Re: pg_xlog not cleaned up - Mailing list pgsql-general

From Tom Lane
Subject Re: pg_xlog not cleaned up
Date
Msg-id 23581.1154633306@sss.pgh.pa.us
Whole thread Raw
In response to Re: pg_xlog not cleaned up  (Wayne Conrad <wconrad@yagni.com>)
Responses Re: pg_xlog not cleaned up  (Christian Kastner <techmail@kvr.at>)
List pgsql-general
Wayne Conrad <wconrad@yagni.com> writes:
> On Mon, Jul 31, 2006 at 10:18:31PM +0200, Christian Kastner wrote:
>> After the online backup script runs, all subsequent attempts to run
>> archive_command fail because the first thing it tries to archive away is
>> the still-existing *.backup file. This fails because a copy already
>> exists in the archive area and the script refuses to overwrite the
>> existing file.

> I'm seeing this too.  I thought I was doing something wrong.

I'm not entirely convinced by Simon's theory that this is already fixed,
because there are some points that you didn't mention.

First off, the thing isn't *supposed* to remove the latest .backup file;
that one, and its .done file, should stick around until your next
successful pg_stop_backup call.  pg_stop_backup removes all but the
just-created .backup file (if they have .done files, showing they've
been archived).  No other operation touches them at all.

Second, the bug Simon mentions consists in trying to create the .ready
file for the new .backup file twice during pg_stop_backup.  If the
archiver is fast enough to have pushed out the .backup file already,
this would result in there being both a .ready and a .done file for the
.backup file.  If that's what you see, then manually removing the .ready
file should clear the problem --- I wouldn't recommend manually removing
either .done or .backup.  If that's *not* what you see, then we have
another problem to deal with.  Please report back.

            regards, tom lane

pgsql-general by date:

Previous
From: Tony Caduto
Date:
Subject: Well, Pervasive is now out....
Next
From: "Joshua D. Drake"
Date:
Subject: Re: Well, Pervasive is now out....