What does the user need to do?
To minimize impact of failover
The HP B6000 Series Backup System has been designed around the principle of no single point
of hardware failure. To gain the most benefit from this:
•
Configure the external network or fibre channel fabric to enable any failover port configuration
to be available in the same network or fibre channel fabric for any nodes in the cluster
•
Configure the VTL(s) on the service set on the node to use both Fibre Channel connections for
its robotics presentation. The drives within the VTL(s) will be evenly distributed between the
two Fibre Channel ports to allow the backup software to still access a number of drives and
the robotics should one of the ports connections fail.
•
Configure the backup application to handle time-outs intelligently
To enable emails and alerts
An alert is automatically generated if failover takes place. The email system should be configured
to automatically notify relevant users that this has happened so that appropriate action can be
taken:
•
To investigate what has caused the node to fail and resolve the issue
•
To check that backup and restore jobs complete successfully and reschedule from the backup
application, if necessary
•
To prepare to invoke fallback after the issue has been resolved
Running failback
Once the issue has been resolved and the failed node is available again, it must be brought back
into the cluster configuration. This must be invoked manually:
•
From the
Failback this server
button on the appropriate Hardware page of the GUI, see
Monitoring a node (page 96)
.
•
Using the CLI command
hardware failback nodeX
where X is the node identifier. See
the HP B6000 CLI Reference Guide for more detail.
110
Failover