Logical recovery

Logging and enqueuing occur with logical recovery transactions (including dynamic backout of the failing task's activity on the transient data queue). Logical recovery is generally used when a group of records have to be processed together for any reason, or when other recoverable resources are to be processed in the same task.

During processing of the transient data request, the destination queue entry is enqueued from the first request, for either input or output, or both (if the queue is to be deleted), until the end of the UOW. This means that none of the other tasks can access the queue for the same purpose during that period of time, thus maintaining the integrity of the queue's status.

At the end of the UOW (sync point or task completion), sync point processing takes place and the queue entry is logged. Any purge requests are processed (during the UOW, a purge only marks the queue ready for purging). The empty control intervals are released for general transient data use. Any trigger levels reached during the UOW cause automatic task initiation to take place for those queues that have a trigger level greater than zero. The buffer is written out to the VSAM data set as necessary.

The DEQUEUE function on the queue entry occurs, releasing the queue for either input or output processing by other tasks. Records written by a task can then be read by another task.