Cause
The two possible causes for this event are:
An unexplained Linux lock-up that starved CPU cycles from all Communication Manager processes for more than 3.3 seconds.
A third main server with a server ID that matches one of the other two main servers was somehow started and is accessible from the network.
In the case of a Linux lock-up, the problem corrects itself by the time the problem is detected and corrective action is not required.
Proposed Solution
Procedure
- On the Linux command line, enter /sbin/arp -a to determine the MAC addresses of the alternate server ethernet ports associated with the CNA, CNB, and duplication links.
- Log in to the alternate server and verify that the MAC addresses match. Do this from both servers.
- Check for a mismatch in the MAC addresses that indicates the presence of a third system posing as a doppelganger.
- Use a network sniffer to find the third main server.