To: Deron <fecastle@gmail.com>
cc: pgsql-admin@postgresql.org
Subject: Re: cache lookup for operator ... 
In-reply-to: <AANLkTinI7mYaqvKVDwTNuHjhfH9UqvGxuGKSuynbpZRY@mail.gmail.com> 
References: <AANLkTinI7mYaqvKVDwTNuHjhfH9UqvGxuGKSuynbpZRY@mail.gmail.com>
Comments: In-reply-to Deron <fecastle@gmail.com>
	message dated "Tue, 13 Jul 2010 10:29:42 -0700"
Date: Tue, 13 Jul 2010 14:22:22 -0400
Message-ID: <4029.1279045342@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>

Deron <fecastle@gmail.com> writes:
> I am trying to determine what might cause an issue like this.   This was a
> large (several hundred GB) data warehouse server with PG 8.2.4.   I can't
> really determine what caused this issue.  We ended up restoring the data
> from a backup.

> Here are some of the errors in the log:

> postmaster.1:Jul 11 14:13:36  postgres[18051]: [8-1] ERROR: cache lookup
> failed for operator 660

> postmaster.1:Jul 11 15:27:00  postgres[26428]: [2-1] ERROR: cache lookup
> failed for operator 93

Evidently you had corruption in the pg_operator catalog.  If you were
lucky, a REINDEX of pg_operator might have fixed it, but unless you kept
a physical copy of the corrupted database there's no way to investigate
further now.

8.2.4 is extremely out of date btw, you should be using 8.2.17 or
something close to that.

			regards, tom lane