Le vendredi 30 novembre 2012 à 09:44 -0500, François Beausoleil a
écrit :
> I'm using 9.1.5 on Ubuntu 11.10, in a streaming replication scenario. On my slave, recovery.conf states:
>
> standby_mode = on
> restore_command = '/usr/local/omnipitr/bin/omnipitr-restore -D /var/lib/postgresql/9.1/main/ --source
gzip=/data/dbanalytics-wal/--remove-unneeded --temp-dir /var/tmp/omnipitr -l /var/log/omnipitr/restore-^Y-^m-^d.log
--streaming-replication--verbose --error-pgcontroldata hang "%f" "%p"'
> archive_cleanup_command = '/usr/local/omnipitr/bin/omnipitr-cleanup --verbose --log
/var/log/omnipitr/cleanup-^Y-^m-^d.log--archive gzip=/data/dbanalytics-wal/ "%r"'
> primary_conninfo = 'host=master port=5432 user=dbrepl password=password'
>
>
> I ran out of disk space on the slave, because the archived WAL records were not removed. The documentation for
archive_cleanup_commandstates[1]:
>
Hi,
I have no personal experience on the matter, but saw this recent post :
http://archives.postgresql.org/pgsql-general/2012-12/msg00129.php
which seems related?
--
Vincent Veyron
http://marica.fr
Logiciel pour département juridique