pg.ddx.io pgsql-performance@postgresql.org mailing list archivehelp / color / mirror / Atom feed
max_fsm_pages and check_points 2+ messages / 2 participants [nested] [flat]
* max_fsm_pages and check_points @ 2006-12-20 05:31 ALVARO ARCILA <arcila_alvaro@hotmail.com> 0 siblings, 1 reply; 2+ messages in thread From: ALVARO ARCILA @ 2006-12-20 05:31 UTC (permalink / raw) To: pgsql-performance HI, I've looking around the log files of my server and lately they indicate that I should consider increase the check_point segments because they're beeing reading too often and also recommend increasing the max_fsm_pages over 169728... those are the config values present in the postgresql.conf shared_buffers = 1000 # min 16 or max_connections*2, 8KB eachwork_mem = 8192 # min 64, size in KBmaintenance_work_mem = 262152 # min 1024, size in KBmax_fsm_pages = 40000 # min max_fsm_relations*16, 6 bytes eachmax_fsm_relations = 2000 # min 100, ~70 bytes each And the new values, acording to the HNIT in log files ... shared_buffers = 1000 # min 16 or max_connections*2, 8KB eachwork_mem = 8192 # min 64, size in KBmaintenance_work_mem = 262152 # min 1024, size in KBmax_fsm_pages = 170000 # min max_fsm_relations*16, 6 bytes eachmax_fsm_relations = 10625 # min 100, ~70 bytes each _________________________________________________________________ Consigue el nuevo Windows Live Messenger http://get.live.com/messenger/overview= ^ permalink raw reply [nested|flat] 2+ messages in thread
* Re: max_fsm_pages and check_points @ 2006-12-20 09:01 Ragnar <gnari@hive.is> parent: ALVARO ARCILA <arcila_alvaro@hotmail.com> 0 siblings, 0 replies; 2+ messages in thread From: Ragnar @ 2006-12-20 09:01 UTC (permalink / raw) To: ALVARO ARCILA <arcila_alvaro@hotmail.com>; +Cc: pgsql-performance On mið, 2006-12-20 at 05:31 +0000, ALVARO ARCILA wrote: > > HI, > > I've looking around the log files of my server and lately they > indicate that I should consider increase the check_point segments > because they're beeing reading too often and also recommend increasing > the max_fsm_pages over 169728... if this has been happening for some time, some tables might possibly have become bloated with dead rows, so a one-time VACUUM FULL or CLUSTER on these might be indicated to speed up reaching the steady state. I think the max_fsm_pages is a minimum recommendation, so you might want to look at VACUUM VERBOSE output after setting it, to see if an even higher value is indicated > those are the config values present in the postgresql.conf > > shared_buffers = 1000 > work_mem = 8192 if you have got a lot of memory, you might want to experiment with these a little gnari ^ permalink raw reply [nested|flat] 2+ messages in thread
end of thread, other threads:[~2006-12-20 09:01 UTC | newest] Thread overview: 2+ messages (download: mbox mbox.gz follow: Atom feed) -- links below jump to the message on this page -- 2006-12-20 05:31 max_fsm_pages and check_points ALVARO ARCILA <arcila_alvaro@hotmail.com> 2006-12-20 09:01 ` Ragnar <gnari@hive.is>
This inbox is served by DDX for PostgreSQL; see mirroring instructions for how to clone and mirror all data and code used for this inbox