On Tue, Jan 4, 2011 at 10:23 AM, sharadov <sreddy at spark.net> wrote:
>
>
>
> One of the slaves is running behind on replication. It is recommended that
> you restart slony when this happens. What are the steps one should take to
> go about restarting it?

There's restarting the slon daemons, there's dropping a node and
recreating it, and there's dropping all replication and starting over.
 In that order.  How to stop and start slon daemons depends on how you
installed the slon daemons.  We used the slony tools to do it.  The
init script basically kill -9s all of the slon daemons to stop it, and
starts each with a script to start them.

After that, if the node can't catch up you may need to drop and
recreate the node.  Pretty simple really, read the reference docs on
drop node command to get started.

Then, there's tearing down all replication on all machines in the
cluster and starting over.  That's probably more involved than you
want or need really.