Text 411, 111 rader
Skriven 2007-01-23 18:13:32 av Mike Luther (1:117/3001.0)
Ärende: XR_C006 OS/2 beta Fix Pack
==================================
Questions abound ..

Not too long ago a curious IBM OS/2 XRC_006 beta 3 Fispack showed up in the
outside world.  Able to be looked at by whomever, as pointed toward for
whatever download by Dink, obviously it has/is attracting attention. Apparently
a fair number of people have tried it and other than one comment about JFS
issues for a larger than 500GB JFS volume, there doesn't seem to be any grief
involved.  Positive comment has appeared about solving the PMMERGE.DLL crashes
that seem to have been happening for some but not at all of the XR_C005 last
formal released IBM Fix Pack for MCP and so on.

The package includes the major fix files, a copy of the required CSF 144 level
Service package needed, as well as a curious floppy content for making a STRESS
group of OS/2 .CMD files I have never seen before.  The new Kernel Level in it
is a 1.05 level two steps above the 1.03 level that is in the XR_C005 release. 
From the README collection, included are kernel boot fixes for high speed AMD
and Intel CPU chips, plus a few other things.  And major changes in the LVM and
JFS operations.

From the Internet Usegroup postings, you'll note that Bob Eager offered a REXX
.CMD script for making the needed floppy diskettes for it.  As well, there has
been serious discussion my Chuck McKinnis about his participation in some of
the APAR's in it as well as from failure data supplied by Paul Smedley.  I've
decided to research this one with the most current box I have, the Intel
915GEVL motherboard, a 2.993 Mhz Intel DuoCore processor, the Intel just fine
Pro100 NIC on it, complete SATA hard drive operations using Dani's latest
driver to sub for the IBM1S506.ADD driver and the SBLive 5.1 audio using a mid
range UNIAUD interface which works fine with it and 1GB of memory.

What I know so far.  I've taken in all of the IBM official fixes off Passport
Advantage which included two LVM releases in 2006, as well as one JFS update in
2006.  But as Chuck McKinnis pointed out in the Usegroup discussion, there are
file differences between the LVM ad JFS files in the hard official releases for
this on P/A and what is in the latests curious beta FP XR_C006 that is out
there.  As well you *ABSOLUTELY* must have the last November 2005 release for
the IDEDASD fixpack for OS/2 off Passport Advantage, with a BLDLEVEL report for
it with the .161 level code to work with not only the FP6, but LVM and JFS
changes as well!

OK, all my boxes are at that IDE 161 level code anyway.  So I asked Chuck
McKinnis what would be the thought about adding in the LVM and JFS fixes before
adding the FP6 for test?  He suggested to add the FP6 first, then do the
LVMUPDT.EXE extract for the large international language fix, plus then the
JVMUPDT.EXE fix after that, and not to replace anything those discrete fixes
found which was later.  I wondered about that but did as suggested.

Duhh!  The discrete fix installers in both these LVM and JFS fixes went
automatically through everything thank you.  Wham, bam, thank you maam! And
when I looked at the logs they produced, all the necessary decisions were
automatically made for the install during the run!  Fascinating.

I made this research run, as said, on a completely cloned duplicate SATA drive
from the technique of an OS/2 Utility Diskette boot run, then using Jan Wijk's
DFSEE 7.1.5 to clone the whole hard drive to the test drive.

POOF!  The FP6 went in perfect, except!  I'd forgotten to think about the fact
that this was a licensed SNAP 3.1.8 video operation.  And all those video
drivers that were found by the FP6 showed later dates than the one in the fix
pack run!  So I told it no, don't overwrite all these video drivers.  Box
booted perfectly.  From there I did the LVM fix,rebooted as required.  Then I
did the JFS fix, rebooted as required. Neither one of these fixes even used the
LOCKED DEVICE DRIVER stuff!

This box has been running solid stable so far for several days now.  But it was
never one with the confusing XR_C005 PMMERGE.DLL lockups anyway for me.  And I
don't have any JFS operations at all here, thought the driver is loaded before
the HPFS.IFS driver in my CONFIG.SYS files anyway on all boxes.

I've given you folks the best guidelines I can about how to test this and
handle the README stuff to see what changes have been made in this curious
archive that is floating around out there now.  I ask one favor from you folks
here!

This VIDEO driver stuff!  Isn't the proper way to do this FP6 install really to
back out the SNAP 3.1.8 video and go back to the IBM generic VGA file
operations on which to do the install?  Then after that boots up go back and
reinstall the SNAP 3.1.8 driver pack?  That so that if you ever really have to
revert to the SETVGA game one way or another,whatever might have been in the
IBM changes in relation to the super high speed CPU changes and the LOADER
would perhaps not stumble should that ever be needed?  I dunno what changes
might have been in there that weren't already in the system anyway for XR_C005.


As a second thought, more complex here.  If anyone does play with this, would
you give some thought to all those interesting .CMD files in the STRESS archive
that is in there?  These seem to be related to a crew in Florida doing this for
IBM's testing use!  Is this part of the old Boca Raton IBM operation, contract
or otherwise that is still in there for testing Fix Packs?

Or is this in relation to whom the Fix Pack was being created for, which
doesn't really seem to be related to eCS, though Chuck McKinnis is sure helping
us maybe go forward into the future and for which I deeply appreciate his and
others' work?  Chuck noted to me that I should post what I found out about my
testing publicly in the Newsgroups.  OK,some of the really best OS/2 people out
there are still in Fido.  I don't see any difference between me and anyone else
out here now that whatever is released out here has surfaced however it got
there..

;)

Here is our opportunity?


--> Sleep well; OS/2's still awake! ;)

Mike @ 1:117/3001



--- Maximus/2 3.01
 * Origin: Ziplog Public Port (1:117/3001)