I have just closed a PMR with IBM on this subject. We have a dynamic process using UPDATE JOBNAME=*. It's worked for many years, but recently it failed big time, knocking out JES2.
The reason for the PMR was to get an explanation of DELAY= parameter on the UPDATE statement. It is not properly documented anywhere. It turns out that this SHOULD do the trick. The new linklist is switched in, but the old linklist is not freed for n seconds, plenty of time four outstanding I/O to complete. IBM still will not give a catagoric assurance about the safety of linklist updates using the feature though. I managed to recreate the problem at will on our test systems. When I used delay=, I got a different abend. IBM took a look at that abend and believe it won't happen with Z/OS 1.10 we are 1.9. So it might be safe with DELAY= and 1.10. The nice thing about our setup is that the dynamic linklist is built using the same PROG members as IPL time, so regression should not be an issue, or syntax errors. Happy to supply more detail directly if you want it. Joe Owens. ---------------------------------------------------------------------- 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
