HP StorageWorks Storage Mirroring Evaluation Guide (T2558-96326, April 2009)

4 - 10
Simulating a failure
To fully evaluate failover, you need to simulate a failure. The Failover Control Center does not have
to be running in order for failover to occur, but for the purpose of this evaluation, make sure that it
is running so that you can see each step of the process.
1. Ping the source’s IP address from a client machine.
2. Ping the source’s machine name from a client machine.
3. Disconnect the network cable(s) on the source. Notice immediately, that the Failover Control
Center Time to Fail counter decreases and never resets. You will see the icons change to yellow
and eventually to red.
4. You will be prompted to determine how to apply the data in queue on the target. Select Apply
Data in Target Queues Then Failover. Once the icons are red and the
Failed Over message
is displayed, failover has occurred.
5. Ping the source’s IP address from a client machine.
6. Ping the source’s machine name from a client machine.
As you can see, the target has taken on the identity of the source. Application and user requests
destined for the source are routed directly to the target. The impact on your end users is minimal.
Simulating data changes after failover
While your source is failed over to your target, end users continue to work without interruption and
the data on the target will be updated. To fully evaluate the next step, restoration, simulate the
changes that the end users would have made on the target while the source was unavailable.
1. Identify the file that you edited earlier on the source.
2. Locate that same file on the target and make edits to it. Save the changes.
3. Modify the other three files from earlier, but this time make the modifications on the target copy
of the file. Save the changes.
If desired, you can also test the target data as you did earlier. See Testing your target data on
page 4-7. You can test user data using the associated application, and you can save the changes if
desired. If you want to test application data, start the application services on the target, and test the
application data by using clients to connect to the application. Because the source is now failed over,
you will not need to worry about pausing the target, taking a snapshot, or configuring clients to
access the application from the target. The clients will continue to access the source, which is now
being handled by the target machine.
NOTE: The Windows Event Viewer on the target provides details on the actual steps that
have occurred during failover.