Specifications

3-42
Cisco Unified Communications Manager Managed Services Guide, Release 8.0(1)
OL-20105-01
Chapter 3 Managing and Monitoring the Health of Cisco Unified Communications Manager Systems
General Health and Troubleshooting Tips
Database Replication Does Not Occur on Lost Node
Database replication does not occur when connectivity is restored on lost node recovery. You can verify
the state of replication by using the methods given in the topic
Database Replication Fails Between the
Publisher and the Subscriber, page 3-39. Only use the following procedure if you have already tried to
reset replication on the node, and have been unsuccessful.
Possible Cause
The CDR check remains stuck in a loop, due to a delete on device table.
Recommended Action
Step 1 Run utils dbreplication stop on the affected subscribers. You can run them all at once.
Step 2 Wait until Step 1 completes, then, run utils dbreplication stop on the affected publisher server.
Step 3 Run utils dbreplication clusterreset from the affected publisher server. When you run the command,
the log name gets listed in the log file. Watch this file to monitor the process status. The path to the
follows:
/var/log/active/cm/trace/dbl/sdi
Step 4 From the affected publisher, run utils dbreplication reset all.
Step 5 Stop and restart all the services on all the subscriber servers [or restart/reboot all the systems (subscriber
servers)] in the cluster to get the service changes. Do this only after utils dbreplication status shows
Status 2.
Database Tables Out of Sync Do Not Trigger Alert
Out of sync means that two servers in the cluster do not contain the same information in a specific
database table.
On Cisco Unified Communications Manager Version 6.x or later, the symptoms include unexpected call
processing behaviors. Calls do get not routed or handled as expected. The symptoms may occur on either
the publisher or on the subscriber servers.
On Cisco Unified Communications Manager Version 5.x, the symptoms include unexpected call
processing behaviors. Calls do not get routed or handled as expected but only when the publisher server
is offline. If you see these symptoms, you can run the utils dbreplication status command “Out of sync”
displays. If “Out of sync” does not display, this is not the problem.
Possible Cause
Database tables remain out of sync between nodes. Replication alerts only indicate failure in the
replication process and do not indicate when database tables are out of sync. Normally, if replication is
working, tables should remain in sync. Instances can occur in which replication appears to be working,
but database tables are “Out of sync”.