Re: Hot_standby WAL archiving question

On Mon, May 2, 2016 at 11:53 AM, Peter Brunnengräber <pbrunnen(at)bccglobal(dot)com
> wrote:

> Hello all,
> So I have been working on setting up an active/standby clustered
> postgresql 9.2 using corosync. The corosync documentation has me enable
> WAL archiving "archive_command = 'cp %p
> /db/data/postgresql/9.2/pg_archive/%f'"
>
> I assume so that the secondary can catch up in async (log shipping) mode
> before going into sync (streaming) mode. But I am noticing as I push data
> through the DB, that my pg_archive keeps growing and I am worried about
> running out of disk space.
>
> Per the postgresql documentation, I set "wal_keep_segments" but I
> believe this only effects the segments in the pg_xlog and not what the
> archive_command is processing.
>
> So, since I am not keeping the archived WAL statements for long term
> use, is it safe to set "archive_cleanup_command = 'pg_archivecleanup
> /db/data/postgresql/9.2/pg_archive %r'" on the master? Also will the
> archive cleanup command even have effect? I only read about it being on
> slaves to clean-up what has already been processed.
>
> Thank you!
>
> -With kind regards,
> Peter Brunnengräber
>
>
> --
> Sent via pgsql-admin mailing list (pgsql-admin(at)postgresql(dot)org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-admin
>

The archive_cleanup_command is something you set on the slave in the
recovery.conf, not on the master. It will run whatever command you place
there to remove WAL files that have been confirmed to have been replayed by
the slave and are no longer needed by the slave itself.

Keith