On Apr 10, 2013, at 5:00 PM, Gerhard Postpischil wrote:
On 4/10/2013 8:05 AM, Peter Relson wrote:
As all should understand, very little of this would be considered
supported in any way shape or form and if anything in this realm
caused a
problem (or could conceivably have caused a problem), IBM service
might
take a hard line about helping.
As a couple of other comments state, user's relations with IBM have
been much of a love-hate relationship.
1) On an early OS/360 system, our resident S.E. (remember those)
spent some agonizing time tracking a system crash, and finally
found some non-reentrant code in DUO. She was very gracious about it.
2) When IBM split out the communications task, they left an
extraneous FREEMAIN that got exercised only when a MOUNT command
was issued with a seven character or longer serial (without the
comm task, the system couldn't even be shut down). It took just
over a year to get the official fix, even though I included a one-
byte ZAP with the original problem report.
3) A new version of the CoBOL compiler produced an 0C4 in execution
of a large program (something about needing more than four base
registers?). IBM had a fix in less than a week.
We had resource accounting and billing before SMF, tape library
code before UCC-1 and TLMS, security before RACF, and were not the
only installation that found IBM's code to be insufficient for our
needs. Many of the features in the current system are due to
complaints, suggestions, and contributions from users and user
groups, and IBM service needs to understand that zOS, despite the
OCO policy, will never assume appliance status.
---------
We too had an full time SE in the early days of MVS. We still printed
dumps and his desk was always piled 6 ft (or over) of many many dumps.
We had a couple of vicious "users" that were constantly crashing our
system(s). One especially nasty problem left Q4 issues dangling and
our MSS suffered because (on count of?) Allocation finally got their
act together after hundreds of PTF's It was definitely an IBM issue.
Our other PITA was a system that ran under CICS that could and did
crash the entire system. It took 8 months of IPL's until the guy
pinned on the CICS "application" when he finally caught them with
their pants down and he pinned it on them conclusively.
He left us I think for the JES3 team in CA. I think he was tired of
finding bugs. I lost track of him so I never heard how he did.
When we were able to point to a culprit we went after them as we were
getting blamed for the outages and it was fun to see others squirm
for a change, as usually we were in the hot seat.
Our DUO issue was of a MP nature (2 PSA's) and DUO would use PSA as a
storage location to put jobnames that ran under DUO. Our group was
able to create a "fix" where it would zap out jobnames that were no
longer running. It took the DUO people a while to fix this although
we had a semi work around it shouldn't have worked that way.
Ed
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
