Tracking phase of the XRF process
Most of the time, your XRF complex is in the tracking phase. The active IMS processes the requests of IMS users, writes to the log, and sends signals across the surveillance mechanisms that were set up in the initialization phase. The alternate IMS tracks the active IMS to reduce the work that the alternate IMS must do at takeover.
The tracking phase looks similar to the XRF synchronization phase that was shown in Synchronization phase of the XRF process.
Tracking activities of the alternate IMS include:
- Reading the log
- Using log information to update control blocks
- If your XRF complex uses USERVAR, opening and closing backup sessions for class-1 terminals
- Checking the surveillance mechanisms and the log records for signs of failure in the active IMS
Tracking activities use few of the non-storage resources of the alternate IMS. The use of real storage depends on how you page fix IMS storage and whether you define a member of DFSFIXxx with page-fixing options that are in effect only when the subsystem is active.
Updating control blocks in the alternate IMS
The alternate IMS reads the OLDS (Use dual OLDSs, because tracking can continue using the other OLDS if one OLDS is lost.) to track:
- Communication network status
In addition to the usual terminal status tracking that is done during the emergency restart process, the alternate IMS monitors the terminal connection status. This is done initially from records in the SNAPQ checkpoint and subsequently from terminal connection status change log records written by the active IMS. For class-1 terminals in an XRF complex that uses USERVAR, backup sessions are maintained by the alternate IMS. When class-1 terminal sessions are terminated on the active IMS, their backup sessions are also terminated on the alternate IMS. The user can never log onto the alternate IMS. Terminal types are described in Terminals in an XRF complex.
- Application program and transaction status
As application programs are scheduled on the active IMS, the alternate IMS builds the control block structure to track the dependent region status.
The operator can prestart dependent regions on the alternate IMS to allow the alternate IMS to complete takeover more quickly.
- Database status
As databases and areas are allocated and opened on the active IMS, the alternate IMS pre-allocates and pre-opens these same databases and data set areas. The IMS macro definitions for all databases and areas must be specified as SHARED. The alternate IMS also tracks all DBRC database authorizations.
Recommendation: Register all databases and areas in DBRC. The alternate IMS also tracks the status of all locks for uncommitted database changes. - Message queue status
Message queues are built on the alternate IMS from the SNAPQ checkpoint records. They are updated on the alternate IMS by OLDS records. A local message queue is used by the alternate IMS to communicate with the operator. During IMS system definition, two master and two secondary master terminals are specified. The master terminals at the active IMS control the active IMS; the master terminals at the alternate IMS control the alternate IMS.
- MFS pool status
MFS blocks are preloaded and released on the alternate IMS as they are loaded and released on the active IMS.
- Dependent region status
The alternate IMS initializes and maintains control blocks to reflect the status of the dependent regions. It monitors all activity in the region so that backout of uncommitted processing can occur at takeover.
With this information, the alternate IMS updates its control blocks to maintain an environment that duplicates the active IMS. It pre-allocates and pre-opens the IMS databases and data areas as the log indicates that the active IMS allocates and opens them. It also loads the MFS control blocks, program specification blocks (PSBs) for backout use, and data management blocks (DMBs).
Surveillance of the active IMS
As described in Surveillance mechanisms, the alternate IMS depends on surveillance to indicate some of the takeover conditions in the active IMS. If surveillance includes the restart data set (RDS), the active IMS periodically places time stamps on the RDS. If surveillance includes the ISC link, the active IMS sends periodic signals over the ISC link. If surveillance includes the log, the alternate IMS periodically checks to ensure that log records continue to arrive.