Alternate Master

The Storage Automated Diagnostic Environment supports the concept of an alternate master. An alternate master is a slave that, on every run of the cron, verifies that the master is still alive and, if the master does not respond, takes over some of the responsibilities of the master. All slaves, including the alternate master, have a copy of the complete configuration. This configuration describes where all the agents are located. This information enables the alternate master to call the slaves and temporarily redirect the flow of events from the master to the alternate master.

Since the master is responsible for sending events and email messages, one of the main functions of the alternate master is to alert the administrator that the master server is no longer operational. The alternate master does not try to become the master; it will, however, remember which agent is the real master and will relinquish its role as temporary master once communication with the real master is regained. This architecture is meant to deal with temporary loss of the master agent. If the master agent is removed from the site, a different server should be made the permanent master (by running ras_install again).


Note -

You must manually change a master's status when it becomes an alternate master by entering the IP address of the new master in the DATA/MASTER file.