Synchronization phase of the XRF process

During the synchronization phase, the alternate IMS becomes an image of the active IMS, as shown in the following figure. The phase begins when the environments are fully initialized.

Figure 1. XRF synchronization phase
Two boxes show two CPCs. Each CPC has OS, IMS, and VTAM regions. Three cylinders show databases, a log, and an restart data set (RDS). A line shows the ISC link connecting the two IMS systems.

The process of building the control blocks in the alternate IMS is very simple. The alternate IMS determines if the last checkpoint was a SNAPQ. If it was not, and you have an ISC link, the alternate IMS requests that the active IMS take a SNAPQ checkpoint. The SNAPQ log records are used to check that the IMS system definitions on both systems match and to build an initial snapshot of the active IMS on the alternate IMS.

If you do not have an ISC link, the alternate IMS issues a message to ask the master terminal operator at the active IMS to request the SNAPQ checkpoint.

The alternate IMS opens the IMS system log, locates the SNAPQ checkpoint, and reads from the log data sets.

If the ISC link exists, a message is sent to the alternate IMS while the checkpoint is being written to the log. If the ISC link does not exist, the alternate IMS monitors the restart data set (RDS) for the completion of SNAPQ.

IMS uses the contents of the dump for several things:
  • The alternate IMS determines which MODSTAT and RDS are being used by the active IMS.
  • The alternate IMS tries to preallocate and pre-open databases and areas that are accessed by the active IMS.
  • MSDBs are loaded by the alternate IMS from the SNAPQ checkpoint records; separate MSDB checkpoint data sets must be available for the alternate IMS.
  • If your XRF complex uses USERVAR, backup sessions for terminals that are opened on the active IMS are initiated on the alternate IMS for those class-1 terminals that are defined with the BACKUP option.
  • Security control blocks are loaded on the alternate IMS from the RACF® data set if you are using RACF for security.
  • The DBRC subsystem record is updated to indicate that an alternate subsystem exists. This suppresses the DBRC ‘SIGNOFF ABNORMAL' message in the case of failure of the active IMS.
  • Status of dependent regions on the active IMS is tracked on the alternate IMS.

Only after a takeover can the alternate IMS write to the log.

Operators can start dependent regions in the alternate IMS to have them ready for a takeover.

When the alternate IMS has processed the SNAPQ checkpoint, the synchronization phase is complete. When the control blocks of the alternate IMS are synchronized with the ones in the active IMS, the alternate IMS is able to take over the processing of the active IMS. From this point on, you can request a planned takeover.