Synchronizing IMS restart after a multiple IMS system failure

Multiple IMS™ system failures can occur in a data-sharing environment. When this happens, you might want to synchronize the startup of all of the failed IMS systems. Consider also synchronizing the startup of all IMS systems at a disaster recovery site.

If a sysplex failure or a restart at a disaster recovery site occurs, any new authorizations to databases or DEDB areas held by a failed IMS system are not granted until the failed IMS system is restarted. If each IMS system that failed is restarted one at a time and begins normal processing, authorization to databases or DEDB areas that were authorized by failed IMS systems will also fail. For full-function databases, message DFS047 is issued with a return code of X'09'. For DEDB areas, message DFS3709 is issued with a return code of X'09'.

After all of the failed IMS systems have been restarted, the databases and DEDB areas have to be manually started. Using the OPTION keyword and SYNCPLEX parameter on the /ERE command, you can serialize the restart of failed IMS systems. Issue the command on each IMS failed system.

After IMS restart processing completes, write-to-operator-with-reply (WTOR) message DFS3067A is issued. When all of the IMS systems receive this message, the operator can reply to each WTOR. Normal processing resumes, and database and DEDB area authorizations can be granted.

The OPTION SYNCPLEX keyword and parameter are valid only an /ERE command.