Checkpoint reconfiguration: An overview

The checkpoint reconfiguration facility allows you to dynamically redefine the checkpoint data set definitions for your JES2 multi-access spool (MAS) configuration. Whether the data sets reside on a DASD volume or on a coupling facility structure, the reconfiguration, through its operator dialog allows you to:

Replace a checkpoint data set.
Discontinue the use of a checkpoint data set.
Resume using a previously suspended checkpoint data set.
Terminate JES2 during the dialog.
Start using a new checkpoint data set.

CAUTION:

JES2 does not support moving the checkpoint data set from one volume serial to another unless the checkpoint dialog is used.

Additionally, you can either enter a checkpoint reconfiguration dialog to change the DASD to any supported type or model or change the DASD while involved in a dialog initiated for other reasons. Device types used to hold the checkpoint data set can be a mix of any JES2-supported DASD (see IBM devices supported by JES2 and how to use them for a list of these devices) or coupling facility structures.

If your checkpoint is on a coupling facility and all you want to do is move it to a different coupling facility, then you can use the SETXCF command to start a rebuild instead of using checkpoint reconfiguration. However, only checkpoint reconfiguration can move the checkpoint from a coupling facility to DASD. See Moving a JES2 checkpoint to a coupling facility using system-managed rebuild for more information.

Depending on the situation, either JES2 or the operator can initiate the checkpoint reconfiguration process. The following situations initiate a checkpoint reconfiguration dialog:

JES2-initiated:
- During initialization processing:
  - Either JES2 could not determine the checkpoint data set specifications, or JES2 requires operator verification of the data set specifications.
  - You specified PARM=RECONFIG as a JES2 start option.
  - You specified RECONFIG=YES on the CKPTDEF initialization statement.
    For a discussion of the JES2-initiated checkpoint reconfiguration at initialization, see Initialization.
- Because of I/O error:
  JES2 automatically initiates a reconfiguration when an I/O error occurs during a read or write of the checkpoint data set. For a discussion of the JES2-initiated checkpoint reconfiguration see I/O error.
- Because of coupling facility structure becomes volatile:
  As result of processing based on your CKPTDEF VOLATILE= specification when a coupling facility on which a checkpoint resides becomes volatile. A coupling facility structure is considered volatile if when power to the coupling facility is lost, all data on the facility is lost. A volatile structure might not be an immediate problem, but is serious enough that JES2 prompts you to take precautions to prevent potential data loss.
  
  Note that a coupling facility is always considered volatile unless you have taken measures to protect against the loss of power to that coupling facility. See ES/9000 and ES/3090 PR/SM Planning Guide for a description of how to monitor the volatility of coupling facilities.
  
  For a discussion of the JES2-initiated checkpoint reconfiguration because of the coupling facility structure becoming volatile, see Volatile coupling facility.
Operator-initiated:
Depending on your installation's use of DASD or coupling facility structures or both to maintain your checkpoint data, you might need to occasionally change your checkpoint definitions. For example, you might choose to initiate a checkpoint reconfiguration because the device on which the checkpoint data set(s) resides is scheduled for service. However if your checkpoint is on a coupling facility, it may be simpler to use system managed rebuild to move your checkpoint to another coupling facility. See Moving a JES2 checkpoint to a coupling facility using system-managed rebuildfor more information

For a discussion of the operator-initiated checkpoint reconfiguration see Operator-initiated entrance into a checkpoint reconfiguration dialog.

The reconfiguration allows minimal disruption across all checkpoint data set failures, thus preventing the need for a JES2 restart.

If your checkpoint is on a coupling facility and all you want to do is move it to a different coupling facility, then you can use the SETXCF command to start a rebuild instead of using checkpoint reconfiguration only if the structure attributes remains the same. In the event that the structure size needs to be enlarged, then the system managed rebuild (SETXCF command) will be unsuccessful.

Additionally, system managed rebuild (SETXCF command) is only supported when the structure is to be moved to a different coupling facility, it cannot be used to move the checkpoint structure to another structure which resides within the same coupling facility.

The reason why JES2 does not support system managed rebuild to move the structure within the same coupling facility, or does not support resize of structure to another coupling facility is because JES2 connects (IXLCONN) to the structure with the ALLOWALTER option set to NO.

If the structure size needs to be changed, then this can only be done using the dynamic reconfiguration dialog process.