Re: Orphan TOAST object

From: "James Farrugia" <james(dot)farrugia(at)gmail(dot)com>
To: "Tom Lane" <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: pgsql-admin(at)postgresql(dot)org
Subject: Re: Orphan TOAST object
Date: 2008-05-12 07:25:34
Message-ID: 8d01be800805120025u3a49fa39rc7033a2480b4982f@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin

Hi Tom,

First of all thanks for the immediate replies!
Was actually waiting for the right moment to upgrade to 8.3 but migrating a
live 1Tb database is a bit daunting especially if you have never done it
before (as in my case). If I'm not mistaken i can upgrade to the latest
minor version without having to dump and restore so I'll do that.

One last thing...can we run into data-loss problems with successfully
vacuumed tables even if there is one unvacuumed database object; what would
have happened if I ignored to vacuum that rogue pg_toast (which was the only
unvacuumed object within the entire database)?

Thanks again.

James

On 5/11/08, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>
> "James Farrugia" <james(dot)farrugia(at)gmail(dot)com> writes:
> > I'm running 8.2.1.
>
> You really need to update to 8.2.latest. There are several known
> data-corruption problems in 8.2.1, and it seems possible that one of
> them ate the pg_depend row you needed.
>
> > I cleanly forgot about pg_depend!
> > Even after re-indexing I wasn't able to find an entry in pg_depend
> having
> > the TOAST's OID. I guess that by creating foo again and linking
> > pg_toast_xxx with foo in pg_depend by hand i can make it go away.
>
> Yeah, that's probably the cleanest recovery strategy.
>
> regards, tom lane
>

In response to

Responses

Browse pgsql-admin by date

  From Date Subject
Next Message Tino Schwarze 2008-05-12 07:35:46 Re: Orphan TOAST object
Previous Message Denison Wright 2008-05-11 18:12:18 pg_dump 8.3 changes from 8.2