Re: WAL archiving disaster

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Mike Atkin <mike(dot)atkin(at)gmail(dot)com>
Cc: pgsql-admin(at)postgresql(dot)org
Subject: Re: WAL archiving disaster
Date: 2011-04-13 00:40:38
Message-ID: 20370.1302655238@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin

Mike Atkin <mike(dot)atkin(at)gmail(dot)com> writes:
> So... normally this would be curtains for the database but the missing
> segment is from a 4 hour period of practically zero activity. It was
> only archived because it hit the archive_timeout. Is there any way I
> can force postgres to ignore this segment and attempt to redo the rest
> of the logs. I realise this may leave the database terminally
> inconsistent but my focus is on retrieving as much of the data as
> possible for re-entry rather than getting a functional system out the
> other end.

If you're willing to assume there were no updates in that interval,
you could try generating a dummy WAL segment having the correct page
headers and no contents except an XLOG_SWITCH record. I don't know
of any handy tools for that unfortunately.

regards, tom lane

In response to

Browse pgsql-admin by date

  From Date Subject
Next Message Donald Fraser 2011-04-13 09:57:05 Re: Missing documentation for error code: 80S01
Previous Message Kevin Grittner 2011-04-12 19:16:02 Re: Missing documentation for error code: 80S01