IBM Mainframe Discussion List <[email protected]> wrote on 12/19/2006 
10:52:44 AM:

> This did not happen to me, but the APAR sounds fairly serious.
> 
> Apparently there is a problem where ACTIVATE can cause overlays within 
the
> IOSAS address space.  The APAR includes the RESTART/BOOT/IPL flag.
> 
> The scope of the problem appears to be the latest z9 machines, and use 
of
> the alternate subchannel set, and a circumstance happening where the 
code
> needs to issue a message with the incorrect length set, and then 
overlaying
> storage.
> 
> I just hate for anyone to get into trouble (perhaps causing themselves 
an
> outage) using a "routine" dynamic change facility within z/OS.  For that
> reason, Heads Up.

  The "who it happened to" was one of our z9 test systems (one where
we do a lot of ACTIVATE testing, including very large changes affecting
many thousands of devices).
  After the first instance of the overlay, I turned on the (undocumented)
CSVRENTPROTECT trap.  And despite running the ACTIVATE back and forth
between the same two IODFs many times, it was several weeks before the
problem reoccurred and hit the trap.  So at least thus far, the problem
has not been highly pervasive.  But it is rather high impact (have to 
reIPL
to get IOSAS functions working), so we thought it prudent to flag the 
APAR as HIPER.  And if you intend to be doing ACTIVATEs which affect
PAV aliases in the alternate subchannel set (which requires a z9 and
at least z/OS 1.7), it would be prudent to install the PTF. 

Jim Mulder   z/OS System Test   IBM Corp.  Poughkeepsie,  NY

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