> I understand that IBM provides a multi-release co-existence support policy 
> for z/OS and other ServerPac products, which very adequately covers much of 
> our IBM inventory, and that IBM will fix any problems discovered.  But, have 
> you actually encountered any problems caused by co-execution of multiple 
> maintenance levels or releases of IBM software?  BTW, IBM transaction and 
> database software is out of the scope of this question for us, because they 
> are maintained separately. 

For years we have used a phased maintenance approach in our sysplexes, which 
consist of 3 systems at the most. Sometimes we have run for 1 week, sometimes 
for 4 weeks with different levels of maintenance or release in the same 
sysplex. It requires to really be up to speed with coexistence maintenance as 
indicated by IBM (and sometimes even by vendors).

Yes, we have had problems in that approach. The most severe was when we went 
from 1.6 to 1.8 (I think) in the sandplex. We had just IPL'd 1.8, it was up for 
the night, and in the morning I was calmly informed by the HMC that IXC402D: 
the 1.8 system was not responding anymore. I took an sadump and had (another) 
bloody row with IBM to get their bug fixed. It turned out that we didn't have 
an apar on the 1.6 system that *wasn't* flagged coexistence. As a result, the 
1.8 system went all over some GRS control blocks. Add to that a stupid (user) 
error in rollout that caused every TSO logon other than the sysprog logon to 
fail, which in turn send a memterm task in master address space to looping back 
on itself because that memterm routine *thought* it had gotten control in its 
'home' address space instead of h<>p<>s. After reading level1/2 the riot act (I 
was given the 'sympathy sickness'-spiel), they eventually got someone to look 
at my sadump who knew what she was doing. 

The less severe problems in that phased approach were easily fixable. The most 
funny was when IBM told me that we were at different MAS levels and I should 
apply a coexistence ptf when in reality we were at the exact same (hihgher) 
level everywhere. In that case it turned out that IBM was right and I was 
right. IBM fixed their apar text to include the 'all systems at the same level' 
description in the apar.

In all, give due diligence, I don't think there are problems with phased (IBM) 
maintenance. For vendors, I have seen one that uses SMP/E but has no clue what 
a prereq across different fmids is. So I would never roll out *their* 
maintenance phased, but then, their software is not sysplex in scope.

Barbara Nitz

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to