Jabber. Cannot assign a user to a server that is not in a valid state

[JBrav0]

There seems to be an issue with my Im Pub server for jabber. 

I noticed that as I was trying to create a Jabber account I could not sign them in. 

Eventually under CUCM Administration User Management > Assign Presence User Assignment > selected user and tried to move the user to the IM publisher for jabber server and it through the error code "Update failed. Cannot assign a user to a server that is not in a valid state". 

So I checked out the server. 

Still under CUCM Admin Servers> IMP server node status and everything is good. 

However, when I go to the Presence Redundancy Group Configuration section (still under CUCM Admin) I notice that Server Status is Running in Backup Mode and Reason is Peer Down. The Subscriber lets me do a fallback. 

When I go to CUCM IM and Presence Admin I noticed that under service for both publisher and subscriber having Cisco XCP Message Archiver being down and cannot start it up. 

What's the protocol here? Im not really familiar with the system and just took over for the person that was. 

Jabber still works with members that already were signed up. 

Do you just do a fallback? or a restart? If I do a fallback will I lose all the members and have to re-add them? 

Thanks for the help in advance. 

[Roger Kallberg]

As your pub looks to be operational I would recommend you to do the fallback. It should not affect an of the things you ask about. Very likely the users will see a small dip in communication with the IM&P system as they are assigned to their “home” server as they are evenly distributed between the two.

About the service not started/starting, do you actually use message archiving? Otherwise you should never be able to start this service.

[Sankar Voleti]

Hi,

   You can just click 'Fallback' button like Roger mentioned to re-balance the users. Also from Node Status pt2.png, XCP Message Archiver service is not monitored if you see the 'Monitored' column. So even if you can't start it, that's fine and won't be responsible for any automatic user failover instances. Only those services that are monitored should always be running failing which, users will failover to the other node.

Hope this helps!

-Sankar