pg.ddx.io pgsql-interfaces@postgresql.org mailing list archivehelp / color / mirror / Atom feed
[GENERAL/INTERFACES] Dynamically detecting need to vacuum? 2+ messages / 2 participants [nested] [flat]
* [GENERAL/INTERFACES] Dynamically detecting need to vacuum? @ 1999-12-13 18:37 Ed Loehr <ELOEHR@austin.rr.com> 1999-12-13 23:28 ` Re: [GENERAL] [GENERAL/INTERFACES] Dynamically detecting need to vacuum? Peter Eisentraut <peter_e@gmx.net> 0 siblings, 1 reply; 2+ messages in thread From: Ed Loehr @ 1999-12-13 18:37 UTC (permalink / raw) To: pgsql-general@postgresql.org; pgsql-interfaces I'm seeing a few intermittent show-stopper errors (some of which I've asked about previously in these forums) which go away after vacuum is run. Pgsql documentation recommends to vacuum nightly or after copying a large class (not sure exactly what that means) or deleting a large number of records. I'm tempted to run vaccum almost anytime I detect Pgsql errors (via perl DBI) in order to try to salvage/retry the query at hand on the fly at the cost of a much slower query from the user's perspective. I'd like to learn how to programmatically calculate/detect when a vacuuming is needed, and would like to hear any advice on how folks are successfully doing this or think would best be done. Any comments? [Hackers: it'd be nice to have pgsql optionally take care of this...] I'm using Pg 6.5.2 on RH Linux 6.0 2.2.5-15smp via perl DBI/DBD::Pg... Thanks in advance, Ed Loehr ^ permalink raw reply [nested|flat] 2+ messages in thread
* Re: [GENERAL] [GENERAL/INTERFACES] Dynamically detecting need to vacuum? 1999-12-13 18:37 [GENERAL/INTERFACES] Dynamically detecting need to vacuum? Ed Loehr <ELOEHR@austin.rr.com> @ 1999-12-13 23:28 ` Peter Eisentraut <peter_e@gmx.net> 0 siblings, 0 replies; 2+ messages in thread From: Peter Eisentraut @ 1999-12-13 23:28 UTC (permalink / raw) To: Ed Loehr <ELOEHR@austin.rr.com>; +Cc: pgsql-general@postgresql.org; pgsql-interfaces On 1999-12-13, Ed Loehr mentioned: > documentation recommends to vacuum nightly or after copying a large class > (not sure exactly what that means) or deleting a large number of records. copying -- as invoked via the COPY command (in general this means inserting, though) large -- probably anything with triple-digit row counts class -- table, relation; in general rows, records, data, stuff > [Hackers: it'd be nice to have pgsql optionally take care of this...] There are some disabled code fragments that take into account a "vacuum daemon". It's even on the TODO list, but vacuum itself needs to be sorted out first. Sorry, no free lunch here, probably. -- Peter Eisentraut Sernanders väg 10:115 peter_e@gmx.net 75262 Uppsala http://yi.org/peter-e/ Sweden ^ permalink raw reply [nested|flat] 2+ messages in thread
end of thread, other threads:[~1999-12-13 23:28 UTC | newest] Thread overview: 2+ messages (download: mbox mbox.gz follow: Atom feed) -- links below jump to the message on this page -- 1999-12-13 18:37 [GENERAL/INTERFACES] Dynamically detecting need to vacuum? Ed Loehr <ELOEHR@austin.rr.com> 1999-12-13 23:28 ` Peter Eisentraut <peter_e@gmx.net>
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