Hi,
I have a customer with CUCM 14 and UCCX 12.5. THe CUCM cluster has a Publisher and two subscribers. The UCCX cluster has a Publisher and a Subscriber.
The servers are hosted on UCS servers with local storage hosted at to different sites.
The UCS at Site 1 hosts the CUCM Pub, a CUCM Sub and the UCCX Pub.
The UCS at Site 2 hosts a CUCM Sub and the UCCX Sub.
CUCM is integrated with LDAP for user provisioning and authentication.
Users including UCCX agents use Jabber for Windows rather than IP phones.
We have been reviewing their DR strategy which has raised some concerns. If the UCS server at site 1 were to fail or be unavailable due to other issues the CUCM Sub and UCCX Sub at site B would still be available.
Users who have recently logged into Jabber should still be able to connect using cached credentials. Users whose credentials have expired or new users who have not yet logged in would be unable to connect. I believe that the credentials are cached for 30 days but have read conflicting information - does anyone have a definitive answer to that?
UCCX agents who are logged in should be moved automatically to the UCCX Sub but no new logins would be possible as the CUCM Pub which handles the LDAP authentication is off line.
We want to improve this. Disabling LDAP authentication and using locally stored passwords would seem to be a solution but has the downside of users having different passwords.
My main concern is how to restore full functionality if site 1 remains unavailable.
I have seen some suggestions that Veaam could be used to replicate the CUCM Pub and UCCX Pub to the VMware host site site 2 (it has enough compute resource to run them).
Is anyone doing this? If yes do you need to power the VMs off before synching them? I have seen something that suggests quiescing DB replication can be used to do this without turning the servers off. Has anyone tried that (utils dbreplication stop/start).
Thanks for reading if you made it this far!
