On Thu, Dec 21, 2017 at 10:50 AM, Seymour J Metz <[email protected]> wrote:
> It's not just DYNALLOC, it's *any* code that explicitly or implicitly > waits. Who wrote the transaction that issued the SVC 99? Why didn't he run > it in a subtask? Presumably he is also doing OPEN, which would be a problem > even without the DYNALLOC. > Yeah, long ago we purchased a company. They had a rarely run transaction which used a COBOL subroutine which did an OPEN / READ / CLOSE on a sequential DSN. For some reason, when we were doing the processing, we ran the transaction a LOT. And exhausted the region; which resulted in a S80A (iirc) abend which took the CICS down. I "cheated" and updated the CICS SRT (System Recovery Table) to basically "do nothing" when an S80A happened. The transaction would then fail until we scheduled a recycle of the CICS. This "moved" the problem from a "CICS is broken" problem to "the transaction is broken" so that applications would fix it. > > > -- > Shmuel (Seymour J.) Metz > http://mason.gmu.edu/~smetz3 > > -- I have a theory that it's impossible to prove anything, but I can't prove it. Maranatha! <>< John McKown ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
