Re: Continual Postgres headaches... - Mailing list pgsql-general

From Tom Lane
Subject Re: Continual Postgres headaches...
Date
Msg-id 5275.1196989256@sss.pgh.pa.us
Whole thread Raw
In response to Continual Postgres headaches...  ("Weber, Geoffrey M." <Geoffrey.Weber@mcleodusa.com>)
Responses Re: Continual Postgres headaches...
List pgsql-general
"Weber, Geoffrey M." <Geoffrey.Weber@mcleodusa.com> writes:
> My problems really are with performance consistency.  I have tweaked the
> execution so that everything should run with sub-second execution times, but
> even after everything is running well, I can get at most a week or two of
> steady running before things start to degrade.

Aside from the other responses: gradual degradation over a period of
time sounds like a bloat problem to me.  You should monitor table and
index sizes to confirm or deny that and find out just what's bloating.
Switching to autovacuum, or a better-designed manual vacuum schedule,
might be the ultimate answer, but first you need some concrete data
about what's going wrong.  Otherwise you're just shooting in the dark.

            regards, tom lane

pgsql-general by date:

Previous
From: Tom Lane
Date:
Subject: Re: Understanding how partial indexes work?
Next
From: Ow Mun Heng
Date:
Subject: Q: using generate_series to fill in the blanks