The mailing list for listmasters using Sympa

[Sylvain Amrani <address@concealed>]

Johann Spies wrote:
> I still have to find the correct tool to replicate the
> postgresql-tables.
>   
PGCluster is your friend.
> How do you configure the setup on the second server to disallow the
> administration of the lists?  Is removing the task_manager daemon
> enough?
>   
I only use SSO in auth.conf and web_tt2 are customized this way. SSO is
only configured on the master. On the spare, apache redirects to an
special page. Of course users see that only at master down time.
> Unfortunately DRBD is not an option for me as the as both the master and
> slave sympa servers are active mailservers at the same time which and
> the postgresql-database is also used by exim.  Is thus not just a case
> of duplicating a block device.
>   
then access your disks via NFS or any GNBD. DRBD give you storage
redundancy, and you need concurrency writes.

With 2 servers you can think at XEN virtualization. Putting your sympa
instances in XEN virtuals hosts, you can move you live  master instance
to the spare physical server, without stopping it, that gives you time
for master upgrade and reboot. As sympa accesses its spool via, say, NFS
on a remote server, there is no trouble to migrate xen domains from one
host to another whithout stopping sympa.

Sylvain.