IM&P Publisher stay in fail over mode after downtime

[nicanor00]

Hi

I have IM&P 12 (2 node in the cluster)

I created presence redundancy group in the CUCM and activated high availability

Everitng was working fine

Then I have an issue on the IM&P publischer, so it was not available for 5h

During the downtime, any new created jabber user was not able to longin, when I check her status in the server, those user are unassigned

Then the publusher get back to UP

see below the status of the 2 node 

publicher :
sever state : Failed over
reason : initialization
Suscriber :
server state Running in backup mode
reason : peer down

Now that the publische get back to up and stack in failover status, any new created jabber user cannot login, when I check those new user status in the server, those user are unassigned

Why user are not automaticaly assigned when the publischer is in failover status ?

How can I assigne those user to the suscriber when publischer has stacked in failover status ?

Why the publischer stack in fail over ?

How can I do to have automatic fallback in this case ?

[Miguel Castillo (miguecas)]

Hi Nicanor,

It is expected that when the HA of the IM&P Servers is in the wrong state, the users cannot be assigned automatically nor manually to one of the servers.

If the IM&P Publisher server is in failed status and you want to assign the users to the subscriber, only two options can be performed:

1. Fix the problem with the IM&P Publisher and recover the HA so it can be normal/normal again.
2. Disable the HA, save, and then assign manually the users.

To know why the IM&P Server is stuck on a failed status, several situations should be taken into account:

• Was there a reboot of a monitored service and HA was not disabled?
  • Keep in mind that HA requires to be disabled if the server is to be rebooted or the following services will be restarted: A Cisco DB, Presence Engine or XCP Router.
• Maybe there was a network disconnection that overpasses the heartbeats keep alive, and that broke the HA
• Are all the services started on the IM&P Publisher?
• Some other options like a core dump or something that should be investigated via the Service Recovery Manager logs

Automatic fallback is disabled by default to avoid saturation in the network traffic:

• But can be enabled by going to CUCM > Service Parameters Configuration > IMP > Cisco Recovery Manager > Enable automatic Fallback > set it to Ture.
• Another solution is to disable the HA > save > Enable HA > save, which will resets the communication and fix the HA wrong state.