开云体育

On Fri, Jun 30, 2023 at 09:43 AM, Fish Fish wrote:

I looked at your log (thank you!) and my analysis (which PROVES that "TPTP" has a serious bug!) is attached.

(ATTACHMENT)

Alejandro olivan Alvarez wrote:

[...]

and re-run it, this time but, following your advice:
no LOG statement in configuration, using -o parameter
to get the whole output (I took a look and now I see
what you meant, it shows many details about the actual
binary build). I attach the new herclog.

Thank you!

Regarding the question of where did I get the file, ...

Never mind. I found it. :)


[...]

it usually gives me around 18-20 mips, with peaks at 26 mips).
I'm aware my environment is somehow not the usual one, so maybe
it may cause issues.

I doubt it's your system.

Instead, as I said in my previous message, I suspect there is a BUG in the "TPTP" program.

Now, based on what you explained, I'm going to repeat again
the experiment, but either declaring the tape as read-only
and/or setting the immutable atribute at the filesystem level
(I guess this last route could cause problems).

Setting the immutable attribute for the file would be the ideal in my opinion. It is a file with valuable READ ONLY data, and should *NEVER* be written to. Not even by root!

As an alternative however, you could simply be careful to always specify the "ro" (read only) option on your 'devinit' command:

devinit 280 ...(tape file name)... ro

See:

*


But if it were me, I personally would recommend setting the immutable attribute on the file. It's much safer in my opinion.

I looked at your log (thank you!) and my analysis (which PROVES that "TPTP" has a serious bug!) is attached.

Is the source code to the TPTP program available? I would sure like to take a look at it!

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

mail: fish@...

On Wed, Jun 28, 2023 at 06:35 PM, Fish Fish wrote:

Where did you download it from anyway? I'd like to take a look at it myself with my AWSBrowse program to see what it looks like. Thanks.

Finally, for the record, since you failed to provide a COMPLETE Hercules log, what version of Hercules are you using, and on which type of host are you running it on? Windows? Linux? Which Windows? Which Linux?

Thanks.

I hope that helps!

Hi!

Thank you for your interest! ... Because of you very detailed explnation, I've repeated the whole experiment (I perform snapshots often, when operating hercules because I tend to run into troubles quite often :-P) so I re-downloaded the aws tape, hetinit a new one, and re-run it, this time but, following your advice: no LOG statement in configuration, using -o parameter to get the whole output (I took a look and now I see what you meant, it shows many details about the actual binary build). I attach the new herclog.

Regarding the question of where did I get the file, I'm using the link provided in the site where the sysgen instructions are presented.
More precissely, the link sits on the page at:
And the link itself ends up pointing to a Dropbox download link. here's the link at the page:?

I would add that, beyond what's showing on herclog about my running environment, I'm running hyperion on the same machine I build it: a Linux Debian 12/Bookworm VM.
I guess this deviates from the typical Windows/Desktop environment where hercules accesses many CPU cores and has pretty fast I/O against an SSD or even NVMe... My VM has only 2 vCores, and they're not high on priority on hte hypervisor. And although the VM itself has its system disks runing on local hyoervisor SSDs, the storage where hercules is run on sits on an NFS mount (so far, this setup has worked flawlessly with Tk4-, VM/370 and some other stuff, it usually gives me around 18-20 mips, with peaks at 26 mips). I'm aware my environment is somehow not the usual one, so maybe it may cause issues.

Now, based on what you explained, I'm going to repeat again the experiment, but either declaring the tape as read-only and/or setting the immutable atribute at the filesystem level (I guess this last route could cause problems).

Best regards!?
?

Alejandro olivan Alvarez wrote:

I attach the configuration file I use, and the hercules log
(which is very long!)

[...]

I noticed two messages appearing just before the I/O error:
I find relevant that, just before the I/O Error, two lines
show up that look very suspicious:

11:15:00 HHC00204E 0:0281 Tape file tape/power-deblock.aws,
type aws: error in function readhdr_awstape(), offset
0x0000000000000D20: end of file (uninitialized tape)

and:

11:15:00 HHC00204E 0:0280 Tape file tape/ibm.360d052006.dos.
power.II.with.rje.opt2.v4m1.aws, type aws: error in function
readhdr_awstape(), offset 0x0000000000000D0A: end of file
(uninitialized tape)

Strangely but, before these errors show up, there was a lot of
activity both on the log regarding devices 280 and 281, and on
the tape files themselves, as I see on the filesystem that their
sizes changed, so it looks like hercules was initially happy
handling the tapes but at some point the thing went wrong.

Yep. I'm seeing the same thing.

But before we get to that, PLEASE REMOVE YOUR:

LOG herclog.log

STATEMENT FROM YOUR CONFIGURATION FILE, and specify a *proper* log file via the '-o' or '--output' or '--logfile' command line option instead!

*

What you sent me was NOT the full/complete Hercules logfile! It is only a partial logfile. It is missing perhaps the most important part of a Hercules logfile: THE VERY BEGINNING, that contains all of Hercules's startup initialization messages!

I would recommend NEVER using the 'log' statement or command EVER! Never ever! Using it make it virtually impossible to provide proper support for a person when they're having problems, since we cannot see what version of Hercules they are using and other important things that get logged during Hercules startup and initialization.

Thanks.

Now... About your problem...

Looking at your log file, it is painfully obvious to me there is an extremely high likelihood that DOS/360's "TPTP" program (or DOS/360 itself) is buggy.

The problem does NOT appear to me to be with Hercules. The problem is instead, with an extremely high degree of likelihood, with TPTP and/or DOS/360.

Tape 280 is the INPUT tape, yes?

And 281 is the OUTPUT tape, yes?

"TPTP" is a simple(?) program that is supposed to READ the specified input tape, deblock each physical block of data that it reads from that input tape, and then write the individual unblocked records from that block to the output tape. Yes? Pretty simple.

Well, for some very strange unknown reason, TPTP (or DOS/360) --in the middle of its processing! -- is deciding to suddenly and unexpectedly switch from *reading* its input tape, to, oddly enough, start *WRITING* to it instead! (WTF?!)

Yes, you read that sentence correctly: It is deciding for reasons unknown to start *WRITING* to its *INPUT* tape! (WTF?!)

AND... it decides at the same time, to also suddenly and unexpectedly switch from writing to its output tape to READING from it instead! (WTF?!)

Yep! It decides for some unknown reason (likely due to the same bug I'm guessing) to start *READING* from its *OUTPUT* tape! (WTF?!)

DUE TO THIS APPARENT BUG, I would *strongly* recommend that you re-download your "ibm.360d052006.dos.power.II.with.rje.opt2.v4m1.aws" input tape file, and *immediately* set it to read-only, to ensure that it cannot be damaged due to this bug.

Where did you download it from anyway? I'd like to take a look at it myself with my AWSBrowse program to see what it looks like. Thanks.

Finally, for the record, since you failed to provide a COMPLETE Hercules log, what version of Hercules are you using, and on which type of host are you running it on? Windows? Linux? Which Windows? Which Linux?

Thanks.

I hope that helps!

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

mail: fish@...

Hi!

I attach the configuration file I use, and the hercules log (which is very long!)

On 6/27/23 20:21, Fish Fish wrote:

[...]

BG 0P10I C EQUIP CHK SYS005=281
CCSW=0210006B080E400CA8 SNS=1040000000C0 CCB=006CD8
BG 0P10I C EQUIP CHK SYS004=280
CCSW=0210006B080E400CA8 SNS=1040000000C0 CCB=006AC8
BG 0P73I JOB POWER01 CANCELED DUE TO I/O ERROR

The aws file is the one referenced above in this thread,
and destination file was successfuly created by hetinit...
so no clue what's going wrong.

Enable CCW tracing for both devices ('t+280' and 't+281' Hercules commands) and try your job again and post the resulting Hercules logfile.

Seeing your Hercules configuration file would help too.



, I noticed two messages appearing just before the I/O error:

I find relevant that, just before the I/O Error, two lines show up that look very suspicious:

11:15:00 HHC00204E 0:0281 Tape file tape/power-deblock.aws, type aws: error in function readhdr_awstape(), offset 0x0000000000000D20: end of file (uninitialized tape)

and

11:15:00 HHC00204E 0:0280 Tape file tape/ibm.360d052006.dos.power.II.with.rje.opt2.v4m1.aws, type aws: error in function readhdr_awstape(), offset 0x0000000000000D0A: end of file (uninitialized tape)

Strangely but, before these errors show up, there was a lot of activity both on the log regarding devices 280 and 281, and on the tape files themselves, as I see on the filesystem that their sizes changed, so it looks like hercules was initially happy handling the tapes but at some point the thing went wrong.


Thank you very much.

Regards.

On Tue, Jun 27, 2023 at 09:35 AM, Alejandro olivan Alvarez wrote:

>> Has anyone succeded in getting POWER completed?

Yes, I have. :-) See below:

[...]

BG 0P10I C EQUIP CHK SYS005=281
CCSW=0210006B080E400CA8 SNS=1040000000C0 CCB=006CD8
BG 0P10I C EQUIP CHK SYS004=280
CCSW=0210006B080E400CA8 SNS=1040000000C0 CCB=006AC8
BG 0P73I JOB POWER01 CANCELED DUE TO I/O ERROR

The aws file is the one referenced above in this thread,
and destination file was successfuly created by hetinit...
so no clue what's going wrong.

Enable CCW tracing for both devices ('t+280' and 't+281' Hercules commands) and try your job again and post the resulting Hercules logfile.

Seeing your Hercules configuration file would help too.



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

mail: fish@...

Hi!

I stumbled with this thread as, precissely, getting POWER on DOS/360 as explained at tutorial ?is the first (and I bet, the only) thing that goes wrong.
Has anyone succeded in getting POWER completed? ... At the very first JOB to deblock the POWER tape into another tape, I got an unexpected job cancellation due to I/O error:

>> I guess It would be therefore reasonable, upon success on the patching, to keep a copy of the resulting, patched, dos/vs 3350s and replace the ones shipped with vm/ce.

Alejandro olivan Alvarez wrote:

[...]

So I needed to read a litle on those queue management
commands ... pdisplay and pdelete, [...]

Just be very careful not to confuse "D" (which is the abbreviation for the PDISPLAY command) with "L" (which is the abbreviation for the PDELETE command)!

I still remember an incident from years and years ago at the shop I used to work at, where an inexperienced newbie mistakenly thought "d" meant "delete" and "l" mean "list", and ended up deleting the entire list queue when she entered the command "l lst,*"!! (Oops!!)

:)

So keep in mind:

"L" = DELETE
"D" = "DISPLAY"

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

mail: fish@...

It works now!!! :-D

Thank you all very very much!

It also turned out that I got quite a bit of old jobs queued there... quite a mess.
So I needed to read a litle on those queue management commands ... pdisplay and pdelete, but also psetup and pgo since (don't know why) one of my past hellocob attempts triggered an 'IQ40A' message and was stuck on the lst queue.... overall, I managed to clean all up, and jobs started to run an print nicely...? a ton of fun!!!

Cheers.

Interesting... some good news/bad news situation here:

The bad news is that Hello Worl in cobol does not work.
The good news (to me) is that it wasn't that different from the one I came up and tried before posting here.
Now, the exact issue is that DOS/VS says this at the console:

BG // JOB HELLO
? ? ? DATE 06/18/23,CLOCK 09/53/27
BG C102I? ? ? ?UNSUPPORTED DEVICE TYPE-SYS004
BG 0S02I JOB HELLO? ? CANCELED DUE TO PROGRAM REQUEST
BG EOJ HELLO? ??
? ? ? DATE 06/18/23,CLOCK 09/53/27,DURATION 00/00/00
F1 1Q34I? ? ? BG WAITING FOR WORK

It is very shocking to me that it doesn't work, since, well, it is a DOS/VS as it comes in VM370/CE, and overall and otherwise it behaves as expected... I read a little on those 'problem program' 'labels' SYS00x but, to be honest, it all sounds to me as fascinating as cryptic.
No need to mention that the other two programs suffer the same doom, as are laconically cancelled for the very same unsupported TYPE-SYS004 device... quite pesky this DOS/VS guy...

Cheers!?

Hi guys.

Thank you for your help! I appreciate this very much.

By one side, I have yet another fork on investigation path, since I have missed that DOS/360 site (probably all my research and googiln has always been '370' focused/filtered) ... I feel that's gonna be fun, since a kick look at it makes me thing this site covers a whole OS sysgen, which sounds lika is a ton of fun!

On the other side, now with 3 'known to work' Jobs, I feel more confident approaching DOS/VS (forgot to mention, I run DOS/VS as a VM370-CE guest). I have a starting point :-)

Cheers!??

On Fri, Jun 16, 2023 at 09:37 AM, Alejandro olivan Alvarez wrote:

Does someone know about some fundamental/basic examples of DOS/VS COBOL JOBs to start to play with? I

I have attached three demo programs for you, including the Hello World program. :-)

Cheers,

Rene FERLAND, Montreal

Hi folks.

Of all the languages around in this Mainframe work, Cobol is the only one I played with a little in the past with OpenCobolIDE (although I find Fortran not that hostile :-) ... But here I'm struggling because at the stacked principles that have to be all right in order for a JOB to work.
Does someone know about some fundamental/basic examples of DOS/VS COBOL JOBs to start to play with? I've not found even the classic Hello World sample to check whether fcobol wakes up.

Thank you very much in advance.
Cheers.

?Hello Rene!

Thank you so much for this explanation. I am going to try this right now.
Hope this would create some load for my personal mainframe. :)

Best wishes,
Andre