Takeover restart processing

IMS Online Reorganization Facility allows a Takeover phase restart to occur if the previous reorganization job failed or was delayed during the Takeover phase. When a takeover restart occurs, the Takeover phase automatically restarts from where it stopped to continue IMS Online Reorganization Facility processing.

The restart key-sequenced data set, which is defined during installation, ensures data integrity and availability if a reorganization fails or is delayed. IMS Online Reorganization Facility stores information that is required to restart during the Takeover phase in this data set. At the beginning of each job, this data set is read to determine if a restart record exists for a specific DBD or partition name.

A restart can be performed when the previous online reorganization job for a specific database (DBD name or partition name) failed during the Takeover phase or was delayed (TAKEOVER(DELAY)) by the previous control statements.

If the control statements or restart data set are not modified, the next time that the same reorganization job runs, IMS Online Reorganization Facility reads the data in the restart data set at the beginning of the job and detects a restart pending. If you specified RESTART(AUTOMATIC), a restart automatically begins. If the restart completes successfully, the database is made available online, and the restart data is removed from the restart data set so that subsequent jobs can start from the beginning.

The default is RESTART(N). In this case, the reorganization job fails with the following error message:

HRF01147E  RESTART(NO) specified for this execution

To allow a restart to begin, you must change the RESTART option to RESTART(AUTOMATIC).

If you recovered the database or partition or if you do not need to perform a restart, specify RESTART(OVERRIDE) to override the restart status. Generally, if you manually recovered the database or the database state, you can consider using the RESTART(OVERRIDE) option. In this case, the reorganization job starts at the first phase as though a restart pending never occurred.

Attention: Use the RESTART(OVERRIDE) option with caution because this option removes the restart data for the database or partition from the restart data set.