For the record, I am at z/OS 1.10. Following up on my own note with wisdom from
IBM in the PMR:
"This behavior is documented in APAR OA31042 which is also HIPER and marks PTF
UA50730 in error. The hung rebuild problem you experienced was introduced by
this PTF and we can see in the dump of xxxx that nucleus module IXLR2MAN is at
the UA50730 level. You can check other system levels of this module by issuing
IP VERBX NUCMAP and doing F for IXLR2MAN to find the address and IP L
address_you_find len(x'1000'). The PTF level will be there. This will enable
you to see where else you might require the fix. This UA50730 introduced a
serious problem. This problem has the potential to any or many lock structures
or serialized list structures. If that happens you need to reset the hung
connector. In the case of some connectors some may be easier than others to
recycle. For instance, were this to occur to ISGLOCK, you'd be looking at a
re-ipl for affected systems.
In this case you may be able to recover by HALTing VTAM long enough for VTAM to
disconnect from the structure from each of the systems that
have outstanding rebuild requests.
I would urge you to apply a fix for OA31042 as soon as possible because it can
affect not only rebuild initiated by the operator but also
system managed as the result of a CF failover."
--------------------------------------------------------------------------------------------------
Many thanks to the excellent IBM support on this issue!
We were able to recover without an IPL using the suggested "HALT of VTAM"
method on the affected LPARs. This allowed the structure REBUILD to complete
its phase successfully and the VTAMs, etc. were restarted. I will be
implementing the suggested fix this weekend.
Hopefully this note may help someone else avoid experiencing this situation.
Bob
-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of
Richards, Robert B.
Sent: Monday, September 13, 2010 4:12 PM
To: [email protected]
Subject: Duplexing ISTGENERIC structure locked up sysplex
Just a heads up to those contemplating doing a duplex of ISTGENERIC. When I
switched to the policy containing the duplex, all hell broke loose.
It is currently caught in REBUILD status and it is preventing "generic" logons
to all systems in the SYSPLEX. If you are already logged on, you are fine.
All attempts to free it using SETXCF commands have been unsuccessful.
I will probably end up performing a sysplex-wide IPL in a few hours to resolve
the situation.
If anyone has experienced duplexing problems with any other IBM MVS components,
I would appreciate if you would drop me a note.
Bob
----------------------------------------------------------------------
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
