CICS has stalled during a run
If a CICS region that has been running normally stalls, so that it produces no output and accepts no input, the scope of the problem is potentially system-wide. The problem might be confined exclusively to CICS, or it could be caused by any other task running under MVS.
Look first on your MVS console for any messages. Look particularly for messages indicating that operator intervention is needed, for example to change a tape volume. The action could be required on behalf of a CICS task, or it could be for any other program that CICS interfaces with.
If there is no operator action outstanding, inquire on active users at the MVS console to see what the CPU usage is for CICS. If you find the value is very high, this probably indicates that a task is looping. Read Dealing with loops for advice about investigating the problem further.
- The system definition parameters are not suitable for your system.
- The system is short on storage, and new tasks cannot be started. This situation is unlikely to last for long unless old tasks cannot, for some reason, be purged.
- The system is at one of the MXT or transaction class limits, and no new tasks can be attached. In such a case, it is likely that existing tasks are deadlocked, and for some reason they cannot be timed out.
- There is an exclusive control conflict for a volume.
- There is a problem with the communications access method.
- There is a CICS system error.
The way you can find out if any of these apply to your system is dealt with in the information that follows. For some of the investigations, you will need to see a system dump of the CICS region. If you do not already have one, you can request one using the MVS console. Make sure that CICS is apparently stalled at the time you take the dump, because otherwise it will not provide the evidence you need. Format the dump using the formatting keywords KE and XM, to get the storage areas for the kernel and the transaction manager.