z/OS JES2 Initialization and Tuning Guide
Previous topic | Next topic | Contents | Contact z/OS | Library | PDF


Moving a JES2 checkpoint to a coupling facility using system-managed rebuild

z/OS JES2 Initialization and Tuning Guide
SA32-0991-00

Managing a JES2 checkpoint in a coupling facility is different from managing a JES2 checkpoint on DASD. The main differences is coupling facilities store data in temporary memory versus the permanent storage of a DASD. The data JES2 stores in a coupling facility is needed for JES2 restart. If CKPT2 is placed on DASD, it can be used by JES2 restart processing if the data in the coupling facility is lost. However, if both checkpoints are on coupling facilities and the coupling facility data is lost (for example because of an extended power outage), then JES2 must be cold started, resulting in the loss of all Jobs and SYSOUT in the JES2 work queues. Because of this, special care must be taken with any coupling facility that contains a JES2 checkpoint structure.

In general, to prevent a total loss of the job queue, you should use 2 checkpoints, specify CKPT DUPLEX=ON on all members, and place CKPT2 on a DASD. You can also use nonvolatile coupling facilities (with battery backup in case of power failure) to minimize the impact of a power failure. However, data on a coupling facility can be lost in ways other than a power failure. The most common way JES2 checkpoints structures in a coupling facility are lost is when the coupling facilities are reinitialized while JES2 structures are still active in them. To prevent this, the JES2 structures should be moved out of a coupling facility before reinitializing it. Depending on the availability of an alternate coupling facility, this can be done in one of two ways.

If there are no suitable coupling facilities available to hold the checkpoint data set, or if this is a planned extended power outage, then the checkpoint data set should be moved to a DASD. This can be done by an operator initiated checkpoint reconfiguration. See Operator-initiated entrance into a checkpoint reconfiguration dialog for details on how to do this.

If there is a suitable alternative coupling facility available, then a system managed rebuild can be used to move the checkpoint structure to the new coupling facility. To accomplish this do the following:
  1. Ensure system managed processes are enabled in your SYSPLEX. See "System-Managed Processes Considerations" in z/OS MVS Setting Up a Sysplex for a complete list of the requirements.
  2. Ensure that the structure definitions for the JES2 checkpoint data sets in the active CFRM policy will direct the checkpoint data set to a suitable coupling facility. If not, update and activate the appropriate policy.
  3. Use the appropriate XCF command to either rebuild only the JES2 checkpoint structure (SETXCF START,REBUILD,STRNAME=) or all structures on a coupling facility (SETXCF START,REBUILD,CFNAME=).
  4. When the rebuild completes, use the D XCF,STRUCTURE,STRNAME= to display the checkpoint structure and verify it is now on a suitable structure.
During the rebuild process, you may notice that JES2 is not responding to commands and not starting new work. Other MAS members may issue messages about not being able to access the checkpoint. JES2 will return to normal operations when the rebuild completes. To minimize the time required to complete the rebuild, you can request the JES2 structures be rebuilt individually by using the SETXCF START,REBUILD,STRNAME= instead of specifying CFNAME=. Alternatively, the structures can be rebuilt when JES2 is not active.
Note: JES2 supports rebuild of its structures only for the SETXCF START,REBUILD command. In the event of any other failure of the coupling facility (such as loss of connectivity), the I/O form of checkpoint reconfiguration will be used for recovery.

Go to the previous page Go to the next page




Copyright IBM Corporation 1990, 2014