@" [name;tld];; Christopher Browne (416) 646 3304 x124 (land) ---------------------------(end of broadcast)--------------------------- TIP 8: explain analyze is your friend ********************************************************************** This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the system manager. This footnote also confirms that this email message has been swept by MIMEsweeper for the presence of computer viruses. www.mimesweeper.com ********************************************************************** From pgsql-performance-owner@postgresql.org Sun Jan 11 13:05:30 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id AE45ED1D236 for ; Sun, 11 Jan 2004 17:05:24 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 73790-07 for ; Sun, 11 Jan 2004 13:04:54 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id B27CAD1B85F for ; Sun, 11 Jan 2004 13:04:53 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i0BH4l19016863; Sun, 11 Jan 2004 12:04:47 -0500 (EST) To: "Spiegelberg, Greg" Cc: "'Christopher Browne '" , "'pgsql-performance@postgresql.org '" Subject: Re: failures on machines using jfs In-reply-to: <387C22290D3FD71195D300508BF7DB5238AE92@colmail01.cranel.com> References: <387C22290D3FD71195D300508BF7DB5238AE92@colmail01.cranel.com> Comments: In-reply-to "Spiegelberg, Greg" message dated "Sun, 11 Jan 2004 11:21:14 -0500" Date: Sun, 11 Jan 2004 12:04:47 -0500 Message-ID: <16862.1073840687@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/107 X-Sequence-Number: 5282 "Spiegelberg, Greg" writes: > PostgreSQL 7.3.3 from source *Please* update to 7.3.4 or 7.3.5 before you get bitten by the WAL-page-boundary bug ... regards, tom lane From pgsql-performance-owner@postgresql.org Sun Jan 11 14:11:38 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1BBD0D1D27E for ; Sun, 11 Jan 2004 18:11:38 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 83282-04 for ; Sun, 11 Jan 2004 14:11:07 -0400 (AST) Received: from swan.mail.pas.earthlink.net (swan.mail.pas.earthlink.net [207.217.120.123]) by svr1.postgresql.org (Postfix) with ESMTP id 0F5EFD1B495 for ; Sun, 11 Jan 2004 14:11:06 -0400 (AST) Received: from user-11204mn.dsl.mindspring.com ([66.32.18.215] helo=shadovitzcmptr) by swan.mail.pas.earthlink.net with smtp (Exim 3.33 #1) id 1Afk3F-0000U2-00 for pgsql-performance@postgresql.org; Sun, 11 Jan 2004 10:11:05 -0800 Received: by localhost with Microsoft MAPI; Sun, 11 Jan 2004 10:11:04 -0800 Message-ID: <01C3D82B.388A7FE0.david@shadovitz.com> From: David Shadovitz Reply-To: "david@shadovitz.com" To: "pgsql-performance@postgresql.org" Subject: COUNT & Pagination Date: Sun, 11 Jan 2004 10:10:52 -0800 X-Mailer: Microsoft Internet E-mail/MAPI - 8.0.0.4211 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/108 X-Sequence-Number: 5283 I understand that COUNT queries are expensive. So I'm looking for advice on displaying paginated query results. I display my query results like this: Displaying 1 to 50 of 2905. 1-50 | 51-100 | 101-150 | etc. I do this by executing two queries. One is of the form: SELECT FROM WHERE LIMIT m OFFSET n > > The other is identical except that I replace the select list with COUNT(*). yes, you need 2 query. Or select it from one: select *, (select count(*) from table) as count from table... pg will optimize this query, and do the count only once > > And an unrelated question: > I'm running PG 7.2.2 and want to upgrade to 7.4.1. I've never upgraded PG > before and I'm nervous. Can I simply run pg_dumpall, install 7.4.1, and then > feed the dump into psql? I'm planning to use pg_dumpall rather than pg_dump > because I want to preserve the users I've defined. My database is the only one > on the system. yes. But check tha faq and the manual for a better explain. C. From pgsql-performance-owner@postgresql.org Sun Jan 11 17:00:34 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D569DD1DA85 for ; Sun, 11 Jan 2004 21:00:33 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 23446-04 for ; Sun, 11 Jan 2004 17:00:03 -0400 (AST) Received: from avocet.mail.pas.earthlink.net (avocet.mail.pas.earthlink.net [207.217.120.50]) by svr1.postgresql.org (Postfix) with ESMTP id 5BE94D1D9EE for ; Sun, 11 Jan 2004 17:00:01 -0400 (AST) Received: from user-v8ldv2j.dsl.mindspring.com ([209.86.252.83]) by avocet.mail.pas.earthlink.net with esmtp (Exim 3.33 #1) id 1Afmgc-0006pt-00; Sun, 11 Jan 2004 12:59:54 -0800 In-Reply-To: <01C3D82B.388A7FE0.david@shadovitz.com> References: <01C3D82B.388A7FE0.david@shadovitz.com> Mime-Version: 1.0 (Apple Message framework v609) Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: <107436B6-4479-11D8-BE00-000A959C6000@cloverpub.com> Content-Transfer-Encoding: 7bit Cc: pgsql-performance@postgresql.org From: Jeff Fitzmyers Subject: Re: COUNT & Pagination Date: Sun, 11 Jan 2004 12:59:36 -0800 To: "david@shadovitz.com" X-Mailer: Apple Mail (2.609) X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/112 X-Sequence-Number: 5287 > So I'm looking for advice on displaying paginated query results. > Displaying 1 to 50 of 2905. > 1-50 | 51-100 | 101-150 | etc. > > I do this by executing two queries. One is of the form: > SELECT FROM WHERE LIMIT m OFFSET n > > The other is identical except that I replace the select list with COUNT(*). > > I'm looking for suggestions to replace that COUNT query. I cannot use the > method of storing the number of records in a separate table because my queries > (a) involve joins, and (b) have a WHERE clause. Well, on all my sites, I do what you do and just live with it :P You can investigate using cursors however (DECLARE, MOVE & FETCH) > And an unrelated question: > I'm running PG 7.2.2 and want to upgrade to 7.4.1. I've never upgraded PG > before and I'm nervous. Can I simply run pg_dumpall, install 7.4.1, and then > feed the dump into psql? I'm planning to use pg_dumpall rather than pg_dump > because I want to preserve the users I've defined. My database is the only one > on the system. I recommend something like this: -- disable access to your database to make sure you have a complete dump -- run dump as database owner account su pgsql (or whatever your postgres user is) -- do compressed dump pg_dumpall > backup.sql -- backup old data dir mv /usr/local/pgsql/data /usr/local/pgsql/data.7.2 -- remove old postgres, install new -- run NEW initdb. replace latin1 with your encoding -- -W specifies a superuser password initdb -D /usr/local/pgsql/data -E LATIN1 -W -- restore dump, watching output VERY CAREFULLY: -- (run as pgsql user again) psql template1 < backup.sql > log.txt -- Watch stderr very carefully to check any errors that might occur. -- If restore fails, re-initdb and re-restore Chris From pgsql-performance-owner@postgresql.org Sun Jan 11 22:56:01 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 589F6D1B46A for ; Mon, 12 Jan 2004 02:56:00 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 63726-04 for ; Sun, 11 Jan 2004 22:55:32 -0400 (AST) Received: from smtp.istop.com (dci.doncaster.on.ca [66.11.168.194]) by svr1.postgresql.org (Postfix) with ESMTP id 6610FD1D576 for ; Sun, 11 Jan 2004 22:55:28 -0400 (AST) Received: from stark.xeocode.com (gsstark.mtl.istop.com [66.11.160.162]) by smtp.istop.com (Postfix) with ESMTP id EC9B436AA6; Sun, 11 Jan 2004 21:55:26 -0500 (EST) Received: from localhost ([127.0.0.1] helo=stark.xeocode.com ident=foobar) by stark.xeocode.com with smtp (Exim 3.36 #1 (Debian)) id 1AfsEg-0006J8-00; Sun, 11 Jan 2004 21:55:26 -0500 To: Richard van den Berg Cc: pgsql-performance@postgresql.org Subject: Re: Explain not accurate References: <3FFEB6DA.8000905@trust-factory.com> In-Reply-To: <3FFEB6DA.8000905@trust-factory.com> From: Greg Stark Organization: The Emacs Conspiracy; member since 1992 Date: 11 Jan 2004 21:55:26 -0500 Message-ID: <87smil3mg1.fsf@stark.xeocode.com> Lines: 27 User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/115 X-Sequence-Number: 5290 Richard van den Berg writes: > Hi there, > > I am quite new to postgresql, and love the explain feature. It enables us to > predict which SQL queries needs to be optimized before we see any problems. > However, I've run into an issue where explain tells us a the costs of a quiry > are tremendous (105849017586), but the query actually runs quite fast. Even > "explain analyze" shows these costs. Do you have any of the optimization parameters off, enable_seqscan perhaps? enable_seqscan works by penalizing plans that use sequential plans, but there are still lots of queries that cannot be done any other way. I'm not sure whether the same holds for all the other parameters. If your tables are all going to grow drastically then this may still indicate a problem, probably a missing index. But if one of them is a reference table that will never grow then perhaps the index will never be necessary. Or perhaps you just need to run analyze. Send the "EXPLAIN ANALYZE" output for the query for starters. You might also send the output of "SHOW ALL". -- greg From pgsql-performance-owner@postgresql.org Sun Jan 11 23:04:43 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 512AAD1B46A for ; Mon, 12 Jan 2004 03:04:43 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 65384-02 for ; Sun, 11 Jan 2004 23:04:16 -0400 (AST) Received: from steelhead.ravensfield.com (unknown [65.222.52.254]) by svr1.postgresql.org (Postfix) with ESMTP id 8F879D1B55D for ; Sun, 11 Jan 2004 23:04:11 -0400 (AST) Received: from [10.4.3.100] (unknown [10.4.3.100]) by steelhead.ravensfield.com (Postfix) with ESMTP id 8E5C866148 for ; Sun, 11 Jan 2004 22:04:15 -0500 (EST) Mime-Version: 1.0 (Apple Message framework v609) Content-Transfer-Encoding: 7bit Message-Id: <22837C75-44AC-11D8-8262-000393A47FCC@ravensfield.com> Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed To: pgsql-performance@postgresql.org From: Andrew Rawnsley Subject: annoying query/planner choice Date: Sun, 11 Jan 2004 22:05:11 -0500 X-Mailer: Apple Mail (2.609) X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/116 X-Sequence-Number: 5291 I have a situation that is giving me small fits, and would like to see if anyone can shed any light on it. I have a modest table (@1.4 million rows, and growing), that has a variety of queries run against it. One is a very straightforward one - pull a set of distinct rows out based on two columns, with a simple where clause based on one of the indexed columns. For illustration here, I've removed the distinct and order-by clauses, as they are not the culprits. Before I go on - v7.4.1, currently on a test box, dual P3, 1G ram, 10K scsi, Slackware 9 or so. The table has been vacuumed and analyzed. Even offered pizza and beer. Production box will be a dual Xeon with 2G ram and RAID 5. When the query is run with a where clause that returns small number of rows, the query uses the index and is quite speedy: rav=# explain analyze select casno, parameter from hai.results where site_id = 9982; QUERY PLAN ------------------------------------------------------------------------ -------------------------------------------------------------- Index Scan using hai_res_siteid_ndx on results (cost=0.00..7720.87 rows=2394 width=30) (actual time=12.118..12.933 rows=50 loops=1) Index Cond: (site_id = 9982) Total runtime: 13.145 ms When a query is run that returns a much larger set, the index is not used, I assume because the planner thinks that a sequential scan would work just as well with a large result set: rav=# explain analyze select casno, parameter from hai.results where site_id = 18; QUERY PLAN ------------------------------------------------------------------------ ---------------------------------------------- Seq Scan on results (cost=0.00..73396.39 rows=211205 width=30) (actual time=619.020..15012.807 rows=186564 loops=1) Filter: (site_id = 18) Total runtime: 15279.789 ms (3 rows) Unfortunately, its way off: rav=# set enable_seqscan=off; SET rav=# explain analyze select casno, parameter from hai.results where site_id = 18; QUERY PLAN ------------------------------------------------------------------------ ----------------------------------------------------------------------- Index Scan using hai_res_siteid_ndx on results (cost=0.00..678587.01 rows=211205 width=30) (actual time=9.575..3569.387 rows=186564 loops=1) Index Cond: (site_id = 18) Total runtime: 3872.292 ms (3 rows) I would like, of course, for it to use the index, given that it takes 20-25% of the time. Fiddling with CPU_TUPLE_COST doesn't do anything until I exceed 0.5, which strikes me as a bit high (though please correct me if I am assuming too much...). RANDOM_PAGE_COST seems to have no effect. I suppose I could cluster it, but it is constantly being added to, and would have to be re-done on a daily basis (if not more). Any suggestions? -------------------- Andrew Rawnsley President The Ravensfield Digital Resource Group, Ltd. (740) 587-0114 www.ravensfield.com From pgsql-performance-owner@postgresql.org Sun Jan 11 23:50:54 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9216FD1DA1E for ; Mon, 12 Jan 2004 03:50:53 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 72705-07 for ; Sun, 11 Jan 2004 23:50:26 -0400 (AST) Received: from zigo.dhs.org (as2-4-3.an.g.bonet.se [194.236.34.191]) by svr1.postgresql.org (Postfix) with ESMTP id 610ADD1D56B for ; Sun, 11 Jan 2004 23:50:21 -0400 (AST) Received: from zigo.zigo.dhs.org (zigo.zigo.dhs.org [192.168.0.1]) by zigo.dhs.org (Postfix) with ESMTP id 6CE968E0D; Mon, 12 Jan 2004 04:50:25 +0100 (CET) Date: Mon, 12 Jan 2004 04:50:25 +0100 (CET) From: Dennis Bjorklund To: Andrew Rawnsley Cc: pgsql-performance@postgresql.org Subject: Re: annoying query/planner choice In-Reply-To: <22837C75-44AC-11D8-8262-000393A47FCC@ravensfield.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=ISO-8859-1 Content-Transfer-Encoding: 8BIT X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/117 X-Sequence-Number: 5292 On Sun, 11 Jan 2004, Andrew Rawnsley wrote: > 20-25% of the time. Fiddling with CPU_TUPLE_COST doesn't do anything > until I exceed 0.5, which strikes me as a bit high (though please > correct me if I am assuming too much...). RANDOM_PAGE_COST seems to have > no effect. What about the effective cache size, is that set properly? -- /Dennis Bj�rklund From pgsql-performance-owner@postgresql.org Mon Jan 12 00:06:24 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id AC660D1B49F for ; Mon, 12 Jan 2004 04:06:22 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 75261-02 for ; Mon, 12 Jan 2004 00:05:51 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id 0A2BCD1B7ED for ; Mon, 12 Jan 2004 00:05:51 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id i0C45nU8038339 for ; Mon, 12 Jan 2004 04:05:50 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id i0C44ref037187 for pgsql-performance@postgresql.org; Mon, 12 Jan 2004 04:04:53 GMT From: Christopher Browne X-Newsgroups: comp.databases.postgresql.performance Subject: Re: annoying query/planner choice Date: Sun, 11 Jan 2004 22:56:59 -0500 Organization: cbbrowne Computing Inc Lines: 42 Message-ID: References: <22837C75-44AC-11D8-8262-000393A47FCC@ravensfield.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Complaints-To: usenet@news.hub.org X-message-flag: Outlook is rather hackable, isn't it? X-Home-Page: http://www.cbbrowne.com/info/ X-Affero: http://svcs.affero.net/rm.php?r=cbbrowne User-Agent: Gnus/5.1006 (Gnus v5.10.6) XEmacs/21.4 (Reasonable Discussion, linux) Cancel-Lock: sha1:kVlrcw0iF1ckQouIwioUuVc8GL0= To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/119 X-Sequence-Number: 5294 Centuries ago, Nostradamus foresaw when ronz@ravensfield.com (Andrew Rawnsley) would write: > I would like, of course, for it to use the index, given that it > takes 20-25% of the time. Fiddling with CPU_TUPLE_COST doesn't do > anything until I exceed 0.5, which strikes me as a bit high (though > please correct me if I am assuming too much...). RANDOM_PAGE_COST > seems to have no effect. I suppose I could cluster it, but it is > constantly being added to, and would have to be re-done on a daily > basis (if not more). > > Any suggestions? The apparent problem is a bad query plan, and for clustering to "fix" it seems a disturbing answer. A problem I saw last week with some query plans pointed to the issue that the statistics were inadequate. We had some queries where indexing on "customer" is extremely worthwhile in nearly all cases, but it often wasn't happening. The problem was that the 10 "bins" in the default stats table would collect up stats about a few _highly_ active customers, and pretty much ignore the less active ones. Because the "bins" were highly dominated by the few common values, stats for the others were missing and pretty useless. I upped the size of the histogram from 10 to 100, and that allowed stats to be kept for less active customers, GREATLY improving the quality of the queries. The point that falls out is that if you have a column which has a bunch of discrete values (rather more than 10) that aren't near-unique (e.g. - on a table with a million transactions, you have a only few hundred customers), that's a good candidate for upping column stats. Thus, you might try: ALTER TABLE MY_TABLE ALTER COLUMN SOME_COLUMN SET STATISTICS 50; ANALYZE MY_TABLE; -- let name="cbbrowne" and tld="ntlug.org" in name ^ "@" ^ tld;; http://www.ntlug.org/~cbbrowne/postgresql.html "There's no longer a boycott of Apple. But MacOS is still a proprietary OS." -- RMS - June 13, 1998 From pgsql-performance-owner@postgresql.org Mon Jan 12 00:04:48 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 739F6D1B7E7 for ; Mon, 12 Jan 2004 04:04:47 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 73502-05 for ; Mon, 12 Jan 2004 00:04:15 -0400 (AST) Received: from steelhead.ravensfield.com (unknown [65.222.52.254]) by svr1.postgresql.org (Postfix) with ESMTP id 6E6DAD1B7ED for ; Mon, 12 Jan 2004 00:04:15 -0400 (AST) Received: from [10.4.3.100] (unknown [10.4.3.100]) by steelhead.ravensfield.com (Postfix) with ESMTP id AD8B766148; Sun, 11 Jan 2004 23:04:14 -0500 (EST) In-Reply-To: References: Mime-Version: 1.0 (Apple Message framework v609) Content-Type: text/plain; charset=ISO-8859-1; format=flowed Message-Id: <83C6D7BA-44B4-11D8-8262-000393A47FCC@ravensfield.com> Content-Transfer-Encoding: quoted-printable Cc: pgsql-performance@postgresql.org From: Andrew Rawnsley Subject: Re: annoying query/planner choice Date: Sun, 11 Jan 2004 23:05:10 -0500 To: Dennis Bjorklund X-Mailer: Apple Mail (2.609) X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/118 X-Sequence-Number: 5293 Low (1000). I'll fiddle with that. I just noticed that the machine only=20 has 512MB of ram in it, and not 1GB. I must have raided it for some other machine... On Jan 11, 2004, at 10:50 PM, Dennis Bjorklund wrote: > On Sun, 11 Jan 2004, Andrew Rawnsley wrote: > >> 20-25% of the time. Fiddling with CPU_TUPLE_COST doesn't do anything >> until I exceed 0.5, which strikes me as a bit high (though please >> correct me if I am assuming too much...). RANDOM_PAGE_COST seems to=20 >> have >> no effect. > > What about the effective cache size, is that set properly? > > --=20 > /Dennis Bj=F6rklund > > > ---------------------------(end of=20 > broadcast)--------------------------- > TIP 1: subscribe and unsubscribe commands go to=20 > majordomo@postgresql.org > -------------------- Andrew Rawnsley President The Ravensfield Digital Resource Group, Ltd. (740) 587-0114 www.ravensfield.com From pgsql-performance-owner@postgresql.org Mon Jan 12 01:40:47 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 2B6CBD1B499 for ; Mon, 12 Jan 2004 05:40:46 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 85187-06 for ; Mon, 12 Jan 2004 01:40:15 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 21E55D1B519 for ; Mon, 12 Jan 2004 01:40:14 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i0C5eD19021521; Mon, 12 Jan 2004 00:40:13 -0500 (EST) To: Andrew Rawnsley Cc: pgsql-performance@postgresql.org Subject: Re: annoying query/planner choice In-reply-to: <22837C75-44AC-11D8-8262-000393A47FCC@ravensfield.com> References: <22837C75-44AC-11D8-8262-000393A47FCC@ravensfield.com> Comments: In-reply-to Andrew Rawnsley message dated "Sun, 11 Jan 2004 22:05:11 -0500" Date: Mon, 12 Jan 2004 00:40:13 -0500 Message-ID: <21520.1073886013@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/120 X-Sequence-Number: 5295 Andrew Rawnsley writes: > I have a situation that is giving me small fits, and would like to see > if anyone can shed any light on it. In general, pulling 10% of a table *should* be faster as a seqscan than an indexscan, except under the most extreme assumptions about clustering (is the table clustered on site_id, by any chance?). What I suspect is that the table is a bit larger than your available RAM, so that a seqscan ends up flushing all of the kernel's cache and forcing a lot of I/O, whereas an indexscan avoids the cache flush by not touching (quite) all of the table. The trouble with this is that the index only looks that good under test conditions, ie, when you repeat it just after an identical query that pulled all of the needed pages into RAM. Under realistic load conditions where different site_ids are being hit, the indexscan is not going to be as good as you think, because it will incur substantial I/O. You should try setting up a realistic test load hitting different random site_ids, and see whether it's really a win to force seqscan off for this query or not. regards, tom lane From pgsql-performance-owner@postgresql.org Mon Jan 12 06:17:14 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 81DA5D1B432 for ; Mon, 12 Jan 2004 10:17:13 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 28168-02 for ; Mon, 12 Jan 2004 06:16:44 -0400 (AST) Received: from anchor-post-31.mail.demon.net (anchor-post-31.mail.demon.net [194.217.242.89]) by svr1.postgresql.org (Postfix) with ESMTP id B52ADD1DA6C for ; Mon, 12 Jan 2004 06:16:41 -0400 (AST) Received: from mwynhau.demon.co.uk ([193.237.186.96] helo=mainbox.archonet.com) by anchor-post-31.mail.demon.net with esmtp (Exim 3.35 #1) id 1Afz7g-0002gY-0V; Mon, 12 Jan 2004 10:16:41 +0000 Received: by mainbox.archonet.com (Postfix, from userid 529) id C08A016E2D; Mon, 12 Jan 2004 10:16:39 +0000 (GMT) Received: from client17.archonet.com (client17.archonet.com [192.168.1.17]) by mainbox.archonet.com (Postfix) with ESMTP id 4A98816DC9; Mon, 12 Jan 2004 10:16:38 +0000 (GMT) From: Richard Huxton To: "david@shadovitz.com" , "pgsql-performance@postgresql.org" Subject: Re: COUNT & Pagination Date: Mon, 12 Jan 2004 10:16:37 +0000 User-Agent: KMail/1.5 References: <01C3D82B.388A7FE0.david@shadovitz.com> In-Reply-To: <01C3D82B.388A7FE0.david@shadovitz.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200401121016.37918.dev@archonet.com> X-Bogosity: No, tests=bogofilter, spamicity=0.000000, version=0.15.3 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/121 X-Sequence-Number: 5296 On Sunday 11 January 2004 18:10, David Shadovitz wrote: > I understand that COUNT queries are expensive. So I'm looking for advice > on displaying paginated query results. > > I display my query results like this: > > Displaying 1 to 50 of 2905. > 1-50 | 51-100 | 101-150 | etc. > > I do this by executing two queries. If you only need the count when you've got the results, most PG client interfaces will tell you how many rows you've got. What language is your app in? -- Richard Huxton Archonet Ltd From pgsql-performance-owner@postgresql.org Mon Jan 12 11:01:44 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id DF875D1D541 for ; Mon, 12 Jan 2004 15:01:42 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 78667-05 for ; Mon, 12 Jan 2004 11:01:15 -0400 (AST) Received: from steelhead.ravensfield.com (unknown [65.222.52.254]) by svr1.postgresql.org (Postfix) with ESMTP id 370E5D1DA1E for ; Mon, 12 Jan 2004 11:01:09 -0400 (AST) Received: from [10.1.1.7] (smallmouth.oh.ia [10.1.1.7]) by steelhead.ravensfield.com (Postfix) with ESMTP id 4A9FC66148; Mon, 12 Jan 2004 10:01:12 -0500 (EST) In-Reply-To: <21520.1073886013@sss.pgh.pa.us> References: <22837C75-44AC-11D8-8262-000393A47FCC@ravensfield.com> <21520.1073886013@sss.pgh.pa.us> Mime-Version: 1.0 (Apple Message framework v609) Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: <4B2B3053-4510-11D8-BF17-000393A47FCC@ravensfield.com> Content-Transfer-Encoding: 7bit Cc: pgsql-performance@postgresql.org From: Andrew Rawnsley Subject: Re: annoying query/planner choice Date: Mon, 12 Jan 2004 10:02:09 -0500 To: Tom Lane X-Mailer: Apple Mail (2.609) X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/122 X-Sequence-Number: 5297 Probably my best solution is to find a better way to produce the information, or cache it on the application side, as it doesn't actually change that much across client sessions. Clustering it occurred to me - it would have to be done on a frequent basis, as the contents of the table change constantly. What I am getting out of it with this operation doesn't change much, so caching in a separate table, in the application layer, or both would probably shortcut the whole problem. Always amazing what occurs to you when you sleep on it...if only I could take a good nap in the middle of the afternoon I would have no problems at all. On Jan 12, 2004, at 12:40 AM, Tom Lane wrote: > Andrew Rawnsley writes: >> I have a situation that is giving me small fits, and would like to see >> if anyone can shed any light on it. > > In general, pulling 10% of a table *should* be faster as a seqscan than > an indexscan, except under the most extreme assumptions about > clustering > (is the table clustered on site_id, by any chance?). What I suspect is > that the table is a bit larger than your available RAM, so that a > seqscan ends up flushing all of the kernel's cache and forcing a lot of > I/O, whereas an indexscan avoids the cache flush by not touching > (quite) > all of the table. The trouble with this is that the index only looks > that good under test conditions, ie, when you repeat it just after an > identical query that pulled all of the needed pages into RAM. Under > realistic load conditions where different site_ids are being hit, the > indexscan is not going to be as good as you think, because it will > incur > substantial I/O. > > You should try setting up a realistic test load hitting different > random > site_ids, and see whether it's really a win to force seqscan off for > this query or not. > > regards, tom lane > > ---------------------------(end of > broadcast)--------------------------- > TIP 4: Don't 'kill -9' the postmaster > -------------------- Andrew Rawnsley President The Ravensfield Digital Resource Group, Ltd. (740) 587-0114 www.ravensfield.com From pgsql-performance-owner@postgresql.org Mon Jan 12 11:39:16 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 05C76D1D27C for ; Mon, 12 Jan 2004 15:39:16 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 89483-01 for ; Mon, 12 Jan 2004 11:38:48 -0400 (AST) Received: from pintail.mail.pas.earthlink.net (pintail.mail.pas.earthlink.net [207.217.120.122]) by svr1.postgresql.org (Postfix) with ESMTP id E9A0BD1B43F for ; Mon, 12 Jan 2004 11:38:43 -0400 (AST) Received: from hsa201.pool031.at101.earthlink.net ([216.249.100.201] helo=shadovitzcmptr) by pintail.mail.pas.earthlink.net with smtp (Exim 3.33 #1) id 1Ag49L-0003wd-00; Mon, 12 Jan 2004 07:38:43 -0800 Received: by localhost with Microsoft MAPI; Mon, 12 Jan 2004 07:39:04 -0800 Message-ID: <01C3D8DF.271378C0.david@shadovitz.com> From: David Shadovitz Reply-To: "david@shadovitz.com" To: "pgsql-performance@postgresql.org" Cc: 'Richard Huxton' Subject: Re: COUNT & Pagination Date: Mon, 12 Jan 2004 07:37:36 -0800 X-Mailer: Microsoft Internet E-mail/MAPI - 8.0.0.4211 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/123 X-Sequence-Number: 5298 > If you only need the count when you've got the results, most PG client > interfaces will tell you how many rows you've got. What language is your app > in? PHP. But I have only a subset of the results, retrieved via a query with a "LIMIT " clause, so $pg_numrows is m. And retrieving all results (i.e. no LIMIT) is at least as expensive as COUNT(*). -David From pgsql-performance-owner@postgresql.org Mon Jan 12 12:13:07 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9438AD1B55D for ; Mon, 12 Jan 2004 16:13:06 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 97380-01 for ; Mon, 12 Jan 2004 12:12:32 -0400 (AST) Received: from fuji.krosing.net (silmet.estpak.ee [194.126.97.78]) by svr1.postgresql.org (Postfix) with ESMTP id EAF3DD1D9F2 for ; Mon, 12 Jan 2004 12:12:15 -0400 (AST) Received: from fuji.krosing.net (localhost.localdomain [127.0.0.1]) by fuji.krosing.net (8.12.8/8.12.8) with ESMTP id i0CGCApI002889; Mon, 12 Jan 2004 18:12:10 +0200 Received: (from hannu@localhost) by fuji.krosing.net (8.12.8/8.12.8/Submit) id i0CGC9R0002887; Mon, 12 Jan 2004 18:12:09 +0200 X-Authentication-Warning: fuji.krosing.net: hannu set sender to hannu@tm.ee using -f Subject: Re: failures on machines using jfs From: Hannu Krosing To: "Spiegelberg, Greg" Cc: "'Christopher Browne '" , "'pgsql-performance@postgresql.org '" In-Reply-To: <387C22290D3FD71195D300508BF7DB5238AE92@colmail01.cranel.com> References: <387C22290D3FD71195D300508BF7DB5238AE92@colmail01.cranel.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit Message-Id: <1073923929.2365.22.camel@fuji.krosing.net> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Mon, 12 Jan 2004 18:12:09 +0200 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/124 X-Sequence-Number: 5299 Spiegelberg, Greg kirjutas P, 11.01.2004 kell 18:21: > It would seem we're experiencing somthing similiar with our scratch > volume (JFS mounted with noatime). Which files/directories do you keep on "scratch" volume ? All postgres files or just some (WAL, tmp) ? ------------- Hannu From pgsql-performance-owner@postgresql.org Mon Jan 12 13:04:26 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 43019D1B43F for ; Mon, 12 Jan 2004 17:04:26 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 05966-07 for ; Mon, 12 Jan 2004 13:03:55 -0400 (AST) Received: from COLSWEEPER.cranel.com (newmail.cranel.com [12.32.71.147]) by svr1.postgresql.org (Postfix) with ESMTP id DDC75D1B432 for ; Mon, 12 Jan 2004 13:03:53 -0400 (AST) Received: from colmail01.cranel.com (colmail01.cranel.com) by COLSWEEPER.cranel.com (Content Technologies SMTPRS 4.3.12) with ESMTP id ; Mon, 12 Jan 2004 12:02:43 -0500 Received: from cranel.com (gspiegelberg.cranel.com [192.168.11.134]) by colmail01.cranel.com with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2657.72) id CZTCJWT0; Mon, 12 Jan 2004 12:07:30 -0500 Message-ID: <4002D375.8040807@cranel.com> Date: Mon, 12 Jan 2004 12:03:49 -0500 From: Greg Spiegelberg Organization: Cranel, Inc. User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030624 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Hannu Krosing Cc: 'Christopher Browne ' , "'pgsql-performance@postgresql.org '" Subject: Re: failures on machines using jfs References: <387C22290D3FD71195D300508BF7DB5238AE92@colmail01.cranel.com> <1073923929.2365.22.camel@fuji.krosing.net> In-Reply-To: <1073923929.2365.22.camel@fuji.krosing.net> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/125 X-Sequence-Number: 5300 Hannu Krosing wrote: > Spiegelberg, Greg kirjutas P, 11.01.2004 kell 18:21: > >>It would seem we're experiencing somthing similiar with our scratch >>volume (JFS mounted with noatime). > > > Which files/directories do you keep on "scratch" volume ? > > All postgres files or just some (WAL, tmp) ? No Postgres files are kept in scratch only the files being loaded into the database via COPY or lo_import. My WAL logs are kept on a separate ext3 file system. Greg -- Greg Spiegelberg Sr. Product Development Engineer Cranel, Incorporated. Phone: 614.318.4314 Fax: 614.431.8388 Email: gspiegelberg@Cranel.com Cranel. Technology. Integrity. Focus. ********************************************************************** This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the system manager. This footnote also confirms that this email message has been swept by MIMEsweeper for the presence of computer viruses. www.mimesweeper.com ********************************************************************** From pgsql-performance-owner@postgresql.org Mon Jan 12 17:47:50 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id DD8DAD1B486 for ; Mon, 12 Jan 2004 21:47:49 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 70224-06 for ; Mon, 12 Jan 2004 17:47:20 -0400 (AST) Received: from bramble.mmrd.com (unknown [65.217.53.66]) by svr1.postgresql.org (Postfix) with ESMTP id E5AC5D1B47A for ; Mon, 12 Jan 2004 17:47:15 -0400 (AST) Received: from thorn.mmrd.com (thorn.mmrd.com [172.25.10.100]) by bramble.mmrd.com (8.12.8/8.12.8) with ESMTP id i0CL1ZcM031493; Mon, 12 Jan 2004 16:01:36 -0500 Received: from gnvex001.mmrd.com (gnvex001.mmrd.com [192.168.3.55]) by thorn.mmrd.com (8.11.6/8.11.6) with ESMTP id i0CLgCl26901; Mon, 12 Jan 2004 16:42:15 -0500 Received: from camel.mmrd.com ([172.25.5.213]) by gnvex001.mmrd.com with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2657.72) id XT870T8W; Mon, 12 Jan 2004 16:42:10 -0500 Subject: Re: COUNT & Pagination From: Robert Treat To: "david@shadovitz.com" Cc: "pgsql-performance@postgresql.org" , "'Richard Huxton'" In-Reply-To: <01C3D8DF.271378C0.david@shadovitz.com> References: <01C3D8DF.271378C0.david@shadovitz.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Ximian Evolution 1.0.8 Date: 12 Jan 2004 16:42:12 -0500 Message-Id: <1073943732.29178.153.camel@camel> Mime-Version: 1.0 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/126 X-Sequence-Number: 5301 On Mon, 2004-01-12 at 10:37, David Shadovitz wrote: > > If you only need the count when you've got the results, most PG client > > interfaces will tell you how many rows you've got. What language is your app > > in? > > PHP. > But I have only a subset of the results, retrieved via a query with a "LIMIT > " clause, so $pg_numrows is m. > And retrieving all results (i.e. no LIMIT) is at least as expensive as > COUNT(*). > Depending on frequency of updates and need for real time info, you could cache the count in session as long as the user stays within the given piece of your app. Robert Treat -- Build A Brighter Lamp :: Linux Apache {middleware} PostgreSQL From pgsql-admin-owner@postgresql.org Fri Jan 16 12:51:58 2004 X-Original-To: pgsql-admin-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 19D91D1B449 for ; Mon, 12 Jan 2004 21:46:15 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 70483-06 for ; Mon, 12 Jan 2004 17:45:46 -0400 (AST) Received: from web21106.mail.yahoo.com (web21106.mail.yahoo.com [216.136.227.108]) by svr1.postgresql.org (Postfix) with SMTP id E4EB4D1B462 for ; Mon, 12 Jan 2004 17:45:43 -0400 (AST) Message-ID: <20040112214545.35952.qmail@web21106.mail.yahoo.com> Received: from [65.206.239.222] by web21106.mail.yahoo.com via HTTP; Mon, 12 Jan 2004 13:45:45 PST Date: Mon, 12 Jan 2004 13:45:45 -0800 (PST) From: Shankar K Subject: Postgres on Netapp To: pgsql-admin@postgresql.org Cc: pgsql-performance@postgresql.org MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/141 X-Sequence-Number: 12004 Hi There, We are considering to use NetApp filer for a highly busy 24*7 postgres database and the reason we chose netapp, mostly being the "snapshot" functionality for backing up database online. The filer would be mounted on a rh linux server (7.3), 4g RAM, dual cpu with a dedicated card for filer. I'd appreciate if anyone could share your experience in configuring things on the filer for optimal performance or any recomendataion that i should be aware of. Thanks, Shankar __________________________________ Do you Yahoo!? Yahoo! Hotjobs: Enter the "Signing Bonus" Sweepstakes http://hotjobs.sweepstakes.yahoo.com/signingbonus From pgsql-performance-owner@postgresql.org Mon Jan 12 23:08:00 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 0865CD1B44A for ; Tue, 13 Jan 2004 03:08:00 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 27894-06 for ; Mon, 12 Jan 2004 23:07:32 -0400 (AST) Received: from mail2000-4.so-net.net.tw (mail2000-4.so-net.net.tw [61.64.127.20]) by svr1.postgresql.org (Postfix) with SMTP id 86419D1B52F for ; Mon, 12 Jan 2004 23:07:27 -0400 (AST) Received: from 10.1.9.27 by mail2000-4.so-net.net.tw with Mail2000 ESMTP Server V2.71S(79708:0:AUTH_RELAY) Tue, 13 Jan 2004 11:07:28 +0800 (CST); (envelope-from ) Received: By OpenMail Mailer;Tue, 13 Jan 2004 11:07:24 +0800 (CST) From: "cnliou" Reply-To: cnliou@so-net.net.tw Subject: Ignore Some Updates Message-ID: <1073963244.78843.cnliou@so-net.net.tw> To: "" Date: Tue, 13 Jan 2004 11:07:24 +0800 (CST) MIME-Version: 1.0 Content-Type: text/plain X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/127 X-Sequence-Number: 5302 Dear developers, I wonder it happens to systems where inefficient update SQL's are used like this: UPDATE MyTable SET MyColumn=1234 Question arises when the value of MyColumn is already 1234 before the update. If I am right, even when the to-be-updated column values equal to the new values, the core never hates to update that row anyway. If so, is it wise or not to adjust the core for lazy SQL users to ignore such "meaningless" updates in order to reduce some disk load and prevent some "holes" resulted from the delete (a consequence of update) in that table? Regards, CN From pgsql-performance-owner@postgresql.org Tue Jan 13 01:00:08 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 201F0D1B55D for ; Tue, 13 Jan 2004 05:00:08 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 44847-06 for ; Tue, 13 Jan 2004 00:59:36 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 192EDD1B494 for ; Tue, 13 Jan 2004 00:59:36 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i0D4xZ19008022; Mon, 12 Jan 2004 23:59:35 -0500 (EST) To: cnliou@so-net.net.tw Cc: "" Subject: Re: Ignore Some Updates In-reply-to: <1073963244.78843.cnliou@so-net.net.tw> References: <1073963244.78843.cnliou@so-net.net.tw> Comments: In-reply-to "cnliou" message dated "Tue, 13 Jan 2004 11:07:24 +0800" Date: Mon, 12 Jan 2004 23:59:35 -0500 Message-ID: <8021.1073969975@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/128 X-Sequence-Number: 5303 "cnliou" writes: > I wonder it happens to systems where inefficient update > SQL's are used like this: > UPDATE MyTable SET MyColumn=1234 > Question arises when the value of MyColumn is already 1234 > before the update. We have to fire UPDATE triggers in any case. > If I am right, even when the to-be-updated column values > equal to the new values, the core never hates to update that > row anyway. If so, is it wise or not to adjust the core for > lazy SQL users to ignore such "meaningless" updates in order Seems like penalizing the intelligent people (by adding useless comparisons) in order to reward the "lazy" ones. regards, tom lane From pgsql-performance-owner@postgresql.org Tue Jan 13 08:46:40 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1FED9D1D8A0 for ; Tue, 13 Jan 2004 12:46:39 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 19360-08 for ; Tue, 13 Jan 2004 08:46:11 -0400 (AST) Received: from fuji.krosing.net (silmet.estpak.ee [194.126.97.78]) by svr1.postgresql.org (Postfix) with ESMTP id 71C1DD1D889 for ; Tue, 13 Jan 2004 08:46:07 -0400 (AST) Received: from fuji.krosing.net (localhost.localdomain [127.0.0.1]) by fuji.krosing.net (8.12.8/8.12.8) with ESMTP id i0DCk8nj002638; Tue, 13 Jan 2004 14:46:08 +0200 Received: (from hannu@localhost) by fuji.krosing.net (8.12.8/8.12.8/Submit) id i0DCk8XV002636; Tue, 13 Jan 2004 14:46:08 +0200 X-Authentication-Warning: fuji.krosing.net: hannu set sender to hannu@tm.ee using -f Subject: Re: failures on machines using jfs From: Hannu Krosing To: Greg Spiegelberg Cc: "'Christopher Browne '" , "'pgsql-performance@postgresql.org '" In-Reply-To: <4002D375.8040807@cranel.com> References: <387C22290D3FD71195D300508BF7DB5238AE92@colmail01.cranel.com> <1073923929.2365.22.camel@fuji.krosing.net> <4002D375.8040807@cranel.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit Message-Id: <1073997967.2528.12.camel@fuji.krosing.net> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Tue, 13 Jan 2004 14:46:08 +0200 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/129 X-Sequence-Number: 5304 Greg Spiegelberg kirjutas E, 12.01.2004 kell 19:03: > Hannu Krosing wrote: > > Spiegelberg, Greg kirjutas P, 11.01.2004 kell 18:21: > > > >>It would seem we're experiencing somthing similiar with our scratch > >>volume (JFS mounted with noatime). > > > > > > Which files/directories do you keep on "scratch" volume ? > > > > All postgres files or just some (WAL, tmp) ? > > No Postgres files are kept in scratch only the files being loaded > into the database via COPY or lo_import. then the speedup does not make any sense ! Is reading from jfs filesystem also 5 times faster than reading from ext3 ? The only explanation I can give to filling database from jfs volume to be so much faster could be some strange filesystem cache interactions. ---------------- Hannu From pgsql-performance-owner@postgresql.org Tue Jan 13 08:59:38 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 2995ED1D876 for ; Tue, 13 Jan 2004 12:59:38 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 19445-08 for ; Tue, 13 Jan 2004 08:59:10 -0400 (AST) Received: from mta11.adelphia.net (mta11.adelphia.net [68.168.78.205]) by svr1.postgresql.org (Postfix) with ESMTP id B673AD1B440 for ; Tue, 13 Jan 2004 08:59:06 -0400 (AST) Received: from potentialtech.com ([68.68.113.33]) by mta11.adelphia.net (InterMail vM.5.01.06.05 201-253-122-130-105-20030824) with ESMTP id <20040113125909.CXQJ21134.mta11.adelphia.net@potentialtech.com>; Tue, 13 Jan 2004 07:59:09 -0500 Message-ID: <4003EB9C.30406@potentialtech.com> Date: Tue, 13 Jan 2004 07:59:08 -0500 From: Bill Moran User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.4) Gecko/20031005 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Hannu Krosing Cc: Greg Spiegelberg , 'Christopher Browne ' , "'pgsql-performance@postgresql.org '" Subject: Re: failures on machines using jfs References: <387C22290D3FD71195D300508BF7DB5238AE92@colmail01.cranel.com> <1073923929.2365.22.camel@fuji.krosing.net> <4002D375.8040807@cranel.com> <1073997967.2528.12.camel@fuji.krosing.net> In-Reply-To: <1073997967.2528.12.camel@fuji.krosing.net> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/130 X-Sequence-Number: 5305 Hannu Krosing wrote: > Greg Spiegelberg kirjutas E, 12.01.2004 kell 19:03: > >>Hannu Krosing wrote: >> >>>Spiegelberg, Greg kirjutas P, 11.01.2004 kell 18:21: >>> >>> >>>>It would seem we're experiencing somthing similiar with our scratch >>>>volume (JFS mounted with noatime). >>> >>> >>>Which files/directories do you keep on "scratch" volume ? >>> >>>All postgres files or just some (WAL, tmp) ? >> >>No Postgres files are kept in scratch only the files being loaded >>into the database via COPY or lo_import. > > > then the speedup does not make any sense ! > > Is reading from jfs filesystem also 5 times faster than reading from > ext3 ? > > The only explanation I can give to filling database from jfs volume to > be so much faster could be some strange filesystem cache interactions. http://www.potentialtech.com/wmoran/postgresql.php -- Bill Moran Potential Technologies http://www.potentialtech.com From pgsql-performance-owner@postgresql.org Tue Jan 13 11:04:54 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 8BA21D1B43F for ; Tue, 13 Jan 2004 15:04:53 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 54046-04 for ; Tue, 13 Jan 2004 11:04:26 -0400 (AST) Received: from COLSWEEPER.cranel.com (newmail.cranel.com [12.32.71.147]) by svr1.postgresql.org (Postfix) with ESMTP id 21EB4D1B519 for ; Tue, 13 Jan 2004 11:04:22 -0400 (AST) Received: from colmail01.cranel.com (colmail01.cranel.com) by COLSWEEPER.cranel.com (Content Technologies SMTPRS 4.3.12) with ESMTP id ; Tue, 13 Jan 2004 10:03:14 -0500 Received: from cranel.com (gspiegelberg.cranel.com [192.168.11.134]) by colmail01.cranel.com with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2657.72) id CZTCJ5KW; Tue, 13 Jan 2004 10:08:01 -0500 Message-ID: <400408F4.5030504@cranel.com> Date: Tue, 13 Jan 2004 10:04:20 -0500 From: Greg Spiegelberg Organization: Cranel, Inc. User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030624 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Hannu Krosing Cc: 'Christopher Browne ' , "'pgsql-performance@postgresql.org '" Subject: Re: failures on machines using jfs References: <387C22290D3FD71195D300508BF7DB5238AE92@colmail01.cranel.com> <1073923929.2365.22.camel@fuji.krosing.net> <4002D375.8040807@cranel.com> <1073997967.2528.12.camel@fuji.krosing.net> In-Reply-To: <1073997967.2528.12.camel@fuji.krosing.net> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/131 X-Sequence-Number: 5306 Hannu Krosing wrote: > Greg Spiegelberg kirjutas E, 12.01.2004 kell 19:03: > >>Hannu Krosing wrote: >> >>>Spiegelberg, Greg kirjutas P, 11.01.2004 kell 18:21: >>> >>> >>>>It would seem we're experiencing somthing similiar with our scratch >>>>volume (JFS mounted with noatime). >>> >>> >>>Which files/directories do you keep on "scratch" volume ? >>> >>>All postgres files or just some (WAL, tmp) ? >> >>No Postgres files are kept in scratch only the files being loaded >>into the database via COPY or lo_import. > > > then the speedup does not make any sense ! We do a lot of preprocessing before the data gets loaded. It's that process that experiences the hiccups I mentioned. -- Greg Spiegelberg Sr. Product Development Engineer Cranel, Incorporated. Phone: 614.318.4314 Fax: 614.431.8388 Email: gspiegelberg@Cranel.com Cranel. Technology. Integrity. Focus. ********************************************************************** This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the system manager. This footnote also confirms that this email message has been swept by MIMEsweeper for the presence of computer viruses. www.mimesweeper.com ********************************************************************** From pgsql-performance-owner@postgresql.org Tue Jan 13 12:05:16 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 3C943D1D8B9 for ; Tue, 13 Jan 2004 16:05:13 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 66791-08 for ; Tue, 13 Jan 2004 12:04:42 -0400 (AST) Received: from wind.mindcry.org (nat-gr.wmis.net [216.109.194.252]) by svr1.postgresql.org (Postfix) with ESMTP id 39926D1B469 for ; Tue, 13 Jan 2004 12:04:26 -0400 (AST) Received: by wind.mindcry.org (Postfix, from userid 1001) id 068E0417B; Tue, 13 Jan 2004 11:04:14 -0500 (EST) Date: Tue, 13 Jan 2004 11:04:13 -0500 From: David Hill To: pgsql-performance@postgresql.org Subject: freebsd 5.2 and max_connections Message-ID: <20040113160413.GB1554@phobia.ms> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.5.5.1i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/132 X-Sequence-Number: 5307 Hello - I am using postgresql to hold aliases, users, and relay_domains for postfix and courier to do lookups from. I am not storing mail in sql. I need postgresql to have fast read performance, so i setup index's on the tables. Also, the queries are basically "select blah from table where domain='domain.com'";, so i dont need to be able to support large results. I will have a lot of mail servers connecting to this postgresql db, so i need to support a lot of connections... but dont need to support large results. I am using FreeBSD 5.2. What are some tuning options and formulas I can use to get good values? Thanks! David From pgsql-performance-owner@postgresql.org Tue Jan 13 13:14:22 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1B3C3D1DAA8 for ; Tue, 13 Jan 2004 17:14:20 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 86424-07 for ; Tue, 13 Jan 2004 13:13:48 -0400 (AST) Received: from anchor-post-37.mail.demon.net (anchor-post-36.mail.demon.net [194.217.242.86]) by svr1.postgresql.org (Postfix) with ESMTP id C1776D1DB3F for ; Tue, 13 Jan 2004 13:13:47 -0400 (AST) Received: from mwynhau.demon.co.uk ([193.237.186.96] helo=mainbox.archonet.com) by anchor-post-37.mail.demon.net with esmtp (Exim 3.35 #1) id 1AgS6s-000IGD-0b; Tue, 13 Jan 2004 17:13:46 +0000 Received: by mainbox.archonet.com (Postfix, from userid 529) id BABE81717D; Tue, 13 Jan 2004 17:13:45 +0000 (GMT) Received: from client17.archonet.com (client17.archonet.com [192.168.1.17]) by mainbox.archonet.com (Postfix) with ESMTP id 4B88817177; Tue, 13 Jan 2004 17:13:43 +0000 (GMT) From: Richard Huxton To: David Hill , pgsql-performance@postgresql.org Subject: Re: freebsd 5.2 and max_connections Date: Tue, 13 Jan 2004 17:13:42 +0000 User-Agent: KMail/1.5 References: <20040113160413.GB1554@phobia.ms> In-Reply-To: <20040113160413.GB1554@phobia.ms> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200401131713.42682.dev@archonet.com> X-Bogosity: No, tests=bogofilter, spamicity=0.000000, version=0.15.3 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/133 X-Sequence-Number: 5308 On Tuesday 13 January 2004 16:04, David Hill wrote: > Hello - > I am using postgresql to hold aliases, users, and relay_domains for postfix > and courier to do lookups from. I am not storing mail in sql. > > I need postgresql to have fast read performance, so i setup index's on the > tables. Also, the queries are basically "select blah from table where > domain='domain.com'";, so i dont need to be able to support large results. > > I will have a lot of mail servers connecting to this postgresql db, so i > need to support a lot of connections... but dont need to support large > results. Firstly - if you don't know about the tuning guidelines/annotated config file, you should go here: http://www.varlena.com/varlena/GeneralBits/Tidbits/index.php Hmm - small result sets accessed directly via indexed fields, so sort_mem probably isn't important to you. Make sure your effective cache setting is accurate though, so PG can estimate whether it'll need to access the disks. Not sure if clustering one or more tables will help - I'm guessing not. What might help is to increase the statistics gathered on important columns. That should give the planner a more accurate estimate of value distribution and shouldn't cost you too much to keep accurate, since I'm guessing a low rate of updating. You might want to play with the random page cost (?or is it random access cost?) but more RAM for a bigger disk cache is probably the simplest tweak. -- Richard Huxton Archonet Ltd From pgsql-performance-owner@postgresql.org Tue Jan 13 13:33:09 2004 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 69CF4D1DB13 for ; Tue, 13 Jan 2004 17:33:08 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 92593-05 for ; Tue, 13 Jan 2004 13:32:37 -0400 (AST) Received: from trade-india.com (unknown [61.16.154.82]) by svr1.postgresql.org (Postfix) with SMTP id 94E43D1DF7E for ; Tue, 13 Jan 2004 13:32:33 -0400 (AST) Received: (qmail 12359 invoked from network); 13 Jan 2004 17:31:37 -0000 Received: from system66.trade-india-local.com (HELO office1.trade-india.com) (192.168.0.66) by system66.trade-india-local.com with SMTP; 13 Jan 2004 17:31:37 -0000 Received: from 192.168.0.100 (SquirrelMail authenticated user mallah) by system67.trade-india-local.com with HTTP; Tue, 13 Jan 2004 23:01:37 +0530 (IST) Message-ID: <1157.192.168.0.100.1074015097.squirrel@system67.trade-india-local.com> In-Reply-To: <01C3D82B.388A7FE0.david@shadovitz.com> References: <01C3D82B.388A7FE0.david@shadovitz.com> Date: Tue, 13 Jan 2004 23:01:37 +0530 (IST) Subject: Re: COUNT & Pagination From: mallah@trade-india.com To: "david@shadovitz.com" Cc: "pgsql-performance@postgresql.org" User-Agent: SquirrelMail/1.4.2 MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Priority: 3 Importance: Normal X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/134 X-Sequence-Number: 5309 > I understand that COUNT queries are expensive. So I'm looking for advice > on > displaying paginated query results. > > I display my query results like this: > > Displaying 1 to 50 of 2905. > 1-50 | 51-100 | 101-150 | etc. > > I do this by executing two queries. One is of the form: > > SELECT