[mod.computers.vax] VMS4.4

JARRELLRA@VTVAX5.BITNET (Ronald A. Jarrell) (05/21/86)

Got mine thursday.  A tape, a mup, and an 80lb box of manual updates.
All people already on support from pre 4.4 vms also got the last free
shipment of fiche.  I'm afraid to look at the manuals. 2 new large binders,
2 small binders, and replacements for almost everything.  HUGE release notes.
The manuals are packed in a "magic" order to make it easy to update, which
means it took me 20 minutes to find the release notes, which is what I
wanted NOW, so I can read them before I put up the tape and upgrade my
manuals a month from now..  To save you some time, find the packet labeled:
"Introduction to Vax/VMS Document Set".  The release notes are the part
after the gray seperator pages in the back of that package.  The old
familiar SPD has been replaced with a 200 page master ordering document..
It *does* give the info you need, but takes so damn long to find it..
About 4 products are NOT supported currently under 4.4, ACMS 1.2, and
Vax Lisp being the two I can remember right now.  It's mandatory to build
a new standalong backup (I think most of us do anyway) and you have to
reinstall decnet.  Oh yes, for cluster sites, if you do a rolling upgrade,
you have to set VMSD1 to 1 so that 4.4 works in 4.3 compatability mode...

Sheesh, I hope it's worth it.

-Ron Jarrell

Tli@USC-ECLB.ARPA (Tony Li) (05/21/86)

Take your time doing that manual upgrade.  At least one of the RTL
updates is missing a few pages.  (RTL 640ish....)  Also, this is a
complete reorganization of your existing manuals.  It's quite a mess.
There are three new binders, which replace existing volumes.  You'll
need to keep the old binders though as they are needed for the
reorganization.  

As far as I can tell everything is included in the new set with the
single exception of the LAT Control Program manual.

For those of you Eagle fans, V4.4 includes a new DRDRIVER and VMB, so
we'll again be waiting for patches from Emulex/SI/other.

;-)

Tli@USC-ECLB.ARPA (Tony Li) (05/23/86)

More installation problems...

The upgrade will not work properly if the distribution is on your
system disk.  The file STARTUP.UP2 tries to mount your system disk
again, and when that fails, it bombs out of the upgrade.  Our fix was
to patch the file with SET NOON ... SET ON around the failing mount
(and dismount).

Also, we keep a copy of DCLTABLES.EXE in SYS$SPECIFIC.  This caused
trivial probelms.  The work-around is to blow that away before
installation.

Waiting for Eagle patches...
Tony ;-)

hobbit@ecsvax.UUCP (Derrell Piper) (05/28/86)

  Some more fun with VMS 4.4:  It seems that the meaning of the 
SET ACC/ENABLE qualifier has changed.  You must now explicitly specify
which keywords you would like enabled.  Kind of trivial, but a
definate GOTCHA.

  More importantly, Accounting seems to suicide every once and a while.
The symptom is that $ SHOW ACCOUNTING lists accounting as being enabled
while in fact no records are being logged to the accounting file.  The
Job Controller doesn't have the file open (as show by a $ SHOW DEV/FILE/NOSYS
on the system disk) and turning it off and back on doesn't help.  The only
solution was to stop all print queues and the queue manager, kill the
job controller and restart it using @SYS$SYSTEM:STARTUP JOBCTL.   

  I called the CO customer support center, and they said that they were
aware of the /ENABLE problem/feature, but they were not aware of Accounting
dying.  It has been SPR'd (I believe in them ;-), but I would appreciate
hearing from anyone else who has encountered this bug

  I second the person's comment about the VMS 4.4 doc set upgrade.  Be
very careful!  The cover letter is hard to find, and several manuals are
shrink wrapped together.  If you're not careful, you'll miss them.

Derrell Piper
120 Rosenau Hall (201H)
School of Public Health
University of North Carolina - Chapel Hill
Chapel Hill, NC 27514          (919) 966-5106

Bitnet:     derrell@uncsphvx.BITNET
Usenet:     ...decvax!mcnc!ecsvax!hobbit 

GOET@HWALHW5.BITNET (11/06/86)

Has anybody experience with running SPSS and/or SAS under VMS4.4. We want
to install 4.4, but not before we are sure that these products will work.

Kees Goet
Agricultural University, Wageningen.
Netherlands.
BITNET: GOET@HWALHW5

CP.PAVER@MCC.COM (Bob Paver) (11/10/86)

SAS works under VMS 4.4 but you must retain some runtime libraries from
VMS 4.2 or VMS 4.3 as an interim workaround.  SAS Institute says there
will be a maintenance release of SAS soon that will make if compatible
with VMS4.4.

The interim workaround does not significantly affect SAS use.  It involves
logicals pointing to old VMS runtime libraries which makes it necessary
to invoke SAS from within a DCL procedure.  I've also had trouble with 
VMS ASSIGN statements executed from with SAS programs.  SAS can't 
translate the logicals unless I use ASSIGN/JOB.  Don't know if this is
a bug or a feature!
-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Bob Paver	(512) 338-3316
Microelectronics and Computer Technology Corp. (MCC)
3500 West Balcones Center Drive
Austin, TX  78759

ARPA:  paver@mcc.com
UUCP:  {ihnp4,seismo,harvard,gatech}!ut-sally!im4u!milano!paver
-------

mic@NGP.UTEXAS.EDU (Mic Kaczmarczik) (11/11/86)

In article <12253823381.38.CP.PAVER@MCC.COM> CP.PAVER@MCC.COM (Bob Paver) writes:
>		[workaround for running SAS under VMS 4.4]
>                                         I've also had trouble with 
>VMS ASSIGN statements executed from with SAS programs.  SAS can't 
>translate the logicals unless I use ASSIGN/JOB.  Don't know if this is
>a bug or a feature!
>-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
>Bob Paver	(512) 338-3316
>Microelectronics and Computer Technology Corp. (MCC)
>3500 West Balcones Center Drive
>Austin, TX  78759
>
>ARPA:  paver@mcc.com
>UUCP:  {ihnp4,seismo,harvard,gatech}!ut-sally!im4u!milano!paver
>-------

This is a necessary side-effect of the SAS `VMS' command.  To execute
general VMS commands from within SAS, SAS spawns a DCL subprocess to
parse and execute the command.  This means that executing

	VMS 'ASSIGN BAR FOO';

inside SAS causes a logical name to be set in the *subprocess'*
logical name table -- SAS can't see it because the change isn't
propagated to the SAS process' own logical name table!  Using
ASSIGN/JOB works because all processes and subprocesses in a job share
the LNM$JOB table, so SAS can see the logical name.

Mic Kaczmarczik
U.T. Austin Computation Center User Services
mic@ngp.utexas.edu