CICS and Automatic Restart Manager

The z/OS® Automatic Restart Manager (ARM) provides the ability to automatically restart address spaces, either after an address space failure or a system failure. You can use ARM to restart a CICS® region, following a region failure.

Whether you use ARM to restart CICS on an alternative system following a system failure depends on whether you use ARM to restart any connected subsystems, for example, DB2®, on another system. Assuming that you have a number of instances of each CICS region type on every member of the sysplex, it should not be necessary to start CICS regions on another image from the perspective of providing a transaction processing capability. However, if you restart a connected subsystem on another system, it might be necessary to also restart any CICS regions that were connected to that subsystem to resolve INDOUBT units of work. If you do not do this, locks associated with these units of work are not released, potentially impacting other transactions that try to use those resources.

In ARM, you can specify a group of address spaces that must all be restarted on the same system if there is a system failure. If you use ARM to restart DB2 elsewhere, include the CICS regions that are associated with that DB2 in the same group.