Dealing with CICS failures

If one of your CICS® regions fails, not only are all BTS processes on the failing region halted, but processes on other regions may also become stuck, because expected events are not generated.

If a CICS region fails, you should perform an emergency restart.

Only in very rare circumstances--for example, if the CICS global catalog or system log is corrupted--should 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

At an emergency restart, CICS automatically restores BTS processes to the state they were in prior to the failure. Any activities that were active at the time of the failure are re-run. 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

At an initial or cold start:

Related concepts
Recovery and restart in BTS
Related tasks
Dealing with stuck processes
Dealing with activity abends
Dealing with unserviceable requests
[[ Contents Previous Page | Next Page Index ]]