Peter Relson wrote:
By the time slip got around to scheduling the actual dump request,
that SRB had already run to completion
The idea's right, but one minor detail: SLIP will have "issued" SDUMP.
SDUMP will have done SUMLST processing, and SDUMP will have "scheduled" the
rest of the dump. All that will happen before the interrupted work unit
continues. It's the "scheduled rest of the dump" that might not have begun
before the SRB completed. Various other factors come into play related to
system and address-space non-dispatchability and the "quiesce" option of
the dump as to whether the SRB might or might not run to completion

We encountered something like this just yesterday...

We asked a customer to send us a dump taken via SDUMP issued from an FRR protecting one of our enclave SRBs. The FRR will SUSPEND until the capture phase ECB is posted. Unfortunately, DAE suppressed the dump. So, rather than tell DAE to "take" the next one, the customer decided to set an error-event SLIP instead. That resulted in lots of "missing" storage in the dump because, although SUMLST ensured many of our control blocks appeared "in-tact", the SYSTRACE showed that the SRB's FRR had already freed all of the "important" storage areas before any of it was dumped.

I usually answer "YES" when prompted by IPCS to use the summary dump data for dump access. After looking at the trace, I dropped the dump and replied "NO" when prompted again. This time I saw what things *really* looked like at the time the dump was finally taken. The SRB routine and all of its storage areas were long gone!

Needless to say, we've asked the customer to instruct DAE to "take" the dump and recreate.

--
Edward E Jaffe
Phoenix Software International, Inc
5200 W Century Blvd, Suite 800
Los Angeles, CA 90045
310-338-0400 x318
[EMAIL PROTECTED]
http://www.phoenixsoftware.com/

----------------------------------------------------------------------
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

Reply via email to