Re: vacuum error - Mailing list pgsql-general

From Ed L.
Subject Re: vacuum error
Date
Msg-id 200703061419.15855.pgsql@bluepolka.net
Whole thread Raw
In response to Re: vacuum error  (Peter Eisentraut <peter_e@gmx.net>)
Responses Re: vacuum error  ("Joshua D. Drake" <jd@commandprompt.com>)
List pgsql-general
On Tuesday March 6 2007 12:20 pm, Peter Eisentraut wrote:
> Ed L. wrote:
> > I am seeing the following error in pgsql 8.1.2:
> >
> > ERROR:  could not access status of transaction 3229475082
> > DETAIL:  could not open file "pg_clog/0C07": No such file or
> > directory
> >
> > What does it mean, and what should I do about it?
>
> 1. Read this thread:
> http://archives.postgresql.org/pgsql-general/2007-02/msg00820.
>php
>
> 2. Upgrade to the latest 8.1.* release.
>
> 3. If that doesn't help, check your system for faulty
> hardware, in particular for bad RAM.

This is a 200gb DB with ~300 transactions/second serving 5 busy
facilities, so downtime comes at a premium.  We have some
maintenance downtime planned for 2 weeks from now.  I'm trying
to understand if this can wait that long.

It appears the only failure occurs during an autovacuum-initiated
"VACUUM FREEZE" on template0 when it hits the pg_statistics
table.  However, that abort appears to be causing autovacuum to
skip all its other duties as it endlessly restarts and fails
again.

Do I care if template0 gets a "VACUUM FREEZE"?

Assuming not, is there a simple way to make autovacuum skip over
template0 so it can tend to the important data in the other
databases?

Is restarting with 8.1.8 a known solution for this problem?  Or
is an initdb required to fix it?

If initdb is required, we might as well move to the latest stable
8.2 version.  I understand my options to minimize downtime to be
limited to async replication.  Other ideas?

BTW, the RAM looks good.

TIA.
Ed

pgsql-general by date:

Previous
From: Tom Lane
Date:
Subject: Re: postgres slower on nested queries
Next
From: "Joshua D. Drake"
Date:
Subject: Re: vacuum error