Restarting Reorganizations
You must manually restart or abort all reorganizations that fail to complete (for example, because of a system crash).
You can always restart replayable reorganizations, which will continue from the last checkpoint established during the reorganization.
You can restart non-replayable reorganizations only if the system failure occurred while the reorganization was waiting for the transition to be initiated or if the failure occurred when the reorganization was committing. In all other cases, you must manually abort non-replayable reorganizations when the system is restarted, to restore the database to its pre-reorganization state.
When a database server node is restarted following a crash that occurred during a schema instantiation, restart recovery recovers all committed transactions including all transactions applied to the metaschema by schema and forms definition file loads.
An interrupted database reorganization is not automatically restarted as part of database recovery. Restarting an interrupted reorganization requires an explicit action. (For details, see "
Restarting an online reorganization interrupted during any of the online phases performs all file reorganizations again. Restarting an offline database reorganization reorganizes only files that have not been previously completed.
