VM system disk arrangement

[Robert O'Hara]

I am looking at reworking the arrangement of disks in the VM 5-pack system.

I have looked at the current disk utilization of MAINT, and I believe that all of his minidisks can comfortably fit on a single 3350.  Also, all of the VM system disks (PERM, TDSK, TEMP) can fit on a single 3350.

Now of course on a real mainframe we would want these disks spread out over several channels, but it seems to me that with Hercules it does not matter.  Is that a correct assumption?

I propose:

  disk 1: PERM, TEMP, TDSK
  disk 2: all of MAINT's disks
  disk 3: OPERATOR, RSCS, RSCS1, CPWATCH, AUTOLOG1, CMSBATCH,
          ALGOL, BREXX, COBOL, DOSVS, OSBASIC, PLI, PL360, RPG,
          SCRIPT, TAPEMAC, WATFIV
  disk 4: GCCCMS
  disk 5: REGINA, CMSUSER

Comments on this layout?

Thanks, Bob

[George-111]

Robert,

I have seen many layouts. Some ranging from what you talk about to the
other extreme where each minidisk is a separate device. Each has it;s
benefits. From a performance aspect, I would separate the system files
across multiple drives. Since 99% of MAINT's disks are rarely used I see
no problem with them all on one disk.

With Hercules ability to create disk devices that do not use the full
allocation (compressed disk) quickly, there is no advantage to packing
everything on a few devices.

George

Robert O'Hara wrote:

>
>
>
> I am looking at reworking the arrangement of disks in the VM 5-pack system.
>
> I have looked at the current disk utilization of MAINT, and I believe
> that all of his minidisks can comfortably fit on a single 3350. Also,
> all of the VM system disks (PERM, TDSK, TEMP) can fit on a single 3350.
>
> Now of course on a real mainframe we would want these disks spread out
> over several channels, but it seems to me that with Hercules it does not
> matter. Is that a correct assumption?
>
> I propose:
>
> disk 1: PERM, TEMP, TDSK
> disk 2: all of MAINT's disks
> disk 3: OPERATOR, RSCS, RSCS1, CPWATCH, AUTOLOG1, CMSBATCH,
> ALGOL, BREXX, COBOL, DOSVS, OSBASIC, PLI, PL360, RPG,
> SCRIPT, TAPEMAC, WATFIV
> disk 4: GCCCMS
> disk 5: REGINA, CMSUSER
>
> Comments on this layout?
>
> Thanks, Bob
>
>

[Dave Wade]

Robert O'Hara wrote:
> I am looking at reworking the arrangement of disks in the VM 5-pack system.
>
> I have looked at the current disk utilization of MAINT, and I believe that all of his minidisks can comfortably fit on a single 3350.  Also, all of the VM system disks (PERM, TDSK, TEMP) can fit on a single 3350.
>

Not sure. Not done the sums...

> Now of course on a real mainframe we would want these disks spread out over several channels, but it seems to me that with Hercules it does not matter.  Is that a correct assumption?
>

If you have several disks then spreading over several disks will help,
if you have RPS. Also I think existing layout is based on sample directs
supplied with original system...

> I propose:
>
>   disk 1: PERM, TEMP, TDSK
>   disk 2: all of MAINT's disks
>   disk 3: OPERATOR, RSCS, RSCS1, CPWATCH, AUTOLOG1, CMSBATCH,
>           ALGOL, BREXX, COBOL, DOSVS, OSBASIC, PLI, PL360, RPG,
>           SCRIPT, TAPEMAC, WATFIV
>   disk 4: GCCCMS
>   disk 5: REGINA, CMSUSER
>
> Comments on this layout?
>

You still need both disk1 and disk 2 to ipl a working system. I think I
would try and make sure I had enough stuff on sysres to ipl a single pack...

> Thanks, Bob
>
>
>
>
> ------------------------------------
>
> Yahoo! Groups Links
>
>
>
>

[Thomas Kern]

I like the idea of a single iplable volume without any of the
non-operations minidisks and then spreading out the non-operational
minidisks across more non-filled compressed Hercules volumes.

Disk 1: Iplable: MAINT(190,191,19E), OPERATOR, AUTOLOG1, CPWATCH,
        Some Temp, Tdisk
Disk 2: Rest of MAINT's minidisk
Disk 3: RSCS, RSCS1, CMSBATCH, Other SVMs, CMSUSER
ALGOL, BREXX, COBOL, DOSVS, OSBASIC, PLI, PL360,
        RPG, SCRIPT, TAPEMAC, WATFIV
Disk 4: GCCCMS
Disk 5: ALGOL, BREXX, COBOL, DOSVS, OSBASIC, PLI, PL360,
        RPG, SCRIPT, TAPEMAC, WATFIV
Disk 6: REGINA
Disk 7: Temp
Disk 8: Tdisk
Disk 9: etc

I also like putting the Waterloo Mods onto a separate volume.

/Tom Kern

Robert O'Hara wrote:

> I am looking at reworking the arrangement of disks in the VM 5-pack system.
>
> I have looked at the current disk utilization of MAINT, and I believe
> that all of his minidisks can comfortably fit on a single 3350. Also,
> all of the VM system disks (PERM, TDSK, TEMP) can fit on a single 3350.
>
> Now of course on a real mainframe we would want these disks spread out
> over several channels, but it seems to me that with Hercules it does not
> matter. Is that a correct assumption?
>
> I propose:
>
> disk 1: PERM, TEMP, TDSK
> disk 2: all of MAINT's disks
> disk 3: OPERATOR, RSCS, RSCS1, CPWATCH, AUTOLOG1, CMSBATCH,
> ALGOL, BREXX, COBOL, DOSVS, OSBASIC, PLI, PL360, RPG,
> SCRIPT, TAPEMAC, WATFIV
> disk 4: GCCCMS
> disk 5: REGINA, CMSUSER
>
> Comments on this layout?
>
> Thanks, Bob

[pfg504]

The best arrangement would be to keep each product on it's own 3350 (could be less than 555 cyls). This would allow development to replace one disk for a given product rather than to replace everything.

As far as an operational system, keep user stuff separate from system stuff, keep perm away from temp(SPOOL, PAGE, TDISK).

The goal would be to update the smallest amount of disk as possible for any given change.

Paul

--- In H390-VM@..., Thomas Kern <tlk_sysprog@...> wrote:

>
> I like the idea of a single iplable volume without any of the
> non-operations minidisks and then spreading out the non-operational
> minidisks across more non-filled compressed Hercules volumes.
>
> Disk 1: Iplable: MAINT(190,191,19E), OPERATOR, AUTOLOG1, CPWATCH,
> Some Temp, Tdisk
> Disk 2: Rest of MAINT's minidisk
> Disk 3: RSCS, RSCS1, CMSBATCH, Other SVMs, CMSUSER
> ALGOL, BREXX, COBOL, DOSVS, OSBASIC, PLI, PL360,
>         RPG, SCRIPT, TAPEMAC, WATFIV
> Disk 4: GCCCMS
> Disk 5: ALGOL, BREXX, COBOL, DOSVS, OSBASIC, PLI, PL360,
>         RPG, SCRIPT, TAPEMAC, WATFIV
> Disk 6: REGINA
> Disk 7: Temp
> Disk 8: Tdisk
> Disk 9: etc
>
> I also like putting the Waterloo Mods onto a separate volume.
>
> /Tom Kern
>
> Robert O'Hara wrote:
> > I am looking at reworking the arrangement of disks in the VM 5-pack system.
> >
> > I have looked at the current disk utilization of MAINT, and I believe
> > that all of his minidisks can comfortably fit on a single 3350. Also,
> > all of the VM system disks (PERM, TDSK, TEMP) can fit on a single 3350.
> >
> > Now of course on a real mainframe we would want these disks spread out
> > over several channels, but it seems to me that with Hercules it does not
> > matter. Is that a correct assumption?
> >
> > I propose:
> >
> > disk 1: PERM, TEMP, TDSK
> > disk 2: all of MAINT's disks
> > disk 3: OPERATOR, RSCS, RSCS1, CPWATCH, AUTOLOG1, CMSBATCH,
> > ALGOL, BREXX, COBOL, DOSVS, OSBASIC, PLI, PL360, RPG,
> > SCRIPT, TAPEMAC, WATFIV
> > disk 4: GCCCMS
> > disk 5: REGINA, CMSUSER
> >
> > Comments on this layout?
> >
> > Thanks, Bob
>

[sqrfolkdnc]

This (many volumes)has the advantage of making it easier to have a test  
(clone of Disk 1) and a production system when making changes, and the
test system can even be run second level if there is a reason to do so
under Hercules, other than for demo purposes.

Having a lot of busy stuff on one volume can still be a performance
issue if CP has to queue up i/os because it knows it can't do two I/Os
to the same disk at the same time.

Thomas Kern wrote:

> I like the idea of a single iplable volume without any of the
> non-operations minidisks and then spreading out the non-operational
> minidisks across more non-filled compressed Hercules volumes.
>
> Disk 1: Iplable: MAINT(190,191,19E), OPERATOR, AUTOLOG1, CPWATCH,
> Some Temp, Tdisk
> Disk 2: Rest of MAINT's minidisk
> Disk 3: RSCS, RSCS1, CMSBATCH, Other SVMs, CMSUSER
> ALGOL, BREXX, COBOL, DOSVS, OSBASIC, PLI, PL360,
>         RPG, SCRIPT, TAPEMAC, WATFIV
> Disk 4: GCCCMS
> Disk 5: ALGOL, BREXX, COBOL, DOSVS, OSBASIC, PLI, PL360,
>         RPG, SCRIPT, TAPEMAC, WATFIV
> Disk 6: REGINA
> Disk 7: Temp
> Disk 8: Tdisk
> Disk 9: etc
>
> I also like putting the Waterloo Mods onto a separate volume.
>
> /Tom Kern
>
> Robert O'Hara wrote:
>  
>> I am looking at reworking the arrangement of disks in the VM 5-pack system.
>>
>> I have looked at the current disk utilization of MAINT, and I believe
>> that all of his minidisks can comfortably fit on a single 3350. Also,
>> all of the VM system disks (PERM, TDSK, TEMP) can fit on a single 3350.
>>
>> Now of course on a real mainframe we would want these disks spread out
>> over several channels, but it seems to me that with Hercules it does not
>> matter. Is that a correct assumption?
>>
>> I propose:
>>
>> disk 1: PERM, TEMP, TDSK
>> disk 2: all of MAINT's disks
>> disk 3: OPERATOR, RSCS, RSCS1, CPWATCH, AUTOLOG1, CMSBATCH,
>> ALGOL, BREXX, COBOL, DOSVS, OSBASIC, PLI, PL360, RPG,
>> SCRIPT, TAPEMAC, WATFIV
>> disk 4: GCCCMS
>> disk 5: REGINA, CMSUSER
>>
>> Comments on this layout?
>>
>> Thanks, Bob
>  

[Robin Brooks-2]

--- In H390-VM@..., "Robert O'Hara" <rohara@...> wrote:
>
> I am looking at reworking the arrangement of disks in the VM 5-pack
system.
>
> I have looked at the current disk utilization of MAINT, and I believe
that all of his minidisks can comfortably fit on a single 3350. Also,
all of the VM system disks (PERM, TDSK, TEMP) can fit on a single 3350.
>
> Now of course on a real mainframe we would want these disks spread out
over several channels, but it seems to me that with Hercules it does not
matter. Is that a correct assumption?

>
> I propose:
>
> disk 1: PERM, TEMP, TDSK
> disk 2: all of MAINT's disks
> disk 3: OPERATOR, RSCS, RSCS1, CPWATCH, AUTOLOG1, CMSBATCH,
> ALGOL, BREXX, COBOL, DOSVS, OSBASIC, PLI, PL360, RPG,
> SCRIPT, TAPEMAC, WATFIV
> disk 4: GCCCMS
> disk 5: REGINA, CMSUSER
>
> Comments on this layout?
>
> Thanks, Bob
>

Could you include FORTRAN G or H as well please

[kerravon86]

--- In H390-VM@..., "Robin Brooks" <robin_brooks@...> wrote:
>
> Could you include FORTRAN G or H as well please

If you don't mind me asking, are you doing something
in particular with Fortran?

I don't think I've seen someone show an interest in
that before.

BFN.  Paul.

[Thomas Kern]

I am always interested in FORTRAN (first good computer language I
learned and all that). I still have the FORTRAN G compiler from VM/370
R6 on my production z/VM 5.3 system at work. I have had to some programs
in FORTRAN there because of the lack of other compilers. I even have the
Scientific Subroutine Package so I can do some statistics and charting.
If there is a free FORTRAN H compiler, I want it.

/Tom Kern

kerravon86 wrote:

>
>
>
> --- In H390-VM@... <mailto:H390-VM%40yahoogroups.com>,
> "Robin Brooks" <robin_brooks@...> wrote:
>>
>> Could you include FORTRAN G or H as well please
>
> If you don't mind me asking, are you doing something
> in particular with Fortran?
>
> I don't think I've seen someone show an interest in
> that before.
>
> BFN. Paul.

[Dave Wade]

I tried to get RATFOR to work (the original Fortran one) and failed...

Dave Wade G4UGM
Illegitimi Non Carborundum

> -----Original Message-----
> From: H390-VM@... [mailto:H390-VM@...] On Behalf
> Of kerravon86
> Sent: 10 June 2009 20:47
> To: H390-VM@...
> Subject: [H390-VM] Re: VM system disk arrangement
>
> --- In H390-VM@..., "Robin Brooks" <robin_brooks@...> wrote:
> >
> > Could you include FORTRAN G or H as well please
>
> If you don't mind me asking, are you doing something
> in particular with Fortran?
>
> I don't think I've seen someone show an interest in
> that before.
>
> BFN.  Paul.
>
>
>
>
> ------------------------------------
>
> Yahoo! Groups Links
>
>
>

[Robert O'Hara]

--- In H390-VM@..., "Robin Brooks" <robin_brooks@...> wrote:

> Could you include FORTRAN G or H as well please

Robin,

The FORTRAN G compiler is on the current 5-pack system, on the Y (19E) disk.

Bob

[Gregg C Levine-2]

Hello!
I am going to take my position clearly.

If it works, leave it alone. Whatever approach all of you decide upon and
test via your own arrangements, and decide to appropriately (Whatever that
is.) distribute will be okay.

I'll then post a note here (in the group) regarding where the appropriate
files can be found.

--
Gregg C Levine hansolofalcon@...
"The Force will be with you always." Obi-Wan Kenobi
 


> -----Original Message-----
> From: H390-VM@... [mailto:H390-VM@...] On Behalf
Of

> Dave Wade
> Sent: Wednesday, June 10, 2009 4:33 PM
> To: H390-VM@...
> Subject: RE: [H390-VM] Re: VM system disk arrangement
>
> I tried to get RATFOR to work (the original Fortran one) and failed...
>
> Dave Wade G4UGM
> Illegitimi Non Carborundum
>
>
> > -----Original Message-----
> > From: H390-VM@... [mailto:H390-VM@...] On Behalf
> > Of kerravon86
> > Sent: 10 June 2009 20:47
> > To: H390-VM@...
> > Subject: [H390-VM] Re: VM system disk arrangement
> >
> > --- In H390-VM@..., "Robin Brooks" <robin_brooks@...> wrote:
> > >
> > > Could you include FORTRAN G or H as well please
> >
> > If you don't mind me asking, are you doing something
> > in particular with Fortran?
> >
> > I don't think I've seen someone show an interest in
> > that before.
> >
> > BFN.  Paul.

[Kevin Leonard-2]

> If there is a free FORTRAN H compiler, I want it.

Tom:

The OS/360 21.8 FORTRAN H compiler is at:

http://kleonard.home.att.net/vm/fthcms.zip

in TAPE DUMP format.  I think I mentioned it when candidate
products were being suggested for the five-pack system, but
it hasn't gotten included.

The existing build assumes you put the FORTRANH MODULE and
IEKAA00 MODULE files on the Y disk.  If you prefer other
alternatives, the source FORTRANH ASSEMBLE and object text
for IEKAA00 is present so you can build your own.

Assembling/compiling all the IEKAA00 source was a serious
challenge, and I ended up delinking the OS load modules to
get object text.

There were problems with the 21.8 FORTRAN H compiler, but
it's the only version we've got as far as I know.  The only
problem I definitely remember was sporadic incorrect base
register selection with OPT=2.  I haven't seen any problems
as long as OPT=0, but that's just my experience.

--
Kevin

[kerravon86]

--- In H390-VM@..., Thomas Kern <tlk_sysprog@...> wrote:
>
> I am always interested in FORTRAN (first good computer language I
> learned and all that). I still have the FORTRAN G compiler from
> VM/370 R6 on my production z/VM 5.3 system at work.

> I have had to [compile] some programs
> in FORTRAN there because of the lack of other compilers.

What sort of environment only has FORTRAN as an
option? And what would you write in if a more
normal suite of options was available?

> If there is a free FORTRAN H compiler, I want it.

Looks like you're in luck there. It's amazing what
pops out of the blue hanging around these lists.

One question though - are either of these Fortran 4?

BFN.  Paul.

[Thomas Kern]

The kind of environment where VM was bought by the users and forced upon
the Data Center. I was hired to babysit the existing MVT/MVS sysprogs to
make sure they did not screw up the VM/PROFS system. They promoted all
development on MVS and only allowed what the users needed for PROFS on
the VM system. I was allowed to load up my private utilities for sysprog
use ONLY, no end-users allowed to use them. So my choice of programming
languages were AssemblerF, EXEC, EXEC2, REXX and my private copies of
FORTRAN and PLI. I also loaded Waterloo Script for me to use instead of DCF.

As much as I like REXX, it is an interpretive language (we did not have
the REXX compiler). So when I needed to write a program to read,
analyze, summarize a Spool File backup tape, I wrote it in FORTRAN. I
redid it in PLI a year later but it used more CPU time. I think the
PLIOPT or VS/PLI could have been better, but I don't have those in my
magic bag of tricks.

I prefer prototyping in REXX and then compiling it. If more efficiency
is needed, then VS/PLI and if even more efficiency is needed, then
FORTRAN or Assembler depending on size (assembler for small stuff).

/Tom Kern


kerravon86 wrote:

> --- In H390-VM@... <mailto:H390-VM%40yahoogroups.com>,
> Thomas Kern <tlk_sysprog@...> wrote:
>>
>> I am always interested in FORTRAN (first good computer language I
>> learned and all that). I still have the FORTRAN G compiler from
>> VM/370 R6 on my production z/VM 5.3 system at work.
>
>> I have had to [compile] some programs
>> in FORTRAN there because of the lack of other compilers.
>
> What sort of environment only has FORTRAN as an
> option? And what would you write in if a more
> normal suite of options was available?
>
>> If there is a free FORTRAN H compiler, I want it.
>
> Looks like you're in luck there. It's amazing what
> pops out of the blue hanging around these lists.
>
> One question though - are either of these Fortran 4?
>
> BFN. Paul.

[kerravon86]

--- In H390-VM@..., Thomas Kern <tlk_sysprog@...> wrote:
>
> the REXX compiler). So when I needed to write a program to read,
> analyze, summarize a Spool File backup tape, I wrote it in FORTRAN. > I redid it in PLI a year later but it used more CPU time.
> I think the
> PLIOPT or VS/PLI could have been better, but I don't have
> those in my magic bag of tricks.

Ok, what is in your magic bag of tricks now, and
would Fortran still be selected from your new
bag in the same circumstance? As we don't have a
rexx compiler for VM/370-380 either.

And can you explain what these Fortrans have in
relation to Fortran 4?

BFN.  Paul.

[Dave Wade]

kerravon86 wrote:

> --- In H390-VM@..., Thomas Kern <tlk_sysprog@...> wrote:
>> the REXX compiler). So when I needed to write a program to read,
>> analyze, summarize a Spool File backup tape, I wrote it in FORTRAN. > I redid it in PLI a year later but it used more CPU time.
>> I think the
>> PLIOPT or VS/PLI could have been better, but I don't have
>> those in my magic bag of tricks.
>
> Ok, what is in your magic bag of tricks now, and
> would Fortran still be selected from your new
> bag in the same circumstance? As we don't have a
> rexx compiler for VM/370-380 either.
>
> And can you explain what these Fortrans have in
> relation to Fortran 4?
>

I would guess they are both "Fortran IV", but at that time the standards
were perhaps more lax and informal. As a compiler Fortran G requires
considerably less resource to compile a program, but Fortran H will
produce more efficient code.

Remember we also have Watfiv with is more or less "load and go"....

> BFN.  Paul.

[Fran Hensler]

Hi Tom -

Before we had VM/SP3 installed in 1984 we did all of our FORTRAN work
in DOS with FORTRAN F and Waterloo WATFIV.  I had the Scientific
Subroutine Package and the BMD & BMDP packages.  I conveted the DOS source
library for BMDP to CMS and did a TAPE DUMP of it.  The files are dated 1978.

I scanned a few of the BMDP programs and there are no copyright
statements so I guess the programs can be redistributed.  If there is
interest in these I could make a VMARC package or create TAPE DUMP on
an .AWS tape.  I have lost the manual for the Biomedical Programs.

/Fran Hensler at Slippery Rock University of Pennsylvania USA for 46 years
    mailto:fjh@...  http://zvm.sru.edu/~fjh  +1.724.738.2153
              "Yes, Virginia, there is a Slippery Rock"
--------------------------------------------------------------------------



On Wed, 10 Jun 2009 16:01:54 -0400 Thomas Kern said:

>I am always interested in FORTRAN (first good computer language I
>learned and all that). I still have the FORTRAN G compiler from VM/370
>R6 on my production z/VM 5.3 system at work. I have had to some programs
>in FORTRAN there because of the lack of other compilers. I even have the
>Scientific Subroutine Package so I can do some statistics and charting.
>If there is a free FORTRAN H compiler, I want it.
>
>/Tom Kern
>
>kerravon86 wrote:
>>
>>
>>
>> --- In H390-VM@... <mailto:H390-VM%40yahoogroups.com>,
>> "Robin Brooks" <robin_brooks@...> wrote:
>>>
>>> Could you include FORTRAN G or H as well please
>>
>> If you don't mind me asking, are you doing something
>> in particular with Fortran?
>>
>> I don't think I've seen someone show an interest in
>> that before.
>>
>> BFN. Paul.
>

[Tony Harminc]

2009/6/10 kerravon86 <kerravon86@...>:

> What sort of environment only has FORTRAN as an
> option? And what would you write in if a more
> normal suite of options was available?

Normal? In the 1960s through most of the 80s, FORTRAN was *the*
scientific and engineering language. What else would you write in?
When I was in 1st year engineering in 1972, FORTRAN was the only
language taught to engineers, except for the Engineering Science
(advanced) students, who were taught PL/I. Computer Science types were
also taught Algol (60), and of course various real and imaginary
machine and assembler languages.

BTW, the university I attended and later worked at did not have a
COBOL compiler installed at any time from 1972 to 1980, as far as I
know. Who would want such a thing?

>> If there is a free FORTRAN H compiler, I want it.
>
> Looks like you're in luck there. It's amazing what
> pops out of the blue hanging around these lists.
>
> One question though - are either of these Fortran 4?

FORTRAN G and H are both FORTRAN IV. FORTRAN II and earlier were
pretty much gone by 1970 or so.

Tony H.

[Thomas Kern]

Thanks Fran,

I would enjoy adding those to my collection.

/Tom

Fran Hensler wrote:

> Hi Tom -
>
> Before we had VM/SP3 installed in 1984 we did all of our FORTRAN work
> in DOS with FORTRAN F and Waterloo WATFIV. I had the Scientific
> Subroutine Package and the BMD & BMDP packages. I conveted the DOS source
> library for BMDP to CMS and did a TAPE DUMP of it. The files are dated 1978.
>
> I scanned a few of the BMDP programs and there are no copyright
> statements so I guess the programs can be redistributed. If there is
> interest in these I could make a VMARC package or create TAPE DUMP on
> an .AWS tape. I have lost the manual for the Biomedical Programs.
>
> /Fran Hensler at Slippery Rock University of Pennsylvania USA for 46 years
> mailto:fjh@... <mailto:fjh%40zvm.sru.edu>
> http://zvm.sru.edu/~fjh <http://zvm.sru.edu/~fjh> +1.724.738.2153
> "Yes, Virginia, there is a Slippery Rock"