PRILOG compression

A data set entry is defined as being inactive when it is older than all of the following criteria:

• Log retention period
• Oldest allocation (ALLOC) for any database updated on that log
• Earliest restart checkpoint for the online IMS

PRILOG record compression deletes inactive data set entries up to the oldest ALLOC on the log or the first gap in the data set entries. A gap occurs when an OLDS has not yet been archived.

When inactive data set entries are deleted from active PRILOGs, they are compressed to a single dummy data set entry that has the same start time as the start time of the log and the same stop time as the stop time of the last inactive data set entry deleted.

PRILOG compression can be performed using the following two methods:

Automatic PRILOG compression

PRILOG record compression is attempted automatically after an OLDS is archived.

Compression includes the deletion of all inactive data set entries in the PRILOG record. When applicable, corresponding entries in the SECLOG, PRISLD, and SECSLD records are also deleted.

Manual process for PRILOG compression

You can initiate PRILOG record compression manually by using the DELETE.LOG INACTIVE command. This command deletes inactive data set entries from active PRILOG records and deletes entire inactive PRILOG records.

If your PRILOG is not getting compressed, you will receive message DSP1150I for one of the following reasons:

• EARLIEST CHECKPOINT

  If this time stamp is within the first DSN entry in the PRILOG, then compression is not possible. Most often, the oldest checkpoint needed for restart causing the problem is the checkpoint to rebuild the message queues. If so, take a SNAPQ checkpoint. When the OLDS that records this SNAPQ checkpoint is archived, the earliest checkpoint time stamp is updated.

• EARLIEST ALLOC TIME

  The databases listed in the LOGALL record are in earliest allocation time stamp order. The earliest allocation time listed in the LOGALL record is associated with the first databases in the listing.

  Concurrent image copies (CICs) and Online image copies (OICs) update the allocation time to reflect the earliest log that is needed for recovery. The checkpoint IDs and checkpoint counts on the logs before the start of the CIC are used to determine where the new allocation time is set. If the allocation times are not being moved forward-in-time, ensure that checkpoint IDs and counts are being recorded in the DSN entries.

  Old allocations for any databases in the LOGALL record can cause compression to fail. Image copy databases on a regular basis to allow old allocations to be deleted.

• LOG RETENTION TIME

  Indicates the minimum amount of time that DBRC is to keep log records in the RECON data set.