Journal data set

The journal data set provides DFSMShsm with a record of each critical change made to a control data set from any host since the last time that CDS was successfully backed up. DFSMShsm recovers control data sets by merging the journal with a backed up version of the control data set. Control data sets cannot be recovered to the point of failure without the journal. Use of the journal is highly recommended.

The journal data set is a special type of data set. It contains a control record at the beginning that points to the last record written to the journal. However, the rest of the data set is processed as a sequential data set, containing a record of each significant change made to the MCDS, BCDS, and OCDS. The journal data set normally does not contain an end-of-file marker. Ensure that you allocate the journal data set only on a DASD volume.

In a multiple DFSMShsm-host environment, if hosts share a single set of control data sets, they must also share a single journal. All DFSMShsm recovery procedures are based on a single journal to merge with a backed up version of a control data set.

Journal Data Set

Required:

No, but strongly recommended.

Allocated by starter set:

Yes, see Starter set example.

Data set type:

Physical sequential data set.

Storage guidance:

The DFSMShsm journal must reside on a DASD device.

The first record of the journal keeps track of the location of the last record in the journal. This pointer is the full disk address. Therefore, the journal data set must not be moved to another location on the current volume or another volume.

To avoid a performance problem, the DFSMShsm journal must not share a volume with any system resource data sets or with any of the DFSMShsm control data sets.

The journal can be defined as a large format data set, which can exceed 65 535 tracks. Doing so can help to reduce the frequency of journal full conditions. For information about large format data sets, see z/OS DFSMS Using Data Sets.

The journal should not share an HDA with any of the CDS data sets. This is important, so that if physical damage occurs, both the journal and a CDS will not be lost at the same time.

The journal is backed up at the beginning of automatic backup (along with the control data sets) and is backed up differently from user data sets. The Journal and Control data sets must be backed up. For information about specifying the journal backup environment, see Defining the backup environment for control data sets.

Also see Considerations for DFSMShsm control data sets and the journal.