VMS 8.4-2L1 on AXPBox

[Bill Gunshannon]

I finally got AXPBox to build and figured this was a good time to
get my VMS System running.

No joy in Mudville....

Portion done: 0%...10%

%PCSI-E-OPENOUT, error opening
DISK$ALPHASYS:[VMS$COMMON.][SYSHLP]HELPLIB.HLB; a
s output
-RMS-E-FNF, file not found
%PCSI-E-OPFAILED, operation failed
Terminating is strongly recommended. Do you want to terminate? [YES]

I have tried it twice with the same result.

The second time I answered "NO' with the following result:

%PCSI-E-OPENIN, error opening
DISK$ALPHASYS:[VMS$COMMON.][SYSLIB]DCLTABLES.EXE;
as input
-RMS-E-FNF, file not found
%PCSI-E-OPFAILED, operation failed
Terminating is strongly recommended. Do you want to terminate? [YES]

tried "NO" again and it seems to be continuing, but even if it
finishes I can't believe it will run correctly with those files
missing.

Any suggestions?

bill

[Ben Lambert]

I have that same installation error, but it runs fine afterwards and the 'missing' files appear to be there.
Others have mentioned it before as well, with no indications of later issues:

https://groups.google.com/g/comp.os.vms/c/dUGJelByoUk/m/AFrHysL8AQAJ

HTH

Ben.

[Robert A. Brooks]

On 1/1/2021 1:28 PM, Bill Gunshannon wrote:
>
> I finally got AXPBox to build and figured this was a good time to
> get my VMS System running.
>
> No joy in Mudville....

[...]

> Any suggestions?

Alpha V8.4-2L1 was designed for upgrade only, not a fresh install.

V8.4-2L2 (which is Alpha-only) can be used for a fresh install.

Having said that, reports are that you can continue past those error messages
and have a working system.

--
-- Rob

[Bill Gunshannon]

OK. I'll let it continue. It is awfully slow, though. I seem
to remember it being a faster install on a VAX. :-)

The ISO is the one I got when I joined the VSI Community Program.
Did I somehow get the wrong file?

bill

[Dave Froble]

It is a bit sad that the V8.4 2L1 distribution has this problem.

So, what EXACTLY are you doing?

Where is your ISO distribution?

What is your emulated system's configuration?

--
David Froble Tel: 724-529-0450
Dave Froble Enterprises, Inc. E-Mail: da...@tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

[Bill Gunshannon]

On 1/1/21 3:11 PM, Dave Froble wrote:
> On 1/1/2021 1:50 PM, Bill Gunshannon wrote:
>> On 1/1/21 1:38 PM, Robert A. Brooks wrote:
>>> On 1/1/2021 1:28 PM, Bill Gunshannon wrote:
>>>>
>>>> I finally got AXPBox to build and figured this was a good time to
>>>> get my VMS System running.
>>>>
>>>> No joy in Mudville....
>>>
>>> [...]
>>>
>>>> Any suggestions?
>>>
>>> Alpha V8.4-2L1 was designed for upgrade only, not a fresh install.
>>>
>>> V8.4-2L2 (which is Alpha-only) can be used for a fresh install.
>>>
>>> Having said that, reports are that you can continue past those error
>>> messages
>>> and have a working system.
>>>
>>
>>
>> OK.  I'll let it continue.  It is awfully slow, though.  I seem
>> to remember it being a faster install on a VAX.  :-)
>>
>> The ISO is the one I got when I joined the VSI Community Program.
>> Did I somehow get the wrong file?
>>
>> bill
>>
>
> It is a bit sad that the V8.4 2L1 distribution has this problem.

Sadly, I'm on my fourth attempt and it is not looking
promising this time either.

>
> So, what EXACTLY are you doing?

Trying to install the version of VMS that I got from VSI for the
Community Program on AXPBox.

>
> Where is your ISO distribution?

Mounted as an emulated CDROM and it boots and starts the install
just fine.

>
> What is your emulated system's configuration?
>

With the exception of filenames for the disk images and a few
additional putty options to make the print easier to read it
is exactly the same as what is found at:

https://github.com/lenticularis39/axpbox/wiki/OpenVMS-installation-guide


bill

[Bill Gunshannon]

Portion done: 0%...10%

%PCSI-E-WRITEERR, error writing
DISK$ALPHASYS:[VMS$COMMON.][SYSLIB]DECW$SERVER_R
IDDGL.EXE;1
-SYSTEM-F-CTRLERR, fatal controller error
-SYSTEM-W-PFMBSY, page fault monitor is in use

%PCSI-E-OPFAILED, operation failed
Terminating is strongly recommended. Do you want to terminate? [YES]

Well, this one looks like an AXPBox problem. I really think this
stuff is just not ready for prime time yet.

Time to move on to another project. :-(

bill

[Timothy Stark]

That is IDE emulation problem. Try SCSI emulation instead. I successfully
installed OpenVMS Alpha on AXPbox.

Tim

_______________________________________________
Info-vax mailing list
Info...@rbnsn.com
http://rbnsn.com/mailman/listinfo/info-vax_rbnsn.com

[Bill Gunshannon]

On 1/1/21 4:07 PM, Timothy Stark wrote:
> That is IDE emulation problem. Try SCSI emulation instead. I successfully
> installed OpenVMS Alpha on AXPbox.
>

Damn, I hadn't thought of that. I have never used an IDE
disk on VMS, can't imagine why I should have been trying now.

I will try that now.

bill

[Dave Froble]

I guess I wasn't specific enough. Do you have emulated disks as files
on only one physical disk? If so, you're beating that disk rather
badly, and that could be some of your "slowness".

Do you have 8.4 distribution prior to 2L1? Maybe install that, then
upgrade to 2L1.

When I installed 2L1 several years ago on my AlphaServer 800 I used
multiple physical disks. Seemed to work fine for me.

However, I reverted to the disk with VMS V8.3, since I had some TCP/IP
problems (too lazy to work them out) and I have HP commercial license,
not this year to year bullshit.

[John H. Reinhardt]

Hi Bill,

As Robert says, the image given to Hobbyists for the CLP is the V8.4-2L1 which is designed to install over the HPE V8.4, hence the errors. There are two. One for HELPLIB and one for DCL. You can safely ignore both and continue with the install. Also as Robert says, it is the V8.4-2L2 which is designed to be a clean install or an upgrade install, however that's not the one VSI gives us. Go figure. It has been mentioned on the VSI OpenVMS forum but other than saying it's okay to continue, no action has been taken. I suppose we should make a little noise over there and ask for the newer version.

And yes, AXPbox is slow. It's lacking the JIT compile and other features of something commercial like Charon or the others.

--
John H. Reinhardt

[Bill Gunshannon]

I have done this for other kinds of systems. Trust me this is slow.
Someone else already said AXPBox is really slow. But then, speed
wasn't the problem. I started in this business when a compile was
a good time to go out for lunch.

>
> Do you have 8.4 distribution prior to 2L1?  Maybe install that, then
> upgrade to 2L1.

I have what VSI provides.

>
> When I installed 2L1 several years ago on my AlphaServer 800 I used
> multiple physical disks.  Seemed to work fine for me.

Someone has already explained that the two missing files is normal
and VMS should run without them. I don't know enough about VMS
internals but it seems to me the the first will mean no HELP
command and the second concerns me because I thought DCLTABLES
was important. I guess we'll see.

>
> However, I reverted to the disk with VMS V8.3, since I had some TCP/IP
> problems (too lazy to work them out) and I have HP commercial license,
> not this year to year bullshit.
>

This is my first serious venture away from VAX. We'll see how it goes.
It is installing now that I got rid of the IDE and went with SCSI. That
I should have figured out myself.


bill

[Bill Gunshannon]

Well, it installed. Next I need to try to set up the network
on AXPBox. But, at least I now have a current version of VMS
to play with. Wonder what Open Source tidbit needs porting?

bill

[Timothy Stark]

> And yes, AXPbox is slow. It's lacking the JIT compile and other features of something commercial like Charon or the others.

Yes, that AXPbox is slow without JIT recompilation. I told AXPbox developers about that. I recommend them to implement JIT recompilation.

I got 63.0 VUPs when run AXPbox on i7-8800K processor. That is same performance with fastest VAX 4000 and VAX 7000/9000 series.
I got 58 VUPs on SIMH VAX 8600 emulator.

CharonVAX runs at 800+ VUPs when JIT enabled according to google searches.

Tim

[Bill Gunshannon]

For anyone still following this and interested...

The system boots. I put the PAKs on. Tried a SHUTDOWN and
it gets to the point of disabling the console and then it hangs there.
Shutdown never happens. No matter how long I wait I can hit ^C and
be back at the $ prompt. An interesting situation I have never seen
on a VAX.

I get a lot of -E-FNF messages on startup, too. Interesting.
Wonder if any of the layered products are even installed at
this point.

bill

[Bill Gunshannon]

Interesting. After a reboot it now successfully shuts down on demand.

BUt, I see there are no layered products. No compilers or anything.
Anybody know what is available and how I can get them?

bill

[Jan-Erik Söderholm]

They should be on "the other" disk image(s). Since you are installing
using a traditional kit (not the pre-built VirtualBox kit), you need
one disk image for the OS installation and one (or two) with LP's.

Just as VMS has always traditionally been distrubuted.

My disk images for 8.4-2L2 are called:

ALPHA0842L2.ISO (8.4-2L2 Base OS)
AVMS842L2LP1.ISO (8.4-2L2 LP disk 1)
AVMS842L2LP2.ISO (8.4-2L2 LP disk 2)

I don't think I ever had any never have used any 2L1 so
I cannot say for sure if the disk names are/were the same...

You should be able to find them at the same download source as
where you fetched the base OS disk image.

[Bill Gunshannon]

I don't remember them being there when I got the base ISO but
they could have been. But then, my credentials would be expired
by now even if I remembered where to go. I suppose I could just
register again and see what happens but I expect that would make
more work for VSI.

bill

[Timothy Stark]

Very good! Be aware when mount disk or tape images with /system parameter.
If you do not mount with /system option and attempt to logout, AXPbox will
crash with unimplemented SCSI 0x1B (unload) command error. I opened tickets
with AXPbox developers last month.

-----Original Message-----
From: Info-vax <info-vax...@rbnsn.com> On Behalf Of Bill Gunshannon via
Info-vax
Sent: Saturday, January 2, 2021 8:37 AM
To: info...@rbnsn.com
Cc: Bill Gunshannon <bill.gu...@gmail.com>
Subject: Re: [Info-vax] VMS 8.4-2L1 on AXPBox

[Timothy Stark]

Where did you get ALPHA0842L2.ISO? I have only ALPHA0842L1.ISO. I checked VSI ftp site and it still have ALPHA842L1.ISO but have ALPHA0842L2LP1/2.ISO.

-----Original Message-----
From: Info-vax <info-vax...@rbnsn.com> On Behalf Of Jan-Erik Söderholm via Info-vax
Sent: Saturday, January 2, 2021 9:47 AM
To: info...@rbnsn.com
Cc: Jan-Erik Söderholm <jan-erik....@telia.com>
Subject: Re: [Info-vax] VMS 8.4-2L1 on AXPBox

[Timothy Stark]

When you access to VSI ftp site, look into LPS directory (folder) where Alpha ISO base images are.

-----Original Message-----
From: Info-vax <info-vax...@rbnsn.com> On Behalf Of Bill Gunshannon via Info-vax
Sent: Saturday, January 2, 2021 11:53 AM
To: info...@rbnsn.com
Cc: Bill Gunshannon <bill.gu...@gmail.com>

Subject: Re: [Info-vax] VMS 8.4-2L1 on AXPBox

[Chris Townley]

My initial credentials still work - go into the LPS directory

Chris

[John H. Reinhardt]

On 1/2/2021 10:52 AM, Bill Gunshannon wrote:

>
> I don't remember them being there when I got the base ISO but
> they could  have been.  But then, my credentials would be expired
> by now even if I remembered where to go.  I suppose I could just
> register again and see what  happens but I expect that would  make
> more work for VSI.
>
> bill
>

Unlike with HPE the VSI credentisls don't expire after a few weeks/months. Mine are still good from the initial email.

--
John H. Reinhardt

[Bill Gunshannon]

I dug around and found the original email. I now have the two
Layered Products ISO files. Thank you everyone.

Now, I have to figure out how to use them. :-)

Another interesting anomaly. I set the timezone during the install.
When I rebooted it went back to GMT. I ran the script to set the
timezone and made it EST5EDT again. It still gives me GMT when I
do SHOW TIME. What did I miss?

bill

[Robert A. Brooks]

On 1/2/2021 2:11 PM, Bill Gunshannon wrote:
> On 1/2/21 1:41 PM, John H. Reinhardt wrote:
>> On 1/2/2021 10:52 AM, Bill Gunshannon wrote:
>>
>>>
>>> I don't remember them being there when I got the base ISO but
>>> they could  have been.  But then, my credentials would be expired
>>> by now even if I remembered where to go.  I suppose I could just
>>> register again and see what  happens but I expect that would  make
>>> more work for VSI.
>>>
>>> bill
>>>
>>
>> Unlike with HPE the VSI credentisls don't expire after a few weeks/months.
>> Mine are still good from the initial email.
>>
>
> I dug around and found the original email.  I now have the two
> Layered Products ISO files.  Thank you everyone.
>
> Now, I have to figure out how to use them.  :-)

Use LD to mount the .iso.

--
-- Rob

[Bill Gunshannon]

Actually, I just mounted the iso as a cdrom. It was the procedure I
wasn't sure I remembered. But I did. :-)

Of course, this brought up another interesting difference from my
beloved VAX. When I was having the original problem I reduced the
disk size thinking that could be the APXBox problem. The example
I was following used a 5GB disk. I reduced it to 1GB. After all,
my VAX could run just fine with 500MB, right? Afer installing VMS
I went to install the C Compiler. "Insufficient Disk Space"!! Wow.

Oh well. Back to the beginning and install it all again using the
bigger disk. Sigh....

bill

[Jan-Erik Söderholm]

Should not need to do a full re-install.
Create and mount a bigger disk.
Backup/image from your current system disk to the new one.
Reboot using the new disk as the system disk.

[Bill Gunshannon]

It's easier to just start a new install (mostly because I don't
know how to do the copy off the top of my head) and go watch some
football. :-)

Sadly, here's another new one. I have done 6 installs so far and never
got this error. Is this one going to be a show stopper?


Portion done: 20%...30%...40%...50%...60%
%PCSI-I-PRCOUTPUT, output from subprocess follows ...
%SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual
address=000000000026
C104, PC=000000007FDC5324, PS=0000001B

%PCSI-E-MODREPLERR, error replacing module INIT$SHR in library
DISK$ALPHASYS:[VM
S$COMMON.][SYSLIB]IMAGELIB.OLB
-SYSTEM-F-ACCVIO, access violation, reason mask=!XB, virtual
address=!XH, PC=!XH
, PS=!XL

%PCSI-E-OPFAILED, operation failed
Terminating is strongly recommended. Do you want to terminate? [YES]

bill

[Timothy Stark]

Did you include DECnet OSI? I tried that and got same error. I tried DECnet IV instead and made it through.

Later I found AXPbox install guide and it explains about that error and simply answer 'no' and continue...

Tim

-----Original Message-----
From: Info-vax <info-vax...@rbnsn.com> On Behalf Of Bill Gunshannon via Info-vax
Sent: Sunday, January 3, 2021 4:25 PM
To: info...@rbnsn.com
Cc: Bill Gunshannon <bill.gu...@gmail.com>
Subject: Re: [Info-vax] VMS 8.4-2L1 on AXPBox

[Bill Gunshannon]

On 1/3/21 5:01 PM, Timothy Stark wrote:
> Did you include DECnet OSI? I tried that and got same error. I tried DECnet IV instead and made it through.

I told the install to include both of them. Not sure which I would
use but then, I may want to connect with my RSTS/E system. :-)

>
> Later I found AXPbox install guide and it explains about that error and simply answer 'no' and continue...

Well, that's what I did guess we'll see what happens.

bill

[Hunter Goatley]

On 1/1/2021 7:14 PM, Bill Gunshannon wrote:
> But then, speed
> wasn't the problem.  I started in this business when a compile was
> a good time to go out for lunch.

When I first started working for Clyde Digital Systems in February 1987,
I was surprised to see several bookcases full of science fiction and
fantasy novels scattered around the programmers' area.

I understood why when I learned that the six developers were all using
one VAX 11/730 for all development work. The 11/730 was painful for a
single user, but multiple builds? The guys would start a build, then
kick back with a novel and read for an hour or two.

I couldn't handle that for long, so I started---without
permission---using the company's business-side VAX 11/750 for my
development work. That flew by comparison, but was still a far cry from
the VAX 11/785 I had been used to using at WKU.

Hunter

[Phillip Helbig (undress to reply)]

In article <85a1ebc7-8eff-577a...@goatley.com>, Hunter

Goatley <goath...@goatley.com> writes:

> When I first started working for Clyde Digital Systems in February 1987,
> I was surprised to see several bookcases full of science fiction and
> fantasy novels scattered around the programmers' area.
>
> I understood why when I learned that the six developers were all using
> one VAX 11/730 for all development work. The 11/730 was painful for a
> single user, but multiple builds? The guys would start a build, then
> kick back with a novel and read for an hour or two.
>
> I couldn't handle that for long, so I started---without
> permission---using the company's business-side VAX 11/750 for my
> development work. That flew by comparison, but was still a far cry from
> the VAX 11/785 I had been used to using at WKU.

Calculations I used to do in an overnight batch job on a VAXstation 3100
I now do interactively on a DS10. But other things I now do in
week-long batch jobs on the DS10, though they would be maybe an hour on
Itanium and perhaps on x86 I could do those interactively as well.

Yesterday while cleaning up around the house I found an add for a 15 MB
disk drive for $2000 or something, and another, somewhat newer, for a 60
MB disk drive "for when 30 MB just isn't enough" for about $3500. True,
that was before my computing time, but not much more than 20 years ago I
bought a 4-GB disk for about $400. Now a GB costs about 10¢. A GB of
RAM will cost more, but much cheaper than the DM 100 per MB we
calculated about 25 years ago.

[Marc Van Dyck]

It happens that Hunter Goatley formulated :

Yup... to the point that, on my first job, in order to avoid builds
during the day (we had one every night), for small modifications, we
would compile /MACHINE_CODE, and then use the compiler listing to PATCH
the modification directly in the existing executable. Was way faster,
even considering mistakes...

--
Marc Van Dyck