Dealing with CICS failures

If one of your CICS® regions fails, all BTS processes on the failing region are halted. Processes on other regions might also become stuck, because expected events are not generated. If a CICS region fails, you must perform an emergency restart.

Only in rare circumstances - for example, if the CICS global catalog or system log is corrupted - might it be necessary to perform an initial or cold start after a failure. If it is necessary, perform a cold start in preference to an initial start.

Emergency starts

During an emergency restart, CICS automatically restores BTS processes to the state they were in before the failure. Any activities that were active at the time of the failure are rerun. That is, if an activation (transaction) was running, it is backed out and restarted. The activity is sent the same reattachment event that caused the failed activation. Its data-containers contain the same data they held at the start of the failed activation.

Initial and cold starts

During an initial or cold start:
  • BTS repository data sets are unchanged.
    Note: Repository data sets are never reinitialized at CICS startup, because they might be shared.
  • The local request queue data set is unchanged. All information about BTS timers, pending, and unserviceable requests is preserved. However, it is likely that some of this information is now irrelevant or invalid, because it refers to processes that no longer exist.


dfhp94j.html | Timestamp icon Last updated: Monday, 5 June 2017