Finding the reason for a DELAY request not completing
If your preliminary investigations have not shown the reason for the wait, you need to look in greater detail at the evidence available.
Before you begin
About this task
Procedure
- Take a system dump, and format it using the keywords CSA, ICP, and AP. These get you the common system area, the interval control program control blocks, and the task control areas, respectively. You might also find information given by the formatting keywords KE (kernel storage areas, including the calling sequence for each task), DS (dispatcher task summary, including details of suspended tasks), and TR (internal trace table) to be useful.
- Locate field CSATODTU in the CSA and make a note of its value. It is the current CICS time of day in internal ‘timer units’.
- Locate the TCA for your task, and read the value of field TCAICEAD. This gives you the address of the interval control element for your task. Use this information to find the ICE (interval control element) for the task, and look at field ICEXTOD. Make a note of its value.
Results
- Your task either did not make the DELAY request you expected, or the interval specified was longer than intended. This could indicate a user error. Check the code of the transaction issuing the request to make sure it is correct.
- Your task’s delay request was not executed correctly. This might indicate an error within CICS code, or a corrupted control block.
If the value of ICEXTOD is less than CSATODTU, the ICE has already expired. The associated task should have resumed. This indicates that some area of storage might have been corrupted, or there is an error within CICS code.