About Failover

Process Manager supports an optional failover feature, which provides redundancy for the Process Manager Server.

The failover feature allows you to configure a second Process Manager Server host to take over the responsibilities of the primary Process Manager Server host if it should fail. The failover feature includes the Enterprise Grid Orchestrator(EGO) or failover daemon (fod, in case of UNIX), which starts the Process Manager Server on the primary Process Manager Server host. The failover daemon monitors the health of the primary Process Manager Server, starting Process Manager Server on the failover host if the primary fails to respond within a certain time period.

The failover feature relies on a shared file system for access to the working directory of the Process Manager Server.

  1. Process Manager Server updates flow status in its working directory based on data it reads from lsb.events.

  2. The fod or EGO on the failover host monitors the primary host. If it receives no response from the heartbeat, it assumes the primary host is down, and starts jfd on the failover host. Process Manager Server is now running on the failover host.

  3. The fod on the failover host continues to monitor for a response from the primary host. When it receives a response, it stops jfd on the failover host, returning control to the primary host.

    The failover host requires access to both the Process Manager working directory JS_TOP/work, and the events file lsb.events.