From: | Markus Schaber <schabi(at)logix-tt(dot)com> |
---|---|
To: | Brendan Duddridge <brendan(at)clickspace(dot)com> |
Cc: | PostgreSQL Performance <pgsql-performance(at)postgresql(dot)org> |
Subject: | Re: Recovery will take 10 hours |
Date: | 2006-04-24 08:29:44 |
Message-ID: | 444C8C78.1000906@logix-tt.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-performance |
Hi, Brandan,
Brendan Duddridge wrote:
> So how do you overlap the restore process with the retrieving of files?
You need a shell script as restore command that does both uncompressing
the current file, and starting a background decompress of the next
file(s). It also has to check whether the current file is already in
progress from a last run, and wait until this is finished instead of
decompressing it. Seems to be a little complicated than it sounds first.
> restore_command = 'gunzip </wal_archive/%f.gz>%p &'
Warning: Don't do it this way!
It will break things because PostgreSQL will try to access a
not-completely-restored wal file.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
From | Date | Subject | |
---|---|---|---|
Next Message | Paul Mackay | 2006-04-24 10:53:29 | Index on function less well cached than "regular" index ? |
Previous Message | Matteo Beccati | 2006-04-24 05:36:57 | Re: Inactive memory Grows unlimited |