For performance reasons the Clustering database does not have any restart / recovery features as would be found on many commercial database systems. In the event of a power failure or disk crash during a long clustering task the database and index will most likely be corrupted and the task would need to begin from the start again.

To avoid such a situation it is possible to create a backup of the necessary files, but the

CLUSTER

step needs to be in the wait state (refer to the

Stopping and Restarting Clustering

section) for you to do this. Once in a wait state, close the Console Client, bring down the DCE Server and create a backup copy of the following files from the Server Working Directory:

If a problem occurs, then these files can be restored to their original location and the Resume Clustering feature of the Console used to start again from the time that the backup was taken.