Discussion:
ipl linux 390 from dasd
(too old to reply)
peter_salous
2003-03-23 13:18:53 UTC
Permalink
Hello,
I already start OS/390 succesfuly, and I initilize one of the dasd
(3390-2) as described in
http://home.iae.nl/users/rvdheij/linuxipl.html after down loading the
three file (tapeipl.ikr, parmfile and initrd from suse) but when i
ipl from this dasd i get the following message
CPU0000: Disabled wait state
PSW=000A0000 0000000E
So apprecaite any help.
Peter salous


------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get 128 Bit SSL Encryption!
http://us.click.yahoo.com/xaxhjB/hdqFAA/xGHJAA/W4wwlB/TM
---------------------------------------------------------------------~->
peter_salous
2003-03-25 13:48:01 UTC
Permalink
This post might be inappropriate. Click to display it.
Ron MacRae
2003-03-26 11:16:06 UTC
Permalink
--- In hercules-390-***@public.gmane.org, "peter_salous"
<peter_salous-F5Bj5G+***@public.gmane.org> wrote:
<<SNIP>>

Peter,
I've just been through this process yesterday. It's not easy or
intuative and I had to get info from lots of different sources.

What happened when you IPLed from the tape.

Did the system detect your DASD.
Did you get the "wecome to Linux" etc?
Did you go through all the stuff about using CTC as your network
device and specifying the addresses and all the rest?

Did you go through all the stuff avout selecting a swap partintion,
initialising a root partition, and loading up optional components etc?

Regards, Ron MacRae.


------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get 128 Bit SSL Encryption!
http://us.click.yahoo.com/xaxhjB/hdqFAA/xGHJAA/W4wwlB/TM
---------------------------------------------------------------------~->
Botrous Thalouth
2003-03-26 14:29:50 UTC
Permalink
Hello Ron,
I create an IPL dasd (01cc) using the sample job
defined in the url
http://home.iae.nl/users/rvdheij/linuxipl.html
and i used the three file as descriped in this job
after shutdwon the OS/390 i prepared another
hercules.cnf file and i change the load parm data also
i put the 1cc as the first on on dasd as you can see
it from my hercules.cnf file.
All the difintion for CTC and the rest was defined as
described for OS/390 and OS/390 working fine, I have
p/390 package and it work fine.
the problem come only when i try to ipl from this tape
to run linux/390 under hercules.
thanks a lot in advance for your help .
Peter



__________________________________________________
Do you Yahoo!?
Yahoo! Platinum - Watch CBS' NCAA March Madness, live on your desktop!
http://platinum.yahoo.com

------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get 128 Bit SSL Encryption!
http://us.click.yahoo.com/xaxhjB/hdqFAA/xGHJAA/W4wwlB/TM
---------------------------------------------------------------------~->
Ron MacRae
2003-03-28 09:30:05 UTC
Permalink
Post by Botrous Thalouth
Hello Ron,
I create an IPL dasd (01cc) using the sample job
<<SNIP>>

Peter,
There is an easier way. Create a logical tape by setting up a
file called linuxipl.tdf containing the following, without the ---s.
--------------------------------------
@TDF
/tapes/tapeipl.ikr FIXED RECSIZE 1024
TM
/tapes/parmline FIXED RECSIZE 1024
TM
/tapes/initrd FIXED RECSIZE 1024
TM
TM
EOT
----------------------------------------
Save this in the tapes directory, along with the 3 files, point the
tape unit at it then IPL from the tape. No other IBM OS required.

What version of Linux are you using? If it's SuSe there should be a
manual.pdf describing what you do next, although it's all prompted
for except the bit where you switch to a telnet client. Also remember
to prefix hardware console command to Linux with '.'.

regards, Ron.


------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get 128 Bit SSL Encryption!
http://us.click.yahoo.com/xaxhjB/hdqFAA/xGHJAA/W4wwlB/TM
---------------------------------------------------------------------~->
Botrous Thalouth
2003-04-03 12:11:30 UTC
Permalink
This post might be inappropriate. Click to display it.
Peter Murray Stockdill (E-mail 2)
2003-04-03 17:00:17 UTC
Permalink
Post by Ron MacRae
@TDF
tapeipl.ikr FIXED RECSIZE 1024
TM
initrd FIXED RECSIZE 1024
TM
parmline FIXED RECSIZE 1024
TM
TM
EOT
Botrous,

For Windows hosts (and for Linux hosts with Hercules 2.17 onwards) this TDF
file should be:

@TDF
tapeipl.ikr FIXED RECSIZE 1024
initrd FIXED RECSIZE 1024
parmline FIXED RECSIZE 1024
TM
EOT

Prior to Hercules 2.17 dummy lines were required in TDF files having only a
newline (LF) character as the end-of-line terminator (the Linux/POSIX text
file convention) rather than a carriage-return newline character sequence
(CRLF) as the end-of-line terminator (the Windows text file convention).
The dummy lines were usually coded as TM lines but they could actually have
contained any character string.

Without this bug a TM or EOT line marks the logical end of the TDF file. In
your case Hercules is emulating a tape containing only tapeipl.ikr which
results in the 'HHCTA050E Attempt to access beyond end of tape
D:/linuxipl.tdf' message.

Hope this helps,
Peter Stockdill.


------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get 128 Bit SSL Encryption!
http://us.click.yahoo.com/W7NydA/hdqFAA/VygGAA/W4wwlB/TM
---------------------------------------------------------------------~->
Marian Gasparovic
2003-04-07 14:23:18 UTC
Permalink
Hi all,
I would like to ask about 3490e tape support in
current Hercules. I created .het tape from real 3490
tape. Now when I try to read it under Hercules, I have
to specify TRTCH=NOCOMP (thank you archive) and it
works fine. But I tried to define 3490 in hcd, tried
both 3480 and 3490 in hercules (3490 should work
according to README.TAPE), but canot get this drive
online. Complains about compact feature not present,
when I defined it in HCD to have compact feature, I
get
PHYSICAL DEVICE INCONSISTENT WITH LOGICAL DEFINITION
HCD DEVICE DEFINITION AT VARIANCE WITH RDC DATA
Not a showstopper, but I am curious what is going on.
Do I try something what is not supposed to work or do
I just make some stupid mistake as usual ?
Thank you

Thank you

=====
===================
Marian Gasparovic
===================
"The mere thought hadn't even begun to speculate about the merest possibility of crossing my mind."



__________________________________________________
Do you Yahoo!?
Yahoo! Tax Center - File online, calculators, forms, and more
http://tax.yahoo.com

------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get 128 Bit SSL Encryption!
http://us.click.yahoo.com/W7NydA/hdqFAA/VygGAA/W4wwlB/TM
---------------------------------------------------------------------~->
herc_fun
2003-04-07 17:27:06 UTC
Permalink
Post by Marian Gasparovic
Hi all,
I would like to ask about 3490e tape support in
current Hercules.
..... (3490 should work
according to README.TAPE), but canot get this drive
online. Complains about compact feature not present,
when I defined it in HCD to have compact feature, I
get
PHYSICAL DEVICE INCONSISTENT WITH LOGICAL DEFINITION
HCD DEVICE DEFINITION AT VARIANCE WITH RDC DATA
All I could see in the docs was 3480 (no 3490 mentioned).
I also had a similar problem with compact in my HCD and had
to re-do the HCD without compaction.

Don't forget that .het is all happening at the device emulation
level and doesn't have any influance on what RDC (read device
characteristics) returns back to the mainframe OS.

I also ran into a case lately when dealing with herc tapes where I
could really have used some of the 3490 capability; in particular
the ability to use NOTE & POINT to do absolute block positioning on
the tape. (Any developers listening ;>)

Charlie.


------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get 128 Bit SSL Encryption!
http://us.click.yahoo.com/xaxhjB/hdqFAA/VygGAA/W4wwlB/TM
---------------------------------------------------------------------~->
ivan_warhead
2003-04-07 17:29:03 UTC
Permalink
Listening..

Will look into it tomorrow.. (sorry.. got a date tonight ;)

--Ivan
Post by herc_fun
Post by Marian Gasparovic
Hi all,
I would like to ask about 3490e tape support in
current Hercules.
..... (3490 should work
according to README.TAPE), but canot get this drive
online. Complains about compact feature not present,
when I defined it in HCD to have compact feature, I
get
PHYSICAL DEVICE INCONSISTENT WITH LOGICAL DEFINITION
HCD DEVICE DEFINITION AT VARIANCE WITH RDC DATA
All I could see in the docs was 3480 (no 3490 mentioned).
I also had a similar problem with compact in my HCD and had
to re-do the HCD without compaction.
Don't forget that .het is all happening at the device emulation
level and doesn't have any influance on what RDC (read device
characteristics) returns back to the mainframe OS.
I also ran into a case lately when dealing with herc tapes where I
could really have used some of the 3490 capability; in particular
the ability to use NOTE & POINT to do absolute block positioning on
the tape. (Any developers listening ;>)
Charlie.
------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get 128 Bit SSL Encryption!
http://us.click.yahoo.com/W7NydA/hdqFAA/VygGAA/W4wwlB/TM
---------------------------------------------------------------------~->
ivan_warhead
2003-04-23 16:05:05 UTC
Permalink
Hello,

Problem solved. Device characteristics were missing the IDRC bit in
byte 9 bit 1.

NOTE : as of today, activating IDRC has no effect on HET compression.

--Ivan
Post by Marian Gasparovic
Hi all,
I would like to ask about 3490e tape support in
current Hercules. I created .het tape from real 3490
tape. Now when I try to read it under Hercules, I have
to specify TRTCH=NOCOMP (thank you archive) and it
works fine. But I tried to define 3490 in hcd, tried
both 3480 and 3490 in hercules (3490 should work
according to README.TAPE), but canot get this drive
online. Complains about compact feature not present,
when I defined it in HCD to have compact feature, I
get
PHYSICAL DEVICE INCONSISTENT WITH LOGICAL DEFINITION
HCD DEVICE DEFINITION AT VARIANCE WITH RDC DATA
Not a showstopper, but I am curious what is going on.
Do I try something what is not supposed to work or do
I just make some stupid mistake as usual ?
Thank you
Thank you
=====
===================
Marian Gasparovic
===================
"The mere thought hadn't even begun to speculate about the
merest possibility of crossing my mind."
Post by Marian Gasparovic
__________________________________________________
Do you Yahoo!?
Yahoo! Tax Center - File online, calculators, forms, and more
http://tax.yahoo.com
------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get 128 Bit SSL Encryption!
http://us.click.yahoo.com/W7NydA/hdqFAA/bW3JAA/W4wwlB/TM
---------------------------------------------------------------------~->
Marian Gasparovic
2003-04-24 12:25:25 UTC
Permalink
I want to report a problem with shadow files merging.
I had five generations of shadow files and tried to
merge it down (without guest OS IPLed, just Hercules
running). It seems it creates some temporary files
that are not removed, after 4th sf-* I got error like
"no space on device" plenty of time for plenty of
offsets. And final message was like "merged with
errors". df showed full filesystem, but after quitting
hercules, there was plenty of room on disk. Then I
tried it again from backup copies and every time
after sf-* I had some amount of free space on
filesystem and after quitting Hercules the amount was
_much_ bigger, those were hundreds of megabytes.
And one more issue regarading cckd. I had one .ckd
disk, tried ckd2cckd and got segmentation fault. Tired
again, it worked. Tried couple of times and got 2:8
success:segmentation fault.
Both Hercules and ckd2cckd are from cvs two or three
days old. ckd2cckd behaved the same both under
Mandrake and Debian, merging issue tried only under
Debian.

Greg ?


=====
===================
Marian Gasparovic
===================
"The mere thought hadn't even begun to speculate about the merest possibility of crossing my mind."



__________________________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo
http://search.yahoo.com

------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get 128 Bit SSL Encryption!
http://us.click.yahoo.com/xaxhjB/hdqFAA/bW3JAA/W4wwlB/TM
---------------------------------------------------------------------~->
Greg Smith
2003-04-24 21:10:55 UTC
Permalink
Post by Marian Gasparovic
I want to report a problem with shadow files merging.
I had five generations of shadow files and tried to
merge it down (without guest OS IPLed, just Hercules
running). It seems it creates some temporary files
that are not removed, after 4th sf-* I got error like
"no space on device" plenty of time for plenty of
offsets. And final message was like "merged with
errors". df showed full filesystem, but after quitting
hercules, there was plenty of room on disk. Then I
tried it again from backup copies and every time
after sf-* I had some amount of free space on
filesystem and after quitting Hercules the amount was
_much_ bigger, those were hundreds of megabytes.
I was able to recreate this problem. The problem seems
to be that the unlink() doesn't release the file space
occupied by the file. (unlink is basically remove or
erase or delete file). This can happen if more than
one file descriptor is opened on the file and the space
won't be released until all file descriptors are closed.
However, I traced all opens and closes and don't see this
happening. And I haven't been able (yet) to write a simple
testcase to duplicate the problem.

However, I *have* updated current cvs to truncate the file
to zero bytes before close/unlink so the space is now
available. I consider this a workaround until I figure out
what the real problem is. It may have to do with the multi-
threaded nature of Hercules. Anyway, it should behave a
lot better now.
Post by Marian Gasparovic
And one more issue regarading cckd. I had one .ckd
disk, tried ckd2cckd and got segmentation fault. Tired
again, it worked. Tried couple of times and got 2:8
success:segmentation fault.
Both Hercules and ckd2cckd are from cvs two or three
days old. ckd2cckd behaved the same both under
Mandrake and Debian, merging issue tried only under
Debian.
Are you familiar with running gdb ?? If so, edit the Makefile
and change the CFLAGS option -fomit-frame-pointer to -g instead.
Then `make clean && make'. Then `gdb ckd2cckd'. Then
`run <parameters>' where <parameters> is what you would type
after `ckd2cckd'. When the error occurs, enter `bt' to
display the stack trace. This should give me an idea where
the error is occurring.

Thanks,
Greg



------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get 128 Bit SSL Encryption!
http://us.click.yahoo.com/W7NydA/hdqFAA/bW3JAA/W4wwlB/TM
---------------------------------------------------------------------~->
s***@public.gmane.org
2003-04-25 01:34:09 UTC
Permalink
Post by Greg Smith
Post by Marian Gasparovic
I want to report a problem with shadow files merging.
I had five generations of shadow files and tried to
merge it down (without guest OS IPLed, just Hercules
running). It seems it creates some temporary files
that are not removed, after 4th sf-* I got error like
"no space on device" plenty of time for plenty of
offsets. And final message was like "merged with
errors". df showed full filesystem, but after quitting
hercules, there was plenty of room on disk. Then I
tried it again from backup copies and every time
after sf-* I had some amount of free space on
filesystem and after quitting Hercules the amount was
_much_ bigger, those were hundreds of megabytes.
I was able to recreate this problem. The problem seems
to be that the unlink() doesn't release the file space
occupied by the file. (unlink is basically remove or
erase or delete file). This can happen if more than
one file descriptor is opened on the file and the space
won't be released until all file descriptors are closed.
However, I traced all opens and closes and don't see this
happening. And I haven't been able (yet) to write a simple
testcase to duplicate the problem.
lsof is your friend.
--
Cheers
John.

Join the "Linux Support by Small Businesses" list at
http://mail.computerdatasafe.com.au/mailman/listinfo/lssb



------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get 128 Bit SSL Encryption!
http://us.click.yahoo.com/W7NydA/hdqFAA/bW3JAA/W4wwlB/TM
---------------------------------------------------------------------~->
Marian Gasparovic
2003-04-25 07:57:14 UTC
Permalink
Greg, I downloaded current cvs and I cannot recreate
the problem. So, it is solved :)
Marian
Post by Greg Smith
Post by Marian Gasparovic
And one more issue regarading cckd. I had one .ckd
disk, tried ckd2cckd and got segmentation fault.
Tired
Post by Marian Gasparovic
again, it worked. Tried couple of times and got
2:8
Post by Marian Gasparovic
success:segmentation fault.
Both Hercules and ckd2cckd are from cvs two or
three
Post by Marian Gasparovic
days old. ckd2cckd behaved the same both under
Mandrake and Debian, merging issue tried only
under
Post by Marian Gasparovic
Debian.
Are you familiar with running gdb ?? If so, edit
the Makefile
and change the CFLAGS option -fomit-frame-pointer to
-g instead.
Then `make clean && make'. Then `gdb ckd2cckd'.
Then
`run <parameters>' where <parameters> is what you
would type
after `ckd2cckd'. When the error occurs, enter `bt'
to
display the stack trace. This should give me an
idea where
the error is occurring.
Thanks,
Greg
__________________________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo
http://search.yahoo.com

------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get 128 Bit SSL Encryption!
http://us.click.yahoo.com/xaxhjB/hdqFAA/bW3JAA/W4wwlB/TM
---------------------------------------------------------------------~->
Ron MacRae
2003-03-28 09:29:41 UTC
Permalink
Post by Botrous Thalouth
Hello Ron,
I create an IPL dasd (01cc) using the sample job
<<SNIP>>

Peter,
There is an easier way. Create a logical tape by setting up a
file called linuxipl.tdf containing the following, without the ---s.
--------------------------------------
@TDF
/tapes/tapeipl.ikr FIXED RECSIZE 1024
TM
/tapes/parmline FIXED RECSIZE 1024
TM
/tapes/initrd FIXED RECSIZE 1024
TM
TM
EOT
----------------------------------------
Save this in the tapes directory, along with the 3 files, point the
tape unit at it then IPL from the tape. No other IBM OS required.

What version of Linux are you using? If it's SuSe there should be a
manual.pdf describing what you do next, although it's all prompted
for except the bit where you switch to a telnet client. Also remember
to prefix hardware console command to Linux with '.'.

regards, Ron.


------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get 128 Bit SSL Encryption!
http://us.click.yahoo.com/xaxhjB/hdqFAA/xGHJAA/W4wwlB/TM
---------------------------------------------------------------------~->
Paul Truscott
2003-04-05 13:57:29 UTC
Permalink
I am trying to install SLES 7 (I think) under Herc. I've created the TDF
file as shown in Digest 1789 below. However, I didn't have a file called
parmline. I did have one called parmfile, and so I copied that and the other
from my cd into my /tape directory (where the TDF file is). When I try to
IPL off /tape/linuxipl.tdf I get the following;

16:39:33.359 00000EEC RAMDISK: Couldn't find valid RAM disk image starting
at 0.
16:39:33.359 00000EEC Kernel panic: VFS: Unable to mount root fs on 01:00
16:39:33.359 00000EEC HHCCP011I CPU0000: Disabled wait state
16:39:33.369 00000EEC PSW=000A0000 8005684C

I've had a brief read through the PDF files that are on the SLES CD's, but
the bits relating to Herc are a little out of date, and don't seem to deal
with trying to IPL off a tape and then pulling the rest of the data from
CD's.

I'm running Herc 2.17-1, with fthreads, Hercgui, etc... all under XP. Can't
complain about anything Tun/Tap oriented as

Can someone explain, how tapeipl.ikr, initrd and parmline relate to booting
up from tape? From my perspective, it would appear that this TDF file
contains enough info to boot linux in the same way that you can create a
Linux boot floppy. Am I correct? If so, how do I get to the CD's then?

Appreciate some tips.

Many thanks in advance.

Paul Truscott
Al Falak SPD
Al Khobar, Saudi Arabia
PH: +966-3-882-9996 Ext 164
Fx: +966-3-882-0826
Mb: +966-54-986-936

The following was pulled from Fridays Digest 1789.


Message: 10
Date: Fri, 4 Apr 2003 03:00:17 +1000
From: "Peter Murray Stockdill \(E-mail 2\)" <petersto-24jMwgp+***@public.gmane.org>
Subject: RE: Re: ipl linux 390 from dasd
Post by Ron MacRae
@TDF
tapeipl.ikr FIXED RECSIZE 1024
TM
initrd FIXED RECSIZE 1024
TM
parmline FIXED RECSIZE 1024
TM
TM
EOT
Botrous,

For Windows hosts (and for Linux hosts with Hercules 2.17 onwards) this TDF
file should be:

@TDF
tapeipl.ikr FIXED RECSIZE 1024
initrd FIXED RECSIZE 1024
parmline FIXED RECSIZE 1024
TM
EOT

Prior to Hercules 2.17 dummy lines were required in TDF files having only a
newline (LF) character as the end-of-line terminator (the Linux/POSIX text
file convention) rather than a carriage-return newline character sequence
(CRLF) as the end-of-line terminator (the Windows text file convention).
The dummy lines were usually coded as TM lines but they could actually have
contained any character string.

Without this bug a TM or EOT line marks the logical end of the TDF file. In
your case Hercules is emulating a tape containing only tapeipl.ikr which
results in the 'HHCTA050E Attempt to access beyond end of tape
D:/linuxipl.tdf' message.

Hope this helps,
Peter Stockdill.


------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get 128 Bit SSL Encryption!
http://us.click.yahoo.com/xaxhjB/hdqFAA/VygGAA/W4wwlB/TM
---------------------------------------------------------------------~->
Continue reading on narkive:
Loading...