In case of an SMSVSAM failure or a CICS® failure, locked records within some VSAM data sets might not be accessible because of retained locks. This lack of accessibility could prevent many transactions and jobs from being restarted. To maintain high availability, you should resolve the retained locks as soon as possible. They will be resolved by a restart of CICS TS and SMSVSAM. The SMSVSAM address space will restart automatically. It would be helpful to use ARM for CICS automatic restart.

If the MVS™ system image fails, ARM restarts a peer recovery instance of DFSMStvs on another MVS system image.

If the SMSVSAM address space fails, all in-flight units of recovery that were using the DFSMStvs instance at the time of failure are backed out. The backouts for these units of recovery are not performed at the time of the failure by the failed DFSMStvs instance. The backouts are performed either at the time of the failure or later by a peer recovery instance of DFSMStvs.

However, in the case of a z/OS® system failure, restart can take a long time in comparison, and the process is not automatic. So in-flight units of recovery can remain for a long time. To prevent this situation, DFSMStvs provides a function known as peer recovery.

When peer recovery occurs, another SMSVSAM instance in the sysplex (a peer) performs the backout for in-flight units of recovery that were being processed by the SMSVSAM address space that failed. Peer recovery will start automatically if there is an ARM policy (created using IXCMIAPU) which includes DFSMStvs and any other resource managers which might be involved in a unit of recovery in a restart group.

VARY SMS,TRANSVSAM(tvsname),PEERRECOVERY,ACTIVE