z/OS and other failures requiring system recovery

If z/OS® fails or other failures occur that require system recovery (for example, a power outage), IMS processing is the same as it is in the nonsysplex data-sharing environment.

For the failing system, operations consist of:

• Restarting z/OS
• Restarting the IRLM
• In the batch environment, backing out batch jobs if necessary
• In the IMS DB/DC or DBCTL environment, emergency restarting IMS (using the /ERESTART command)

You can restart z/OS and the IRLM on a different system. Also, batch backout and emergency restart can be done on any z/OS and IRLM in the data-sharing group.

For the non-failing systems:

• Locks are retained for the failed IMS.
• Locks are released as the failed IMS restarts and recovers its databases.
• Database buffers for the failed IMS are invalidated.
• Database error conditions are sent to the other IRLMs.