Date: Tue, 28 Aug 2018 15:34:50 +0200 (CEST)
From: 066ce286@free.fr
To: wambacher@posteo.de
Cc: pgsql-admin@lists.postgresql.org
Message-ID: 
 <198952269.367235503.1535463290956.JavaMail.root@zimbra82-e14.priv.proxad.net>
In-Reply-To: <2be8ebca-fecc-51f4-39d7-9869941c679a@posteo.de>
Subject: Re: tuple concurrently updated
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

I don't think a vacuum can handle what a pg_dump can't. But I'm maybe wrong=
.

----- Mail original -----
De: wambacher@posteo.de
=C3=80: pgsql-admin@lists.postgresql.org
Envoy=C3=A9: Mardi 28 Ao=C3=BBt 2018 15:19:28
Objet: Re: tuple concurrently updated

because it's an OpenStreetMap Full planet database, which is huge (~ 2
TB), i don't have many backups. the last is from 20180608

ok, i can restore that and reload the incremential updates or reload
full data starting from raw data. both will need some days.

if i could get rid of the bad data-record, it's much easier to recover that=
.

Running a vacuum or a vacuum full, is my last try.

Regards
walter


Am 28.08.2018 um 15:10 schrieb 066ce286@free.fr:
> BTW you have data corruption.
>
> So forget your problem about concurrent updates ; that's not the clue.
>
> Your new question should be how to recover corrupted table.
>
> Sorry, I've no skill for that problem other that "how ran your backups/ar=
chivelogs ?"
>