On 3/13/2017 7:00 AM, Barbara Nitz wrote:
Apparently we ran into another of those PDSE latch contention bugs. While
terminating the holder of the latch, we had to force the batch job since cancel
had absolutely no effect.
We got messages
IEF743I INIT FORCED - CODE SA22 - IN ADDRESS SPACE 0040
IEF743I AUTO39P FORCED - CODE SA22 - IN ADDRESS SPACE 0040
$HASP310 AUTO39P TERMINATED AT END OF MEMORY
$HASP310 INIT TERMINATED AT END OF MEMORY
According to the books, IEF743I means "System action: The job and address space
end." Unfortunately that does not seem to be the case. There is no indication that I
can find in hardcopy log that the address space (Initiator) got restarted, and the dump
that I took 17 minutes later shows the address space still up with ASCBEWST (address
space last lost control) about 1s after $HASP310 INIT TERMINATED AT END OF MEMORY. I
do see the initiator restarted 25 minutes after.
I don't have a batch job that is non-cancelable, so I cannot test the force
command on a batch job running in an initiator. Does anyone know for sure how
memterm processing due to force goes for an initiator? Does the initiator
really terminate?
Our problem turned from Latch contention into ENQ contention, with that INIT
holding a vital ENQ and not releasing it.
Thanks and regards, Barbara
Hi Barbara,
PDSE's causing issues? Say it ain't so! Did you try a $PI for the
offending initiator before the FORCE? I've never had to FORCE an
initiator before. Hopefully you got some dumps and sent this to IBM
with a PMR. I'd be interested to see how this works out.
Regards,
Tom Conley
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
