Recovering a cluster after a complete system loss
Use this information with the appropriate checklist in the Recovering your system topic for recovering your entire system after a complete system loss when your system loses power unexpectedly.
Scenario 1: Restoring to the same system
- In order to prevent inconsistencies in the device domain information between
the Licensed Internal Code and IBM® i,
it is recommended that you install the Licensed Internal Code by using option
3 (Install Licensed Internal Code and Recover Configuration). Note: For the Install Licensed Internal Code and Recover Configuration operation to succeed, you must have the same disk units -- with exception of the load source disk unit if it has failed. You must also be recovering the same release.
- After you have installed the Licensed Internal Code, follow the Recovering Your Disk Configuration procedure in the Recovering your system topic. These steps will help you avoid having to reconfigure the disk pools.
- After you have recovered your system information and are ready to start clustering on the node you just recovered, you must start clustering from the active node. This will propagate the most current configuration information to the recovered node.
Scenario 2: Restoring to a different system
After you have recovered your system information and checked the job log to make sure that all objects have restored, you must perform the following steps to obtain the correct cluster device domain configuration.
- From the node you just restored, delete the cluster.
- From the active node, perform these steps:
- Remove the recovered node from the cluster.
- Add the recovered node back into the cluster.
- Add the recovered node to the device domain.
- Create the cluster resource group or add the node to the recovery domain.