Table Of Contents
Field Notice: Cisco Unified MeetingPlace Application Server Can Go Down Due to Replication Issues
Field Notice: Cisco Unified MeetingPlace Application Server Can Go Down Due to Replication Issues
Published: June 21, 2010This field notice is for customers who have not already encountered a replication issue. We recommend that all such customers apply the noted hotfix as a preventative measure.
Problem
Due to an issue with the IBM Informix database management system, the sblob_dbspace may get full during the replication process and cause the Cisco Unified MeetingPlace Application Server to fail. The failure will see the Application Server reboot multiple times when attempting to start MeetingPlace services, and the services will fail to start when attempting to use the CLI command mpx_sys restart. Additionally, output from the CLI command mp_replication status on either node 1 or node 2 may show that the connected (remote) server has a very large value in the Queue column.
Workaround/Solution
Cisco has made a hotfix available called sblob_dbspace-fix.sh. The hotfix creates a second database chunk (sblob_dbspace2), which is approximately 1GB in size or 20x larger than sblob_dbspace. Once applied, the system will automatically start filling sblob_dbspace2 once sblob_dbspace is full.
To apply the hotfix, contact your Cisco TAC representative.
Note
Without the hotfix, the system raises an alarm when sblob_dbspace is 90% full. Under heavy load, this gives you approximately three minutes to create additional space for the database replication queue. If the additional space is not created within this time, the system will fail.
