"Consider using the FORCE command as a last resort when the CANCEL command still fails to perform its function after you have issued it several times."
"WARNING: Never use the FORCE command without understanding that: - After issuing FORCE, you might have to re-IPL." "ARM The system is to terminate the specified job, time-sharing user, or started procedure if it is non-cancellable. If the FORCE ARM command fails to terminate the address space within a reasonable time, reissue FORCE with the ARM parameter. The ARM parameter executes normal task termination routines without causing address space destruction. The system rejects this parameter if the address space for the specified job, time-sharing user, or started procedure cannot be terminated or should be terminated via the CANCEL command. If the command still fails after several attempts, try issuing FORCE without the ARM parameter. Note: This keyword is not related to the ARMRESTART parameter and the functions of the automatic restart manager." --------- All of the above is out of: z/OS MVS System Commands SA22-7627-13 starting on page 4-284. ARM is on 4-286. (THIS ONE IS PDF, not BOOKTRASHED). >From years gone by, FORCE is the command of last resort. If you have to use FORCE, you better to be ready to IPL. And after all the nice work that BCP/RTM personnel have done to solve many of those issues, we just had to IPL last week after a FORCE was used on a JOB (initiator controlled as opposed to STC). This occurred on a z/OS 1.7 sysplex system where a JOB would not cancel and terminate. FORCE was used. We had to IPL because JES2 got knotted up dealing with the spooled output from the job, DUMP processing seemed to not function quite right (for EVERYONE). My personal opinion was the "C" code that we have to deal with for "P-Thread Create". It was that section (IBM's Code I believe) that did not accept the CANCEL and terminate gracefully. But I do not work on that area of our product (at this point) and so can only provide my observations based on the messages. But to continue with your argument: If I remember the logic correctly, FORCE actually terminates ALL control blocks in the address space (NOT including the ones in COMMON storage such as the ASCB). And it doesn't play nice about it. One moment you have page and segment tables, the next you don't. And I think the ASCB gets marked as NON-REUSE. I think the comment about error recovery also includes the DUMP TASK taking either SDUMP or SYSMDUMP - the DUMP address space can become locked, or ENQues on certain system resources may not get released. But it has been a very long time since I played with RTM at that level. Regards, Steve Thompson <snip> I'm having a disagreement with IBM Support for a product. I need to quote chapter and verse (of doc that may not even exist) on the use of FORCE. I have found a situation that requires use of FORCE if parts of Unix System Services are brought down prior to shutting the product down. The product can be canceled (but not stopped) up to the point it gets an EC6 abend; once it gets an EC6 abend, a cancel is ignored. Level 2 says the problem is in the order we are doing our shutdown. I contend he is right but that doesn't justify need of the FORCE. <snip> ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
