Re: [PERFORM] Bad choice of query plan from PG 7.3.6 to PG 7.3.9
Sorry Tom, misread your mail! My bad :-( I believe the following is the data you need ? Live Server relname relpages ctp_statcon 72 statcon_pk 135 Test Server relname relpages ctp_statcon 34 statcon_pk 28 Have executed the following query to obtain that data: SELECT relname, relpages FROM pg_class WHERE relname = 'statcon_pk' OR relname = 'sc2ctp_fk' OR relname = 'sc2mtp_fk' OR relname = 'sc2sc_fk' OR relname = 'ctp_statcon' The size difference for the index is surprisingly big I think, considering that there's only around 1000 rows more in the table on the live server than on the server. Count for Live Server: 12597 Count for Test Server: 11494 Any insight into this? Cheers Jona PS: The meta data for the table is: CREATE TABLE statcon_tbl ( id serial NOT NULL, data bytea, wm bool DEFAULT 'FALSE', created timestamp DEFAULT now(), modified timestamp DEFAULT now(), enabled bool DEFAULT 'TRUE', bitsperpixel int4 DEFAULT 0, mtpid int4, sctid int4, ctpid int4, CONSTRAINT statcon_pk PRIMARY KEY (id), CONSTRAINT sc2ctp_fk FOREIGN KEY (ctpid) REFERENCES contype_tbl (id) ON UPDATE CASCADE ON DELETE CASCADE, CONSTRAINT sc2mtp_fk FOREIGN KEY (mtpid) REFERENCES mimetype_tbl (id) ON UPDATE CASCADE ON DELETE CASCADE, CONSTRAINT sc2sct_fk FOREIGN KEY (sctid) REFERENCES statcontrans_tbl (id) ON UPDATE CASCADE ON DELETE CASCADE ) WITHOUT OIDS; CREATE INDEX ctp_statcon ON statcon_tbl USING btree (sctid, ctpid); Tom Lane wrote: Jona [EMAIL PROTECTED] writes: anyway, here's the info for relpages: Live Server: 424 Test Server: 338 I was asking about the indexes associated with the table, not the table itself. regards, tom lane
Re: [PERFORM] Bad choice of query plan from PG 7.3.6 to PG 7.3.9
Wouldn't the VACUUM have made them equivalent?? anyway, here's the info for relpages: Live Server: 424 Test Server: 338 Please note though that there're more rows on the live server than on the test server due to recent upload. Total Row counts are as follows: Live Server: 12597 Test Server: 11494 When the problems started the tables had identical size though. Cheers Jona Tom Lane wrote: Jona [EMAIL PROTECTED] writes: Test Server: comm=# VACUUM ANALYZE VERBOSE StatCon_Tbl; INFO: --Relation public.statcon_tbl-- INFO: Pages 338: Changed 0, Empty 0; Tup 11494: Vac 0, Keep 0, UnUsed 0. Total CPU 0.02s/0.00u sec elapsed 1.98 sec. INFO: --Relation pg_toast.pg_toast_179851-- INFO: Pages 85680: Changed 0, Empty 0; Tup 343321: Vac 0, Keep 0, UnUsed 0. Total CPU 1.75s/0.23u sec elapsed 30.36 sec. INFO: Analyzing public.statcon_tbl VACUUM Live Server: comm=# VACUUM ANALYZE VERBOSE StatCon_Tbl; INFO: --Relation public.statcon_tbl-- INFO: Pages 424: Changed 0, Empty 0; Tup 12291: Vac 0, Keep 0, UnUsed 6101. Total CPU 0.00s/0.01u sec elapsed 0.01 sec. INFO: --Relation pg_toast.pg_toast_891830-- INFO: Pages 89234: Changed 0, Empty 0; Tup 352823: Vac 0, Keep 0, UnUsed 5487. Total CPU 3.21s/0.47u sec elapsed 18.03 sec. INFO: Analyzing public.statcon_tbl VACUUM Hm, the physical table sizes aren't very different, which suggests that the problem must lie with the indexes. Unfortunately, VACUUM in 7.3 doesn't tell you anything about indexes if it doesn't have any dead rows to clean up. Could you look at pg_class.relpages for all the indexes of this table, and see what that shows? regards, tom lane ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org
Re: [PERFORM] Bad choice of query plan from PG 7.3.6 to PG 7.3.9
Jona [EMAIL PROTECTED] writes: anyway, here's the info for relpages: Live Server: 424 Test Server: 338 I was asking about the indexes associated with the table, not the table itself. regards, tom lane ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send unregister YourEmailAddressHere to [EMAIL PROTECTED])
Re: [PERFORM] Bad choice of query plan from PG 7.3.6 to PG 7.3.9
Thank you for the swift reply. The test server is hardly ever vacuumed as it in general sees very limited traffic. vacuum is only necessary if the server sees a lot of write operations, i.e. update, delete, insert right? What explains the different choice of query plans then? As can be seen from the following snippets the test server decides to use an index twice in Query 2, where as the live server decides to do a full scan of tables with 38.5k and 5.5k records. In Query 3 it's vice versa. Seems strange to me... Query 2: --- Bad idea, price_tbl hold 38.5k records Test: - Index Scan using aff_price_uq on price_tbl (cost=0.00..6.01 rows=1 width=4) (actual time=0.01..0.01 rows=1 loops=2838)" Live: - Seq Scan on price_tbl (cost=0.00..883.48 rows=2434 width=4) (actual time=0.86..67.25 rows=4570 loops=1)" Filter: (affid = 8)" --- Bad idea, sct2subcattype_tbl hold 5.5k records Test: - Index Scan using subcat_uq on sct2subcattype_tbl (cost=0.00..79.26 rows=26 width=8) (actual time=0.01..0.17 rows=59 loops=48) Live: - Seq Scan on sct2subcattype_tbl (cost=0.00..99.26 rows=5526 width=8) (actual time=0.01..30.16 rows=5526 loops=1)" Query 3: - Bad idea, sct2lang_tbl has 8.6k records Test: - Seq Scan on sct2lang_tbl (cost=0.00..150.79 rows=8679 width=8) (actual time=0.03..10.70 rows=8679 loops=1)" Live: - Index Scan using sct2lang_uq on sct2lang_tbl (cost=0.00..8.13 rows=2 width=8) (actual time=1.10..2.39 rows=2 loops=69)" Will get a VACUUM VERBOSE of StatCon_Tbl Cheers Jona PS: The query plans are extracted using pgAdmin on Windows, if you can recommend a better cross-platform postgre client I'd be happy to try it out. Tom Lane wrote: Jona [EMAIL PROTECTED] writes: I'm currently experiencing problems with long query execution times. What I believe makes these problems particularly interesting is the difference in execution plans between our test server running PostGreSQL 7.3.6 and our production server running PostGreSQL 7.3.9. The test server is an upgraded "home machine", a Pentium 4 with 1GB of memory and IDE disk. The production server is a dual CPU XEON Pentium 4 with 2GB memory and SCSI disks. One should expect the production server to be faster, but appearently not as the outlined query plans below shows. I think the plans are fine; it looks to me like the production server has serious table-bloat or index-bloat problems, probably because of inadequate vacuuming. For instance compare these entries: - Index Scan using ctp_statcon on statcon_tbl (cost=0.00..6.01 rows=1 width=4) (actual time=0.05..0.31 rows=39 loops=4) Index Cond: ((statcon_tbl.sctid = "outer".sctid) AND (statcon_tbl.ctpid = 1)) - Index Scan using ctp_statcon on statcon_tbl (cost=0.00..20.40 rows=5 width=4) (actual time=27.97..171.84 rows=39 loops=4) Index Cond: ((statcon_tbl.sctid = "outer".sctid) AND (statcon_tbl.ctpid = 1)) Appears to be exactly the same task ... but the test server spent 1.24 msec total while the production server spent 687.36 msec total. That's more than half of your problem right there. Some of the other scans seem a lot slower on the production machine too. 1) How come the query plans between the 2 servers are different? The production server's rowcount estimates are pretty good, the test server's are not. How long since you vacuumed/analyzed the test server? It'd be interesting to see the output of "vacuum verbose statcon_tbl" on both servers ... regards, tom lane PS: if you post any more query plans, please try to use software that doesn't mangle the formatting so horribly ... ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [PERFORM] Bad choice of query plan from PG 7.3.6 to PG 7.3.9
Results of VACUUM VERBOSE from both servers Test server: comm=# VACUUM VERBOSE StatCon_Tbl; INFO: --Relation public.statcon_tbl-- INFO: Pages 338: Changed 338, Empty 0; Tup 11494: Vac 0, Keep 0, UnUsed 0. Total CPU 0.02s/0.00u sec elapsed 0.04 sec. INFO: --Relation pg_toast.pg_toast_179851-- INFO: Pages 85680: Changed 85680, Empty 0; Tup 343321: Vac 0, Keep 0, UnUsed 0. Total CPU 4.03s/0.40u sec elapsed 70.99 sec. VACUUM Live Server: comm=# VACUUM VERBOSE StatCon_Tbl; INFO: --Relation public.statcon_tbl-- INFO: Pages 424: Changed 0, Empty 0; Tup 12291: Vac 0, Keep 0, UnUsed 6101. Total CPU 0.01s/0.00u sec elapsed 0.60 sec. INFO: --Relation pg_toast.pg_toast_891830-- INFO: Pages 89234: Changed 0, Empty 0; Tup 352823: Vac 0, Keep 0, UnUsed 5487. Total CPU 4.44s/0.34u sec elapsed 35.48 sec. VACUUM Cheers Jona Tom Lane wrote: Jona [EMAIL PROTECTED] writes: I'm currently experiencing problems with long query execution times. What I believe makes these problems particularly interesting is the difference in execution plans between our test server running PostGreSQL 7.3.6 and our production server running PostGreSQL 7.3.9. The test server is an upgraded "home machine", a Pentium 4 with 1GB of memory and IDE disk. The production server is a dual CPU XEON Pentium 4 with 2GB memory and SCSI disks. One should expect the production server to be faster, but appearently not as the outlined query plans below shows. I think the plans are fine; it looks to me like the production server has serious table-bloat or index-bloat problems, probably because of inadequate vacuuming. For instance compare these entries: - Index Scan using ctp_statcon on statcon_tbl (cost=0.00..6.01 rows=1 width=4) (actual time=0.05..0.31 rows=39 loops=4) Index Cond: ((statcon_tbl.sctid = "outer".sctid) AND (statcon_tbl.ctpid = 1)) - Index Scan using ctp_statcon on statcon_tbl (cost=0.00..20.40 rows=5 width=4) (actual time=27.97..171.84 rows=39 loops=4) Index Cond: ((statcon_tbl.sctid = "outer".sctid) AND (statcon_tbl.ctpid = 1)) Appears to be exactly the same task ... but the test server spent 1.24 msec total while the production server spent 687.36 msec total. That's more than half of your problem right there. Some of the other scans seem a lot slower on the production machine too. 1) How come the query plans between the 2 servers are different? The production server's rowcount estimates are pretty good, the test server's are not. How long since you vacuumed/analyzed the test server? It'd be interesting to see the output of "vacuum verbose statcon_tbl" on both servers ... regards, tom lane PS: if you post any more query plans, please try to use software that doesn't mangle the formatting so horribly ... ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [PERFORM] Bad choice of query plan from PG 7.3.6 to PG 7.3.9
You didn't do analyze. Chris Jona wrote: Results of VACUUM VERBOSE from both servers Test server: comm=# VACUUM VERBOSE StatCon_Tbl; INFO: --Relation public.statcon_tbl-- INFO: Pages 338: Changed 338, Empty 0; Tup 11494: Vac 0, Keep 0, UnUsed 0. Total CPU 0.02s/0.00u sec elapsed 0.04 sec. INFO: --Relation pg_toast.pg_toast_179851-- INFO: Pages 85680: Changed 85680, Empty 0; Tup 343321: Vac 0, Keep 0, UnUsed 0. Total CPU 4.03s/0.40u sec elapsed 70.99 sec. VACUUM Live Server: comm=# VACUUM VERBOSE StatCon_Tbl; INFO: --Relation public.statcon_tbl-- INFO: Pages 424: Changed 0, Empty 0; Tup 12291: Vac 0, Keep 0, UnUsed 6101. Total CPU 0.01s/0.00u sec elapsed 0.60 sec. INFO: --Relation pg_toast.pg_toast_891830-- INFO: Pages 89234: Changed 0, Empty 0; Tup 352823: Vac 0, Keep 0, UnUsed 5487. Total CPU 4.44s/0.34u sec elapsed 35.48 sec. VACUUM Cheers Jona Tom Lane wrote: Jona [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] writes: I'm currently experiencing problems with long query execution times. What I believe makes these problems particularly interesting is the difference in execution plans between our test server running PostGreSQL 7.3.6 and our production server running PostGreSQL 7.3.9. The test server is an upgraded home machine, a Pentium 4 with 1GB of memory and IDE disk. The production server is a dual CPU XEON Pentium 4 with 2GB memory and SCSI disks. One should expect the production server to be faster, but appearently not as the outlined query plans below shows. I think the plans are fine; it looks to me like the production server has serious table-bloat or index-bloat problems, probably because of inadequate vacuuming. For instance compare these entries: - Index Scan using ctp_statcon on statcon_tbl (cost=0.00..6.01 rows=1 width=4) (actual time=0.05..0.31 rows=39 loops=4) Index Cond: ((statcon_tbl.sctid = outer.sctid) AND (statcon_tbl.ctpid = 1)) - Index Scan using ctp_statcon on statcon_tbl (cost=0.00..20.40 rows=5 width=4) (actual time=27.97..171.84 rows=39 loops=4) Index Cond: ((statcon_tbl.sctid = outer.sctid) AND (statcon_tbl.ctpid = 1)) Appears to be exactly the same task ... but the test server spent 1.24 msec total while the production server spent 687.36 msec total. That's more than half of your problem right there. Some of the other scans seem a lot slower on the production machine too. 1) How come the query plans between the 2 servers are different? The production server's rowcount estimates are pretty good, the test server's are not. How long since you vacuumed/analyzed the test server? It'd be interesting to see the output of vacuum verbose statcon_tbl on both servers ... regards, tom lane PS: if you post any more query plans, please try to use software that doesn't mangle the formatting so horribly ... ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org
Re: [PERFORM] Bad choice of query plan from PG 7.3.6 to PG 7.3.9
Now with analyze Test Server: comm=# VACUUM ANALYZE VERBOSE StatCon_Tbl; INFO: --Relation public.statcon_tbl-- INFO: Pages 338: Changed 0, Empty 0; Tup 11494: Vac 0, Keep 0, UnUsed 0. Total CPU 0.02s/0.00u sec elapsed 1.98 sec. INFO: --Relation pg_toast.pg_toast_179851-- INFO: Pages 85680: Changed 0, Empty 0; Tup 343321: Vac 0, Keep 0, UnUsed 0. Total CPU 1.75s/0.23u sec elapsed 30.36 sec. INFO: Analyzing public.statcon_tbl VACUUM Live Server: comm=# VACUUM ANALYZE VERBOSE StatCon_Tbl; INFO: --Relation public.statcon_tbl-- INFO: Pages 424: Changed 0, Empty 0; Tup 12291: Vac 0, Keep 0, UnUsed 6101. Total CPU 0.00s/0.01u sec elapsed 0.01 sec. INFO: --Relation pg_toast.pg_toast_891830-- INFO: Pages 89234: Changed 0, Empty 0; Tup 352823: Vac 0, Keep 0, UnUsed 5487. Total CPU 3.21s/0.47u sec elapsed 18.03 sec. INFO: Analyzing public.statcon_tbl VACUUM Have done some sampling running the same query a few times through the past few hours and it appears that the VACUUM has helped. The following are the results after the Vacuum: After VACUUM VERBOSE: Index Scan using ctp_statcon on statcon_tbl (cost=0.00..21.29 rows=5 width=4) (actual time=0.07..0.37 rows=39 loops=4) Index Cond: ((statcon_tbl.sctid = outer.sctid) AND (statcon_tbl.ctpid = 1)) After VACUUM ANALYZE VERBOSE: Index Scan using ctp_statcon on statcon_tbl (cost=0.00..20.03 rows=5 width=4) (actual time=0.09..0.37 rows=39 loops=4) Index Cond: ((statcon_tbl.sctid = outer.sctid) AND (statcon_tbl.ctpid = 1)) Only question remains why one server uses its indexes and the other don't eventhough VACUUM ANALYZE has now been run on both servers? And even more interesting, before the VACUUM ANALYZEit was the server where no vacuum had taken place that used its index. Cheers Jona Christopher Kings-Lynne wrote: You didn't do analyze. Chris Jona wrote: Results of VACUUM VERBOSE from both servers Test server: comm=# VACUUM VERBOSE StatCon_Tbl; INFO: --Relation public.statcon_tbl-- INFO: Pages 338: Changed 338, Empty 0; Tup 11494: Vac 0, Keep 0, UnUsed 0. Total CPU 0.02s/0.00u sec elapsed 0.04 sec. INFO: --Relation pg_toast.pg_toast_179851-- INFO: Pages 85680: Changed 85680, Empty 0; Tup 343321: Vac 0, Keep 0, UnUsed 0. Total CPU 4.03s/0.40u sec elapsed 70.99 sec. VACUUM Live Server: comm=# VACUUM VERBOSE StatCon_Tbl; INFO: --Relation public.statcon_tbl-- INFO: Pages 424: Changed 0, Empty 0; Tup 12291: Vac 0, Keep 0, UnUsed 6101. Total CPU 0.01s/0.00u sec elapsed 0.60 sec. INFO: --Relation pg_toast.pg_toast_891830-- INFO: Pages 89234: Changed 0, Empty 0; Tup 352823: Vac 0, Keep 0, UnUsed 5487. Total CPU 4.44s/0.34u sec elapsed 35.48 sec. VACUUM Cheers Jona Tom Lane wrote: Jona [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] writes: I'm currently experiencing problems with long query execution times. What I believe makes these problems particularly interesting is the difference in execution plans between our test server running PostGreSQL 7.3.6 and our production server running PostGreSQL 7.3.9. The test server is an upgraded home machine, a Pentium 4 with 1GB of memory and IDE disk. The production server is a dual CPU XEON Pentium 4 with 2GB memory and SCSI disks. One should expect the production server to be faster, but appearently not as the outlined query plans below shows. I think the plans are fine; it looks to me like the production server has serious table-bloat or index-bloat problems, probably because of inadequate vacuuming. For instance compare these entries: - Index Scan using ctp_statcon on statcon_tbl (cost=0.00..6.01 rows=1 width=4) (actual time=0.05..0.31 rows=39 loops=4) Index Cond: ((statcon_tbl.sctid = outer.sctid) AND (statcon_tbl.ctpid = 1)) - Index Scan using ctp_statcon on statcon_tbl (cost=0.00..20.40 rows=5 width=4) (actual time=27.97..171.84 rows=39 loops=4) Index Cond: ((statcon_tbl.sctid = outer.sctid) AND (statcon_tbl.ctpid = 1)) Appears to be exactly the same task ... but the test server spent 1.24 msec total while the production server spent 687.36 msec total. That's more than half of your problem right there. Some of the other scans seem a lot slower on the production machine too. 1) How come the query plans between the 2 servers are different? The production server's rowcount estimates are pretty good, the test server's are not. How long since you vacuumed/analyzed the test server? It'd be interesting to see the output of vacuum verbose statcon_tbl on both servers ... regards, tom lane PS: if you post any more query plans, please try to use software that doesn't mangle the formatting so horribly ... ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org
[PERFORM] Bad choice of query plan from PG 7.3.6 to PG 7.3.9 part 1
Hi I'm currently experiencing problems with long query execution times. What I believe makes these problems particularly interesting is the difference in execution plans between our test server running PostGreSQL 7.3.6 and our production server running PostGreSQL 7.3.9. The test server is an upgraded home machine, a Pentium 4 with 1GB of memory and IDE disk. The production server is a dual CPU XEON Pentium 4 with 2GB memory and SCSI disks. One should expect the production server to be faster, but appearently not as the outlined query plans below shows. My questions can be summoned up to: 1) How come the query plans between the 2 servers are different? 2) How come the production server in general estimates the cost of the query plans so horribly wrong? (ie. it chooses a bad query plan where as the test server chooses a good plan) 3) In Query 2, how come the production server refuses the use its indexes (subcat_uq and aff_price_uq, both unique indexes) where as the test server determines that the indexes are the way to go 4) In Query 3, how come the test server refuses to use its index (sct2lang_uq) and the production server uses it? And why is the test server still faster eventhough it makes a sequential scan of a table with 8.5k records in? Please note, a VACUUM ANALYSE is run on the production server once a day (used to be once an hour but it seemed to make no difference), however there are generally no writes to the tables used in the queries. If anyone could shed some light on these issues I would truly appreciate it. Cheers Jona PS. Please refer to part 2 for the other queries and query plans Query 1: EXPLAIN ANALYZE SELECT DISTINCT StatConTrans_Tbl.id, Code_Tbl.sysnm AS code, PriceCat_Tbl.amount AS price, Country_Tbl.currency, CreditsCat_Tbl.amount AS credits, Info_Tbl.title, Info_Tbl.description FROM (SCT2SubCatType_Tbl INNER JOIN SCT2Lang_Tbl ON SCT2SubCatType_Tbl.sctid = SCT2Lang_Tbl.sctid INNER JOIN Language_Tbl ON SCT2Lang_Tbl.langid = Language_Tbl.id AND Language_Tbl.sysnm = UPPER('us') AND Language_Tbl.enabled = true INNER JOIN Info_Tbl ON SCT2SubCatType_Tbl.sctid = Info_Tbl.sctid AND Language_Tbl.id = Info_Tbl.langid INNER JOIN SubCatType_Tbl ON SCT2SubCatType_Tbl.subcattpid = SubCatType_Tbl.id AND SubCatType_Tbl.enabled = true INNER JOIN CatType_Tbl ON SubCatType_Tbl.cattpid = CatType_Tbl.id AND CatType_Tbl.enabled = true INNER JOIN SuperCatType_Tbl ON CatType_Tbl.spcattpid = SuperCatType_Tbl.id AND SuperCatType_Tbl.enabled = true INNER JOIN StatConTrans_Tbl ON SCT2SubCatType_Tbl.sctid = StatConTrans_Tbl.id AND StatConTrans_Tbl.enabled = true INNER JOIN Price_Tbl ON StatConTrans_Tbl.id = Price_Tbl.sctid AND Price_Tbl.affid = 8 INNER JOIN PriceCat_Tbl ON Price_Tbl.prccatid = PriceCat_Tbl.id AND PriceCat_Tbl.enabled = true INNER JOIN Country_Tbl ON PriceCat_Tbl.cntid = Country_Tbl.id AND Country_Tbl.enabled = true INNER JOIN CreditsCat_Tbl ON Price_Tbl.crdcatid = CreditsCat_Tbl.id AND CreditsCat_Tbl.enabled = true INNER JOIN StatCon_Tbl ON StatConTrans_Tbl.id = StatCon_Tbl.sctid AND StatCon_Tbl.ctpid = 1 INNER JOIN Code_Tbl ON SuperCatType_Tbl.id = Code_Tbl.spcattpid AND Code_Tbl.affid = 8 AND Code_Tbl.cdtpid = 1) WHERE SCT2SubCatType_Tbl.subcattpid = 79 ORDER BY StatConTrans_Tbl.id DESC LIMIT 8 OFFSET 0 Plan on PostGre 7.3.6 on Red Hat Linux 3.2.3-39 Limit (cost=178.59..178.61 rows=1 width=330) (actual time=22.77..28.51 rows=4 loops=1) - Unique (cost=178.59..178.61 rows=1 width=330) (actual time=22.77..28.50 rows=4 loops=1) - Sort (cost=178.59..178.60 rows=1 width=330) (actual time=22.76..22.85 rows=156 loops=1) Sort Key: statcontrans_tbl.id, code_tbl.sysnm, pricecat_tbl.amount, country_tbl.currency, creditscat_tbl.amount, info_tbl.title, info_tbl.description - Hash Join (cost=171.19..178.58 rows=1 width=330) (actual time=3.39..6.55 rows=156 loops=1) Hash Cond: (outer.cntid = inner.id) - Nested Loop (cost=170.13..177.51 rows=1 width=312) (actual time=3.27..5.75 rows=156 loops=1) Join Filter: (inner.sctid = outer.sctid) - Hash Join (cost=170.13..171.48 rows=1 width=308) (actual time=3.12..3.26 rows=4 loops=1) Hash Cond: (outer.crdcatid = inner.id) - Hash Join (cost=169.03..170.38 rows=1 width=300) (actual time=3.00..3.11 rows=4 loops=1) Hash Cond: (outer.spcattpid = inner.spcattpid) - Hash Join (cost=167.22..168.56 rows=1 width=253) (actual time=2.88..2.97 rows=4 loops=1) Hash Cond: (outer.id = inner.prccatid) - Seq Scan on pricecat_tbl (cost=0.00..1.29 rows=12 width=12)
Re: [PERFORM] Bad choice of query plan from PG 7.3.6 to PG 7.3.9 part 1
Jona [EMAIL PROTECTED] writes: I'm currently experiencing problems with long query execution times. What I believe makes these problems particularly interesting is the difference in execution plans between our test server running PostGreSQL 7.3.6 and our production server running PostGreSQL 7.3.9. The test server is an upgraded home machine, a Pentium 4 with 1GB of memory and IDE disk. The production server is a dual CPU XEON Pentium 4 with 2GB memory and SCSI disks. One should expect the production server to be faster, but appearently not as the outlined query plans below shows. I think the plans are fine; it looks to me like the production server has serious table-bloat or index-bloat problems, probably because of inadequate vacuuming. For instance compare these entries: - Index Scan using ctp_statcon on statcon_tbl (cost=0.00..6.01 rows=1 width=4) (actual time=0.05..0.31 rows=39 loops=4) Index Cond: ((statcon_tbl.sctid = outer.sctid) AND (statcon_tbl.ctpid = 1)) - Index Scan using ctp_statcon on statcon_tbl (cost=0.00..20.40 rows=5 width=4) (actual time=27.97..171.84 rows=39 loops=4) Index Cond: ((statcon_tbl.sctid = outer.sctid) AND (statcon_tbl.ctpid = 1)) Appears to be exactly the same task ... but the test server spent 1.24 msec total while the production server spent 687.36 msec total. That's more than half of your problem right there. Some of the other scans seem a lot slower on the production machine too. 1) How come the query plans between the 2 servers are different? The production server's rowcount estimates are pretty good, the test server's are not. How long since you vacuumed/analyzed the test server? It'd be interesting to see the output of vacuum verbose statcon_tbl on both servers ... regards, tom lane PS: if you post any more query plans, please try to use software that doesn't mangle the formatting so horribly ... ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
[PERFORM] Bad choice of query plan from PG 7.3.6 to PG 7.3.9 part 1b
Please refer to part 1a for questions and part 2 for more queries and query plans. Why won't this list accept my questions and sample data in one mail??? /Jona Query 1: EXPLAIN ANALYZE SELECT DISTINCT StatConTrans_Tbl.id, Code_Tbl.sysnm AS code, PriceCat_Tbl.amount AS price, Country_Tbl.currency, CreditsCat_Tbl.amount AS credits, Info_Tbl.title, Info_Tbl.description FROM (SCT2SubCatType_Tbl INNER JOIN SCT2Lang_Tbl ON SCT2SubCatType_Tbl.sctid = SCT2Lang_Tbl.sctid INNER JOIN Language_Tbl ON SCT2Lang_Tbl.langid = Language_Tbl.id AND Language_Tbl.sysnm = UPPER('us') AND Language_Tbl.enabled = true INNER JOIN Info_Tbl ON SCT2SubCatType_Tbl.sctid = Info_Tbl.sctid AND Language_Tbl.id = Info_Tbl.langid INNER JOIN SubCatType_Tbl ON SCT2SubCatType_Tbl.subcattpid = SubCatType_Tbl.id AND SubCatType_Tbl.enabled = true INNER JOIN CatType_Tbl ON SubCatType_Tbl.cattpid = CatType_Tbl.id AND CatType_Tbl.enabled = true INNER JOIN SuperCatType_Tbl ON CatType_Tbl.spcattpid = SuperCatType_Tbl.id AND SuperCatType_Tbl.enabled = true INNER JOIN StatConTrans_Tbl ON SCT2SubCatType_Tbl.sctid = StatConTrans_Tbl.id AND StatConTrans_Tbl.enabled = true INNER JOIN Price_Tbl ON StatConTrans_Tbl.id = Price_Tbl.sctid AND Price_Tbl.affid = 8 INNER JOIN PriceCat_Tbl ON Price_Tbl.prccatid = PriceCat_Tbl.id AND PriceCat_Tbl.enabled = true INNER JOIN Country_Tbl ON PriceCat_Tbl.cntid = Country_Tbl.id AND Country_Tbl.enabled = true INNER JOIN CreditsCat_Tbl ON Price_Tbl.crdcatid = CreditsCat_Tbl.id AND CreditsCat_Tbl.enabled = true INNER JOIN StatCon_Tbl ON StatConTrans_Tbl.id = StatCon_Tbl.sctid AND StatCon_Tbl.ctpid = 1 INNER JOIN Code_Tbl ON SuperCatType_Tbl.id = Code_Tbl.spcattpid AND Code_Tbl.affid = 8 AND Code_Tbl.cdtpid = 1) WHERE SCT2SubCatType_Tbl.subcattpid = 79 ORDER BY StatConTrans_Tbl.id DESC LIMIT 8 OFFSET 0 Plan on PostGre 7.3.6 on Red Hat Linux 3.2.3-39 Limit (cost=178.59..178.61 rows=1 width=330) (actual time=22.77..28.51 rows=4 loops=1) - Unique (cost=178.59..178.61 rows=1 width=330) (actual time=22.77..28.50 rows=4 loops=1) - Sort (cost=178.59..178.60 rows=1 width=330) (actual time=22.76..22.85 rows=156 loops=1) Sort Key: statcontrans_tbl.id, code_tbl.sysnm, pricecat_tbl.amount, country_tbl.currency, creditscat_tbl.amount, info_tbl.title, info_tbl.description - Hash Join (cost=171.19..178.58 rows=1 width=330) (actual time=3.39..6.55 rows=156 loops=1) Hash Cond: (outer.cntid = inner.id) - Nested Loop (cost=170.13..177.51 rows=1 width=312) (actual time=3.27..5.75 rows=156 loops=1) Join Filter: (inner.sctid = outer.sctid) - Hash Join (cost=170.13..171.48 rows=1 width=308) (actual time=3.12..3.26 rows=4 loops=1) Hash Cond: (outer.crdcatid = inner.id) - Hash Join (cost=169.03..170.38 rows=1 width=300) (actual time=3.00..3.11 rows=4 loops=1) Hash Cond: (outer.spcattpid = inner.spcattpid) - Hash Join (cost=167.22..168.56 rows=1 width=253) (actual time=2.88..2.97 rows=4 loops=1) Hash Cond: (outer.id = inner.prccatid) - Seq Scan on pricecat_tbl (cost=0.00..1.29 rows=12 width=12) (actual time=0.04..0.08 rows=23 loops=1) Filter: (enabled = true) - Hash (cost=167.21..167.21 rows=1 width=241) (actual time=2.80..2.80 rows=0 loops=1) - Nested Loop (cost=3.77..167.21 rows=1 width=241) (actual time=1.31..2.79 rows=4 loops=1) Join Filter: (inner.sctid = outer.sctid) - Nested Loop (cost=3.77..161.19 rows=1 width=229) (actual time=1.19..2.60 rows=4 loops=1) Join Filter: (outer.sctid = inner.sctid) - Hash Join (cost=3.77..155.17 rows=1 width=44) (actual time=1.07..2.37 rows=4 loops=1) Hash Cond: (outer.langid = inner.id) - Nested Loop (cost=2.69..154.06 rows=7 width=40) (actual time=0.90..2.18 rows=8 loops=1) Join Filter: (outer.sctid = inner.sctid) - Nested Loop (cost=2.69..21.30 rows=1 width=32) (actual time=0.78..1.94 rows=4 loops=1)