ROBLEM DESCRIPTION
====================
IMP 10.5.2.22900-2 (IMP 10.5.2SU1)
1) XCP Router service not running on IMP node
2) Presence Redundancy group page shows Assigned users is not same as Active users.
According to help page:
Assigned Users: Displays the number of users who are assigned to this IM and Presence Service node.
Active Users: Displays the number of users that are homed to this IM and Presence Service node in any given high availability state. This number only changes when a high availability event occurs in the redundancy group. In the Normal state, the active user count equals the assigned users count.
TROUBLESHOOTING
================
> We found that XCP Router issue was result of services set to debug level. Whenever XCP Router service is set to debug level for extended period of time, that service can
intermittently crash. This is related to this bug:
https://bst.cloudapps.cisco.com/bugsearch/bug/CSCuq31057/?reffering_site=dumpcr
> The XCP router issue may have subsequently caused an IMDB replication issue. You are hitting the following bug:
https://bst.cloudapps.cisco.com/bugsearch/bug/CSCuy70097/?reffering_site=dumpcr
> When checking IMDB replication with the below SQL commands. we found that "TTSOFT" output was in the 4000s on pub and 5000s on sub. Output should be 0 on both nodes.
run pe sql ttlogin select count(*) from typesysreplication
run pe sql ttsoft select count(*) from typesysreplication
run pe sql ttreg select count(*) from typesysreplication
> IMDB replication being out of sync can cause inconsistent HA status (as was seen on your server where assigned users was not equal active users)
> to resolve IMDB replication issue, we ran the following workaround. This resolved both the SQL output and the issue with active users not equal to assigned users
Restart all IMDB datastores.
Disable High Availability (HA) for the IMP sub-cluster.
Stop Cisco Presence Engine on all Nodes.
utils service stop Cisco Presence Engine
Verify all the Datastore Services are running: Cisco Login Datastore, Cisco Route Datastore, Cisco Presence Datastore, Cisco SIP Registration Datastore.
utils service list
Restart Cisco Config Agent on each node one at a time.
utils service restart Cisco Config Agent
Start Cisco Presence Engine.
utils service start Cisco Presence Engine
Enable HA for the Sub-cluster.
ACTION PLAN
============
Right now, all appears to be good on your servers. However, please note that the fix was only a workaround. In the event of another network outage
or something similar, it's possible that you could hit another IMDB replication failure. To permanentaly fix the IMDB replication bug, IMP needs to
be upgraded to 10.5.2 SU2 and have ciscocm.cup-imdb-update-v1.1.cop.sgn installed. The cop file, ciscocm.cup-imdb-update-v1.1.cop.sgn, can be dowloaded
from Cisco.com and will be found in the IMP services Utils folder.