Re: ZODB replication with DRBD

On 2018-08-28 21:41, Héctor Velarde wrote:
> Jim, I was last week in Brasília and learn about a replication schema
> with multiple ZEO server in master mode that sys admins of some bodies
> have implemented there: they don't use ZRS, but a stack with Corosync,
> Pacemaker and DRBD.
> 
> AFAIK, they replicate the ZODB with DRBD and, in case of failure of a
> ZEO server, they mount it on other machine and switch the IP to it;
> that's completely transparent to the clients, as the same IP is used.
> 
> they told me they have sites running with this configuration for years
> with almost no problems, away from sporadic ones related with sessions.

Just FYI, I have exact such an setup running for about 7 years now at a
client with >40 Plone instances in Austria. I even held a talk (in
German language) in 2012 about that topic (and all around), slides/video
at https://de.slideshare.net/jensens/pk2012-noekubda

Today I would probably go the k8s/docker/container way or use ZRS. But I
must admit, the setup with DRBD/corosync/pacemaker is very very robust.

HTH Jens

-- 
Klein & Partner KG, member of BlueDynamics Alliance

-- 
You received this message because you are subscribed to the Google Groups "zodb" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
For more options, visit https://groups.google.com/d/optout.