Populating the common recall queue

A host that is participating in CRQplex-wide recall processing attempts to place all of its recall requests onto the CRQ. Requests are placed on the CRQ if the following conditions are true:

If recall requests are unexpectedly not placed onto the CRQ, issue the QUERY ACTIVE command and examine message ARC1540I. It will indicate which of the above conditions is not being met.

Any request that cannot be placed onto the CRQ is placed on the local queue of the originating host.

Recall requests that can be placed onto the CRQ are sorted by WAIT or NOWAIT status and priority. WAIT requests are always a higher priority than NOWAIT requests. Requests of the same priority are sorted in FIFO order, with the exception that NOWAIT requests of priority 50 are subsorted with requests from multiple users. This subsorting prevents a single user with many requests from blocking other users' requests for an extended period of time.

The host that initiates the recall request invokes user exit ARCRPEXT, if the exit is enabled. The processing host invokes user exit ARCRDEXT, if the exit is enabled. You can ensure consistent processing by using common RD exits throughout the CRQplex. The RP exit may be used to assign requests that originate on one host a higher or lower priority in relation to the priorities set by RP exits on other hosts. The default priority for these requests is 50. Refer to z/OS DFSMS Installation Exits for a description of these user exits.

The SETSYS EMERGENCY command does not affect the placement of recall requests onto the CRQ. This allows a host that is in emergency mode to place its requests onto the CRQ and enables other active hosts to process those requests. Use the HOLD COMMONQUEUE(PLACEMENT) command to prevent recalls from being placed onto the CRQ when the emergency mode is in effect.