failing 8.1.4 autovacuum

Lists: pgsql-general
From: "Ed L(dot)" <pgsql(at)bluepolka(dot)net>
To: pgsql-general(at)postgresql(dot)org
Subject: failing 8.1.4 autovacuum
Date: 2006-10-23 22:03:25
Message-ID: 200610231603.25894.pgsql@bluepolka.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-general

I have an 8.1.4 autovac process running on HP-UX 11.23 IA64 and
repeatedly failing with the following error:

ERROR: failed to re-find parent key in "audit_idx1"

Will a reindex or drop index make this problem go away?

Is there anything I can do to help identify the underlying issue?
(This is an intensely used production system, so I have very
limited room to operate.)

TIA.

Ed


From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: "Ed L(dot)" <pgsql(at)bluepolka(dot)net>
Cc: pgsql-general(at)postgresql(dot)org
Subject: Re: failing 8.1.4 autovacuum
Date: 2006-10-23 22:58:15
Message-ID: 721.1161644295@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-general

"Ed L." <pgsql(at)bluepolka(dot)net> writes:
> I have an 8.1.4 autovac process running on HP-UX 11.23 IA64 and
> repeatedly failing with the following error:
> ERROR: failed to re-find parent key in "audit_idx1"

> Will a reindex or drop index make this problem go away?

Probably.

> Is there anything I can do to help identify the underlying issue?

Can you send me a copy of the physical index file?

regards, tom lane


From: "Ed L(dot)" <pgsql(at)bluepolka(dot)net>
To: pgsql-general(at)postgresql(dot)org
Subject: Re: failing 8.1.4 autovacuum
Date: 2006-10-24 00:51:06
Message-ID: 200610231851.06874.pgsql@bluepolka.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-general

On Monday October 23 2006 4:03 pm, Ed L. wrote:
> I have an 8.1.4 autovac process running on HP-UX 11.23 IA64
> and repeatedly failing with the following error:
>
>         ERROR:  failed to re-find parent key in "audit_idx1"
>
> Will a reindex or drop index make this problem go away?

BTW, it'd be nice if autovac didn't simply abort and restart with
the prior tables that were successfully vacuumed, but rather was
able to continue on after the error to the next table. As it
is, it starves the others but autovac just keeps banging it's
head against the wall, consuming lots of I/O and accomplishing
little. Maybe that's part of the 8.2 fixes I haven't kept up
with?

Ed