Date: Fri, 27 Nov 2015 16:53:28 -0300
From: Alvaro Herrera <alvherre@2ndquadrant.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Kouber Saparev <kouber@gmail.com>, pgsql-admin@postgresql.org
Subject: Re: VACUUM unable to accomplish because of a non-existent MultiXactId
Message-ID: <20151127195328.GE4320@alvherre.pgsql>
References: <62FE8F79-7278-4AE0-9FDE-0C0277A7B609@gmail.com>
 <20151127184936.GD4320@alvherre.pgsql>
 <5403.1448652828@sss.pgh.pa.us>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <5403.1448652828@sss.pgh.pa.us>
User-Agent: Mutt/1.5.23 (2014-03-12)
Precedence: bulk
Sender: pgsql-admin-owner@postgresql.org

Tom Lane wrote:
> Alvaro Herrera <alvherre@2ndquadrant.com> writes:
> > Kouber Saparev wrote:
> >> Recently I spotted an auto vacuum that is constantly being run over a pg_toast table in the database. Interestingly the underlying table did not have that amount of writes that would trigger the auto vacuum every minute.
> >> 
> >> When I tried to run the VACUUM manually it died with a non-existent MultiXactId error:
> >> 
> >> db=# vacuum analyze verbose pg_toast.pg_toast_376621;
> >> INFO:  vacuuming "pg_toast.pg_toast_376621"
> >> ERROR:  MultiXactId 2915905228 does no longer exist -- apparent wraparound
> 
> > Hmm, I don't think there can be multixacts in toast tables at all,
> > normally.
> 
> Couldn't creation and deletion of a toasted object within the same
> transaction do it?

I tried and couldn't find a way to cause one to appear.  When a row is
created, it has xmax=0 so when it's deleted the xmax can just be set to
the deleting xact.  And we don't ever update pg_toast rows, AFAIK.

-- 
Álvaro Herrera                http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services


-- 
Sent via pgsql-admin mailing list (pgsql-admin@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-admin