IBM MQ requests
At a high level, all processor-unique queues are maintained in system heap and IBM MQ memory to obtain extremely high access rates to z/TPF queues.
To assist with the recovery of persistent messages after a system IPL, the contents of all queues are checkpointed on a time-initiated basis to processor-unique DASD records. All updates made by using the MQPUT, MQPUT1, or MQGET functions between checkpoints are written to the recovery log.
The recovery log is used during restart for the recovery of IBM MQ requests that might have been
in transition or lost. The size of the recovery log must be large enough to accommodate all logging
activity for all z/TPF resource managers (not just
the IBM MQ resource manager) between each successfully completed checkpoint. The checkpoint interval
is 5 seconds. (If your z/TPF system is running under
VPARS, the checkpoint interval is 1 second.) Enter the ZMQSC DISPLAY command with the QMGR parameter
specified to view the number of seconds that were taken to complete the longest IBM MQ checkpoint
since the last z/TPF system IPL. (This value is
labeled CHECKPOINT HIGH WATER MARK in the display.) This value can help to ensure
that your recovery log can accommodate all the necessary logging activity.
All processor shared queues are filed on z/TPF collection support (z/TPFCS).