We have had problems with batch jobs doing multiple concurrent updates of
PDSE's. When the failure occurs the failing job holds a PDSE latch thus
causing other PDSE users to fail. Serializing the updates seems to have
resolved the problems. IBM recommended applying the fixes for APARs OA11965,
OA12189, and OA12657, we are also tracking OA12727. The following was cut
from our ETR with IBM and is their recommendation on resoving latch problems:
- Issue the V SMS,PDSE, ANALYSIS command against all systems in the plex
- Get a dump of SMSPDSE, MASTER, and any jobs holding the PDSE resource
(from the ANALYSIS output) on any systems that show problems
- Cancel the holding job(s), forcing them out if necessary
- Re-issue the ANALYSIS command. See if the PDSE resources are cleared
If so, resume operating and provide us the dump(s), syslog, and
detailed EREP for the system(s) that had problems
- If cancelling the job(s) doesn't free the PDSE resource(s), issue
the FREELATCH command using info from the ANALYSIS output
At this point, ANALYSIS shouldn't show any latch/lock contention. If
it still does, get another dump of the problem system and contact us.
Andy Corpes <[EMAIL PROTECTED]> wrote:
I would be grateful for any group members experiences with PDSE's, or any
links to documents describing issues or problems that may be encountered.
with them.
---------------------------------
Click here to donate to the Hurricane Katrina relief effort.
----------------------------------------------------------------------
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
