Forward recovery of data sets accessed in RLS mode
A recoverable data set that is updated in RLS mode can have retained locks held for individual records.
In the event of a data set failure, it is important to ensure that you preserve any retained locks as part of the data set recovery process. This is to enable the locks associated with the original data set to be attached to the new data set. If the data set failure is caused by anything other than a volume failure, retained locks can be "unbound" using the SHCDS FRUNBIND subcommand. The data set can then be recovered and the locks rebound to the recovered data using the SHCDS FRBIND subcommand.
If a data set failure is caused by the loss of a volume, it is not possible to preserve retained locks using FRUNBIND and FRBIND because SMSVSAM no longer has access to the failed volume. When recovering from the loss of a volume, you can ensure data integrity only by deleting the entire IGWLOCK00 lock structure, which forces CICS to perform lost locks recovery. CICS uses information from its system log to perform lost locks recovery. (For more information about lost locks processing, see Lost locks recovery.) Recovering data from the loss of a volume requires a different procedure from the simple loss of a data set.
- For recovery of failed data sets where the volume remains operational, see Recovery of data set with volume still available
- For recovery of failed data sets where the volume itself has failed, see Recovery of data set with loss of volume.