开云体育

On Mon, Mar 20, 2023 at 07:56 PM, Gregg Levine wrote:

Drew? I know you don't, (that is work for them), but what did happen this time?

I'm not sure, and I prefer not to describe what little I do know of the internal configuration due to security.

Suffice to say a configuration got reset during a power related reboot (I honestly don't know what rebooted!)

-ahd-

Hello!
Drew? I know you don't, (that is work for them), but what did happen this time?
-----
Gregg C Levine gregg.drwho8@...
"This signature fought the Time Wars, time and again."

On Mon, Mar 20, 2023 at 8:56?PM Drew Derbyshire <swhobbit@...> wrote:

If you're trying to log into the LCM+L VM/SP machine and can't, the Highs Elves of the LCM+L are aware of the problem; they working on it. It may take a couple weeks due to an Elf central to the solution being out of town.

(I do check remote access to the VM/SP image daily, but my powers are limited at best.)

Your humble VM/SP valet,
-ahd-

p.s. All Together now: I do not work for the LCM+L, and I do not speak for them.

And this message was found attached to a recorder buoy somewhere in
the Wild Space areas, and was thought to have been released by an
about to be destroyed Imperial Cruiser.

If you're trying to log into the LCM+L VM/SP machine and can't, the Highs Elves of the LCM+L are aware of the problem; they working on it. It may take a couple weeks due to an Elf central to the solution being out of town.

(I do check remote access to the VM/SP image daily, but my powers are limited at best.)

Your humble VM/SP valet,
-ahd-

p.s.?All Together now: I do not work for the LCM+L, and I do not speak for them.

> Strange... Isn't TODENABLE just to allow the guest to *set* the TOD
> clock? Anyone can read it. Do you actually need to set it?

Kevin explained in the OS360 group:
"MVT, unlike OS/VS systems, doesn't maintain a local offset from the TOD clock; that is, every SET command tries to set the clock directly, not just an offset.? And VM/370 treats the Set Clock instruction as a no-op:? it appears to work but doesn't do anything."

So, I assumed I was attempting to set the TOD clock when I replied with the date and clock values for the IEE166A message.? Any "SET DATE/CLOCK" commands I try are also ineffective:

- D T????????????????????????????????? ?
? IEE136I??? TIME=10.55.56?? DATE=83.068
- T DATE=95.066??????????????????????? ?
? IEE118I SET PARAMETER(S) ACCEPTED??? ?
- D T????????????????????????????????? ?
? IEE136I??? TIME=10.56.05?? DATE=83.068
- T CLOCK=01.01.01???????????????????? ?
? IEE118I SET PARAMETER(S) ACCEPTED?
- D T????????????????????????????????? ?
? IEE136I??? TIME=11.31.32?? DATE=83.068???
- T DATE=95.066,CLOCK=01.01.01???????? ?
? IEE118I SET PARAMETER(S) ACCEPTED??? ?
- D T????????????????????????????????? ?
? IEE136I??? TIME=11.00.50?? DATE=83.068

(and yes, I noticed the time changed on the "SET CLOCK", just not in obviously predictable way :-) )

I remember hitting the "TOD clock enable" button on our 360/50 whenever we changed the value, so that is probably the equivalent of VM/ESA's TODENABLE.

> It'll be interesting to see of the bug I encountered 20 years ago will
> go away if, as Fish suggests, you try LEGACYSENSEID, or if it's
> elsewhere in VM/ESA.

I am running with LEGACYSENSEID enabled.? It does not appear to help, so it looks like the bug is still there.? What I have is a P390 image, so that makes sense.? I've encountered the loop at IPL time if the devices are preconfigured in Hercules, Otherwise, if I attach them later, the loop occurs at "vary online" time.

> Does a Y2K version imply that it'll support 4-digit years? That sounds
> like a lot of fiddly work.

Kevin's comment in hercules-os360 says that it will support dates above 1999.? He didn't mention 4-digit years, so I'm not sure.

Yep, I run with it in all my images.? You showed it to me years ago when I had issues sharing 2314s.

On Thu, 9 Mar 2023 at 19:40, Jeff Snyder <jsnyder1369@...> wrote:

I tried testing this on VM/ESA and ran into some problems.

When defining the guest in USER DIRECT, I gained the TODENABLE option, however I lost
REALTIMER and ISAM, which the manual for VM/SP says are required for MVT.

Strange... Isn't TODENABLE just to allow the guest to *set* the TOD
clock? Anyone can read it. Do you actually need to set it?

When I IPLed VM with the 2314 and 3330 DASD in the Hercules configuration, VM appeared
to go into a loop sensing those devices (high CPU utilization, no other activity).

Heh... I remember around 1999-2000 on a P390 (actually an Integrated
Server, but really the same thing), defining some 3330s (allowed by
the P390 config), and firing up VM/ESA. It went into a loop! I
complained to IBM (back when it was easy to do so), and once past the
"why would you possibly want to define a 3330?" etc. I was told that
they had removed most but not quite all of the support for such old
devices from VM/ESA, and that it was known to loop, but since such
devices were no longer supported and couldn't be found on any Real
Iron, they wouldn't fix it.

(The reason I wanted to have a 3330 on VM/ESA was because I had some
DDR tapes of 3330s that I wanted to restore, and this was before the
Hercules era,)

I started VM without them in the configuration and manually attached them to Hercules. I
then tried using "SET RDEVICE" to indicate the devices should not sensed and still couldn't
get them online.

set rdevice 4150 type dasd mdc off shared no notsensed
16:26:42 HCPZRP6722I Created RDEV for device 4150.
16:26:42 1 RDEV(s) specified; 0 RDEV(s) changed; 1 RDEV(s) created
Ready; T=0.01/0.01 16:26:42
vary online 4150
16:27:57 HCPCPN6895I Device 4150 cannot be varied online because the device
16:27:57 HCPCPN6895I identification data is inconclusive.
16:27:57 HCPCPN6785E Unable to identify device 4150 dynamically.
16:27:57 1 device(s) specified; 0 device(s) successfully varied online

It'll be interesting to see of the bug I encountered 20 years ago will
go away if, as Fish suggests, you try LEGACYSENSEID, or if it's
elsewhere in VM/ESA.

It looks like MVT under VM/ESA is going to be a non-starter or at least more work than it's worth if Kevin Leonard has a Y2K version of MVT on the way.

Does a Y2K version imply that it'll support 4-digit years? That sounds
like a lot of fiddly work.

Tony H.

Jeff Snyder wrote:

[...]

vary online 4150
16:27:57 HCPCPN6895I Device 4150 cannot be varied online because
the device
16:27:57 HCPCPN6895I identification data is inconclusive.
16:27:57 HCPCPN6785E Unable to identify device 4150 dynamically.
16:27:57 1 device(s) specified; 0 device(s) successfully varied
online

Does your Hercules configuration file specify the LEGACYSENSEID option?

*

--
"Fish" (David B. Trout)
Software Development Laboratories

mail: fish@...

I tried testing this on VM/ESA and ran into some problems.

When defining the guest in USER DIRECT, I gained the TODENABLE option, however I lost REALTIMER and ISAM, which the manual for VM/SP says are required for MVT.

When I IPLed VM with the 2314 and 3330 DASD in the Hercules configuration, VM appeared to go into a loop sensing those devices (high CPU utilization, no other activity).? I started VM without them in the configuration and manually attached them to Hercules.? I then tried using "SET RDEVICE" to indicate the devices should not sensed and still couldn't get them online.

?set rdevice 4150 type dasd mdc off shared no notsensed
16:26:42 HCPZRP6722I Created RDEV for device 4150.
16:26:42 1 RDEV(s) specified; 0 RDEV(s) changed; 1 RDEV(s) created
Ready; T=0.01/0.01 16:26:42
?vary online 4150
16:27:57 HCPCPN6895I Device 4150 cannot be varied online because the device
16:27:57 HCPCPN6895I identification data is inconclusive.
16:27:57 HCPCPN6785E Unable to identify device 4150 dynamically.
16:27:57 1 device(s) specified; 0 device(s) successfully varied online

It looks like MVT under VM/ESA is going to be a non-starter or at least more work than it's worth if Kevin Leonard has a Y2K version of MVT on the way.

On Thu, 9 Mar 2023 at 17:45, Jeff Snyder <jsnyder1369@...> wrote:

Hi Chris,

It looks like that gets us outside MVT's sanity check range:

| 00 IEE116A TOD CLOCK INVALID- REPLY WITH SET PARAMETERS
| R 00,'CLOCK=01.01.01,DATE=55.066,AUTO=N'
| IEE600I REPLY TO 00 IS; 'CLOCK=01.01.01,DATE=55.066,AUTO=N'
IEE306I SET INVALID NUMERICS
*01 IEE116A TOD CLOCK INVALID- REPLY WITH SET PARAMETERS

The year 1955 cannot be represented in a TOD clock using the MVT epoch
of 1960, so - although the wording is odd - it doesn't surprise me
that it won't accept it.

But I don't remember how the TOD clock was integrated into the OS/360
(MVT) timing scheme. Keep in mind that OS/360 used the interval timer
for *all* timing, since that's the only timer there is on S/360, and
presumably could in theory handle any 2-digit years. With S/370 IIRC
the interval timer continued to be used for all timing, with the
newfangled TOD clock optionally used to initialize the timer.

Tony H.

Hi Chris,

It looks like that gets us outside MVT's sanity check range:

| 00 IEE116A TOD CLOCK INVALID- REPLY WITH SET PARAMETERS?? ?
| R 00,'CLOCK=01.01.01,DATE=55.066,AUTO=N'????????????????? ?
| IEE600I REPLY TO 00 IS; 'CLOCK=01.01.01,DATE=55.066,AUTO=N'
? IEE306I SET????? INVALID NUMERICS???????????????????????? ?
?*01 IEE116A TOD CLOCK INVALID- REPLY WITH SET PARAMETERS???

On Wed, Mar 8, 2023 at 12:23 AM, Bertram Moshier wrote:

I'm wondering; is it possible to automate VM/CMS file editing a file by using KEdit on Windows, instead.

In 2008, Robert O'Hara wrote a small Windows program to achieve that purpose. It is called VMconsole.exe. It is essentially a modified telnet client.

You connect to VM/370 using it with a command like this:

vmconsole.exe 127.0.0.1 3270

It will started a 3215 connection to VM/370. You logon to your virtual with it and you edit a file with the command:

redit fn ft [ fm | A ]

This will transfer the CMS file to Windows and open the Windows local copy with Notepad. After you save the local copy and quit Notepad, it will be sent back to CMS to replace the current file.

I just tried it, replacing Notepad with KEdit, and it seems to work.

If you want to try it on your side, download the ZIP archive VMconsole.zip from the Files section of this group.

The ZIP contains vmconsole.exe, vmconsole.cfg (in which you may change notepad par kedit), and two EXECs (REDIT and WINMSG) that you need to transfer to your VM/370 virtual machine, either via the card reader or the tape vmconsole.aws.

Cheers,

Rene FERLAND, Montreal

P.S. -- This program was written before EE was made available.

Jeff,

It looks most odd, as the time is also corrupted.

What happens if you change the order of the reply thus:-

R 0,'DATE=55.066,CLOCK=01.01.01,AUTO=N'

(that's the format in my notes for MVT).

Chris
--
<cjar1950@...>



----------------------------------------------------------------------------------------------------------------------------------
On Wed, 08 Mar 2023 15:09:31 -0800
"Jeff Snyder" <jsnyder1369@...> wrote:

Hi Uwe.? Using 28 years set the date back to 1995, which should have fit within MVT's sanity check for dates.? When I did that, MVT thought it was 1955.? I'm afraid if I go back any further, I might drop out of the early end of MVT's date range.

Using "YROFFSET -28":

VM says it's 1995
q t
14:47:30 TIME IS 14:47:30 PST TUESDAY 03/07/95
14:47:30 CONNECT= 00:04:23 VIRTCPU= 000:00.25 TOTCPU= 000:02.42

However the MVT guest thinks it's 1955!

| 00 IEE116A TOD CLOCK INVALID- REPLY WITH SET PARAMETERS
| R 00,'clock=01.01.01,date=95.057,auto=n'
| IEE600I REPLY TO 00 IS; 'clock=01.01.01,date=95.057,auto=n'
IEE118I SET PARAMETER(S) ACCEPTED
- D T
*IEE136I??? TIME=01.42.24?? DATE=55.066

* By the way, using 28 years works fine when running MVT directly under Hercules.? It's only when I try it under VM that it has issues.

Hi Uwe.? Using 28 years set the date back to 1995, which should have fit within MVT's sanity check for dates.? When I did that, MVT thought it was 1955.? I'm afraid if I go back any further, I might drop out of the early end of MVT's date range.

Using "YROFFSET -28":

VM says it's 1995
?q t
14:47:30 TIME IS 14:47:30 PST TUESDAY 03/07/95
14:47:30 CONNECT= 00:04:23 VIRTCPU= 000:00.25 TOTCPU= 000:02.42

However the MVT guest thinks it's 1955!

| 00 IEE116A TOD CLOCK INVALID- REPLY WITH SET PARAMETERS?? ?
| R 00,'clock=01.01.01,date=95.057,auto=n'????????????????? ?
| IEE600I REPLY TO 00 IS; 'clock=01.01.01,date=95.057,auto=n'
? IEE118I SET PARAMETER(S) ACCEPTED???????????????????????? ?
- D T?????????????????????????????????????????????????????? ?
? IEE136I??? TIME=01.42.24?? DATE=55.066?

By the way, using 28 years works fine when running MVT directly under Hercules.? It's only when I try it under VM that it has issues.

Tony,

Unfortunately, this option is not available to VM/SP5.? However, I have a VM/ESA image available to me, I may try looking over there and see if this is an option.? I'm afraid that if I get too current on the VM, it may not support MVT any longer.

Hi Fish, thanks for the input.? Unfortunately, this approach has a couple drawbacks that limit its usability for me:

1.? When submitting the jobs in the fashion you describe, they are not presented to MVT until the punch is closed.? As soon as that happens, all the jobs are read in, EOF is sent, and I have to restart the reader.

2. This would not handle input from multiple users.? The first person to close their punch would close the reader for everyone.? While the jobs would still be in the VM SPOOL awaiting input, they would not be read by MVT until the reader was restarted.

Oh, well, I guess I'll just have to run HASP when I want to allow outside input... :)

Jeff-

Have you tried to set the VM date set back 56 years? A multiple of 28 and sets the date back before 2000 for MVT.

-Uwe

Hi Tony,

Thanks for the feedback.? Kevin Leonard has said that the next MVT Turnkey will have support for dates beyond 1999.? For now, I'm just working to get MVT to coexist with VM and get the best results possible until the new Turnkey comes out.? I can live with the "random" dates and times, I just wanted the time to be accurate if possible and having the day of week match up by setting the year back by 28 would have been a plus, too.