[bit.listserv.vmxa-l] 8904 sup .....

SPGJAF@UCBCMSA.BITNET (Joseph A. Faracchio {415} 642-7638 {w}) (02/14/90)

Sometime back someone asked about 8904SUP and I said we were on and stable.
Well .....    I just put this on VMSHARE ...
Well at this point I can't recommend 8904SUP.    We are up to our ears
in alligators (CP Abends).  I'll try to be brief in the history but this
is gonna end up long.

We were on 8902SUP experiencing a couple of abends (IOL002 and PRG005)
and a lot of abends (DSP003).  Level 2 gave us a trap for the DSP003's and
we shipped them an abend with the trap on.  Then we went to 8904SUP.

8904SUP was very stable for our MVS/370 and HPO 4.2 guests.  Then we migrated.

When we migrated all our CMS users to the XA level,     things started getting
interesting.   In fact I'd say I'm living the Chinese curse of interesting times
First we would abend/hang and (as noted previous) not be able to get a dump.
This was the PSW loop invalid that required a clear storage (O7) to RE-IPL.
It seems a SYSTEM RESET (O3) would have enabled a standalone dump but we
were never sure the registers would still be valid.  We eliminated this problem
by cold starting and migrating all spool files via RSCS. (I.E. SPTAPE problem)

With the above fixed we started experiencing IOL008 abends - about 4-6 of em
in just a few days.  This was fixed by adding a second part to the circumvention
recommended by Dick Newson on the list & PROB VMXASP2 for bug4 - module HCPVMS

Then we had various and assorted FRF002's (the new Sel007's).  Some were
due to "user errors" i.e. we weren't retrieving ACNT records fast enough.
One was "fixes in hiding" (not in the bucket) for MDC (35933) and one we have
a circumvention while they shuffle the logic - raise RSMXT+8 from 2 to x'10'.

We       expect to       see FRF002's again but IBM seems to be better
managing them with changes to whole setup.

BUT NOW for the WHOPPER:  We're experiencing PRG004's or  FRE016's on a
BI-DAILY basis (about every other day.)  They seem to like to alternate.

And in between these, we're still seeing 1 (so far) PTF004 which I think Dick
Newson's latest recommendation for HVC may fix. And the DSP003's have returned.
And we had one more IOL002 which I think is related to our fiber optic
channel extender and/or I/O errors on 3274s on MVS at the other end.

IBM's recommendations (besides praying) is to replace Dick's fixes
        Bug1 - module HCPUDU - vm38883 (and pre: 37417)
        Bug2 - module HCPGAL - vm38741
        Bug3 - module HCPVMS - vm39345
        Bug4 - module HCPVSP - vm39929

Only 38883 is on now.   38741 is 'on the way', 39345 has just arrived
and 39929 is not closed.
Plus I have a circumvention that will retry a few times before going down
with a Fre016 (although I've seen 2 more since I've put it on, no PRG4's)
They've accepted an APAR for our problems (my very own APAR) VM40436.

One other problem has been printer hangs til next IPL after a FLUSH
command.  Fixed by : VM37797

That's 5 FRE016's, 4 PRG004's, 6 FRF002's, 1 IOL002, 4 IOL008's, 1 PTF004.
and 1 DSP003 since we went to 8904SUP on Jan 14, 1990.  (and a DSP002 for prt)

The last two weeks have been a bit of a nightmare and I was wondering:
Is anybody else on 8904SUP besides me and Dick Newson?????? :) ????????
cheers (as in bottoms up) ... joe.f.