Installation of sarge on hercules

[Frans Pop]

Hi,

I've done several installations on Hercules without problems. I also
happen to be a member of the d-i team. Hope I can help.

Some questions:
- Can you still reproduce the problem?
- Does selecting an other mirror help?
- How did you set up forwarding on the host machine?
- What selections did you make when configuring the network in d-i?

Cheers,
FJP

P.S. I am subscribed to the list, but must have missed your post. Just
came across it in the archives, so that's why the late reply.

[Guenther Bergmann]

Hi Frans,

Post by Frans Pop
Hi,
I've done several installations on Hercules without problems. I
also happen to be a member of the d-i team. Hope I can help.
- Can you still reproduce the problem?

Yes, this problem is persistent.

Post by Frans Pop
- Does selecting an other mirror help?

Yes, i've tried different mirrors located in germany and a few in
the USA and Australia.

Post by Frans Pop
- How did you set up forwarding on the host machine?

echo "1" >/proc/sys/net/ipv4/ip_forward

Post by Frans Pop
- What selections did you make when configuring the network in
d-i?

1. network device type: ctc
2. CTC-device: 0x0800-0x081
3. protocol for this connection: S/390
4. primary network interface: ctc0
5. ip-address: 192.168.253.1
6. point-to-point-address: 192.168.253.2
7. nameserver: 192.168.254.1
8. hostname: andromeda
9. domain: gbergmann.de

I have tested network connectivity as follows:
- ssh ***@andromeda
- open shell
- then:

BusyBox v1.00-pre10 (Debian 20040623-1) Built-in shell (ash)
Enter 'help' for a list of built-in commands.

~ # wget http://www.debian.org
Connecting to www.debian.org[194.109.137.218]:80
index.html 100% |*****************************| 15636
00:00 ETA

So i think network connectivity is fine.

Post by Frans Pop
P.S. I am subscribed to the list, but must have missed your post.
Just came across it in the archives, so that's why the late
reply.

No Problem, any help is appreciated.

best regards

Guenther

[Frans Pop]

Hello Günther,

I've tried to reproduce your problem, but I'm afraid everything works fine
for me.
I did a tape boot using the daily d-i image of Jan 27. from:
http://raptor.debian.org/d-i/images/daily/tape/

I used the same mirror you did (http://ftp2.de.debian.org) and both
selecting the mirror and downloading installer components in the next
step worked perfectly.
The debug message you see in the logs is normal, but the error _is_
related to that command being executed and the results from it verified.

Did you try on different days or was it all on the same day? In that case
it may be that there was just a general problem with the archives.

FWIW, I compiled hercules myself using the upstream 3.01 version (as the
version in Debian does not like 2.6 kernels).

I suggest you give it one more try with the latest images as well.
I don't really see what else I can do to help you here.

Cheers,
Frans

Post by Guenther Bergmann

Post by Frans Pop
- How did you set up forwarding on the host machine?

echo "1" >/proc/sys/net/ipv4/ip_forward

I have additionally set up forwarding using IP-tables, but as networking
works for you otherwise...

Post by Guenther Bergmann

Post by Frans Pop
- What selections did you make when configuring the network in d-i?

1. network device type: ctc
2. CTC-device: 0x0800-0x081
3. protocol for this connection: S/390
4. primary network interface: ctc0
5. ip-address: 192.168.253.1
6. point-to-point-address: 192.168.253.2
7. nameserver: 192.168.254.1
8. hostname: andromeda
9. domain: gbergmann.de

I use the same options, except that my CTC-device is on 0x0a00-0x0a01.

[Ivan Warren]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Post by Frans Pop
FWIW, I compiled hercules myself using the upstream 3.01 version
(as the version in Debian does not like 2.6 kernels).
I suggest you give it one more try with the latest images as well.
I don't really see what else I can do to help you here.

Frans,

Side note : Current upstream hercules version is 3.02, not 3.01
(maybe you should give it a try !)

Guenther,

What version of hercules are you trying with ? there is a slight
possibility the version you are using may be incorrectly emulating
the architecture - to a point it is giving incorrect results...

Maybe you should try the latest 3.02 hercules version.. (which
incidently also performs *much* better than 3.01 performance wise - a
non-negligeable point when talking about architecture emulators).

...

<OT>
And... Maybe someone should try convincing MZ into re-injecting
hercules in the debian loop - the old 2.16 deb version is just way
too outdated to be of any significant value !
</OT>

- --Ivan

-----BEGIN PGP SIGNATURE-----
Version: PGP 8.1

iQA/AwUBQfrZJodFv4a3KlzEEQI0bgCgjdMTYJOQkMbD+hujioSql/OV04gAnA4F
afxdKmT12iZXROlgsFI6jvKs
=CFdv
-----END PGP SIGNATURE-----

[Bastian Blank]

Post by Ivan Warren
Maybe you should try the latest 3.02 hercules version.. (which
incidently also performs *much* better than 3.01 performance wise - a
non-negligeable point when talking about architecture emulators).

I never saw this problem in z/VM on the real hardware.

Bastian

[Ivan Warren]

Post by Bastian Blank
I never saw this problem in z/VM on the real hardware.
Bastian

It's a totaly different issue.. z/VM and PR/SM are virtualisation
layers, while hercules is a hardware emulator.

But the initial problem *might* not have anything to do with the fact
Guenther is trying to install debian on emulated hardware (I
personally never ran into the issue described either, although I run
debian/390 exclusivelly on emulated hardware).

- --Ivan

[Frans Pop]

Post by Ivan Warren
Side note : Current upstream hercules version is 3.02, not 3.01
(maybe you should give it a try !)

Thank you for alerting me to that. I now get over 30 MIPS sometimes.
The maximum I used to see with 3.01 was about 24, so an improvement
of about 25%. Impressive!
Makes a lot of difference when running aptitude and such.

However, I did get what looks like a serious error during boot of
my existing installation:
Starting MTA:
ctc0 Illegal packet type 0xffff received, dropping
/usr/sbin/update-exim4.conf: line 380: 243 Segmentation fault mv -f ${UPE
X4C_outputfile}.tmp ${UPEX4C_outputfile}
exim4.

There was a new version of exim4 available, and during the upgrade
I got a similar error:
Installing new version of config file /etc/exim4/exim4.conf.template ...
/var/lib/dpkg/info/exim4-config.postinst: line 182: 815 Segmentation fault mv /etc/exim4/update-exim4.conf.conf.tmp /etc/exim4/update-exim4.conf.conf
dpkg: error processing exim4-config (--configure):
subprocess post-installation script returned error exit status 139

I've never seen similar errors while running 3.01.
(Weird that moves should cause segfaults.)

After restarting hercules with only 1 processor (I've been running with 2),
everything was fine. After that I started with 2 processors again and the
error during boot did not reoccur.
So I'm not sure if the problem was fixed by restarting hercules or by reducing
the number of processors. I'll keep an eye out for other weirdness.

Cheers,
FJP