The spool is the repository for all input jobs and most system output (SYSOUT) that JES2 manages. Because of fluctuating workload requirements, such as month-end processing, your installation might need to add spool space, or for maintenance reasons such as I/O errors, hardware failure, or utilization needs, you might need to delete idle spool packs. In order to optimize system availability, JES2 provides a facility to dynamically add and delete spool data sets. Through the use of operator commands ($S SPL, $P SPL, and $Z SPL) spool data sets can be added, drained, or halted without interrupting JES2 processing.

Spool data sets can be managed by the operator without affecting the normal processing of work, or any accompanying online system or processor availability. Therefore, the operator has the flexibility of adding or deleting spool data sets to meet your installation's varying workload needs. This facility permits multiple addition and deletion of spool data sets simultaneously, provided the maximum of 253 spool data sets is not exceeded at any one time. (Your maximum number of data sets can be less than 253. This number is dependent upon your JES2 track group size, JES2 buffer size, and the amount of space allocated on each spool data set.)

On a cold start, all mounted volumes whose first four or five characters match the VOLUME= parameter on the SPOOLDEF statement and contain the JES2 data set SYS1.HASPACE are automatically started. Spool volumes that do not contain the SYS1.HASPACE data set that are required for processing can be started dynamically in the future by use of the operator commands after the SYS1.HASPACE data set has been added to that volume.

Following a warm start, JES2 restores the status of the spool configuration. A spool data set can be deleted at an all-member warm start if that data set is not available (for example, the volume is not mounted). However, if you delete a spool data set during an all-member warm start, you will lose all work which resides on those spool data sets (which can include: JESNEWS and spooled messages). If the data set is put into inactive status by the operator response (GO) to the $HASP853 message, all work on that data set can later be recovered by the $S SPL command. However, if the data set is drained by the operator response (PURGE) to the $HASP853 message, all work on that data set is lost.

The operator can use the $D SPL command to display the status of the data set and the jobs currently using a specific data set. The operator should be aware of the contents of a spool data set before deleting it. The deletion of a spool volume will not complete while jobs allocated to that volume are active in one of the phases of job processing. For example, if jobs allocated to the volume are in execution, the deletion command will not complete until the job finishes execution. Never- or seldom-ending jobs such as SYSLOG, IMS™, CICS^* and VTAM® generally must be given special consideration. You must establish procedures to prevent such situations from occurring, and, if appropriate, establish procedures to handle redistribution of long-running jobs and system data sets. Refer to Performance considerations and recommendations for the description of device partitioning procedures. Additionally, the $D J,SPOOL command will display those data sets which are currently allocated to specific jobs.