Skip to contentSkip to conversation

Where to start from?

parisc-linux Bartlomiej Ochman 3 participants 5 messages Sep 1, 2004 – Sep 2, 2004 1 message not in the archive

lore
Share & tools
lore
#0 lore raw

Hello.

For some time now, I installed debian/testing on some spare C200+.
I didn't have much time to play around with it until now. Recently I discovered, it hangs completly, without any message, under heavy disc load (eg. bzip2 -cd kernel-source-2.6.8.tar.bz | tar xf -, or apt-get upgrade cycle).

As I am not familiar with hppa architecture, where should I start to snoop from? I've tried both 2.4.x and 2.6.x kernels, both stock and custom compiled and they both behave the same way.

TIA,
BO
_______________________________________________
parisc-linux mailing list
[email protected]
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

#2 lore raw
In reply to a message not in the archive
Quoted 1 line · Joel Soete

Just to be sure that I well understand: do you mean that you tried the same operation when runing 2.4.x and 2.6.x?

Yes, I tried to untar kernel source both under 2.4 and 2.6. No difference, if it was stock, or custom compiled kernel. It hangs randomly under heavy disc load.

Quoted 1 line

The very first step would be that you send us a dmesg.

Attached. This particular kernel seems to be most stable on my machine, which surprises me after few months on the list.

Quoted 1 line

As it seems to be a disk pb: check the scsi chain [...]

I've double checked this, but everything seems to be fine.

Quoted 1 line

If you still have a diagnostic hp cd, boot with this one and check the disks.

I'll try to figure out this, but I'm afraid the only CDs we have are with HP-UX 10.20 :(

This machine worked as network monitoring station with HP-UX/HP OV on it, it has never reported any problems with any disk.

Regards,
BO

_______________________________________________
parisc-linux mailing list
[email protected]
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

#3 lore raw

Tx BO,

Quoted 6 lines · Bartlomiej Ochman
Quoted 1 line · Joel Soete

Just to be sure that I well understand: do you mean that you tried the same operation when runing 2.4.x and 2.6.x?

Yes, I tried to untar kernel source both under 2.4 and 2.6. No difference, if it was stock, or custom compiled kernel. It hangs randomly under heavy disc load.

mmm in normal operation, can the system run severall hours without pb? (in case of over heating (fan hs), system quartz failure, ...)

And when system hung, if you wait some minutes (about 3 min iirc), don't you read SCSI BUS RESET messages at the console?

That said, it would make me thought to a pb I encounter with my c110 (exactely the same decription) but the driver pb was ncr53c720 (not concerned here) and for some 2.6 only (no pb with 2.4)

Quoted 3 lines
Quoted 1 line

The very first step would be that you send us a dmesg.

Attached. This particular kernel seems to be most stable on my machine,

mmm I just noticed:
...
Cujo version 2.0 found at 0xf1000000
Enabling Cujo 2.0 bug workaround
Don't recognise Cujo at address 0xf1000000, not enabling workaround
Cujo: No PCI devices enabled.
...

duno if it's important or not?

Quoted 3 lines

I've double checked this, but everything seems to be fine.

Nice :)

Quoted 4 lines
Quoted 1 line

If you still have a diagnostic hp cd, boot with this one and check the disks.

I'll try to figure out this, but I'm afraid the only CDs we have are with HP-UX 10.20 :(

Iirc it should stand on "hpux Support+" cd

Quoted 2 lines

This machine worked as network monitoring station with HP-UX/HP OV on it, it has never reported any problems with any disk.

Trust me this logic has no sence in electronic: 3 weeks ago we (I mean an Hp engineer) power off a N system to add a FC card, when we try to restart, ... too bad the boot disk was definitively HS (even thought we check it carefully with mstm before shutdown: the disk never presents any pb before?)

That said, the final step is to 'TOC' your system when it hung (do you know the bleue button behind your box?)
then grab the 'ser piminfo' and analyse what you grab on your system when it reboot with <http://cvs.parisc-linux.org/build-tools/dump_analyser.sh>

hth,
Joel
_______________________________________________
parisc-linux mailing list
[email protected]
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

#4 lore raw
Quoted 1 line · Joel Soete

mmm in normal operation, can the system run severall hours without pb?

Not only hours, but weeks or months. I didn't have troubles during instalation, I think 1.5 years ago, when the disks worked hard too. I also think, everything started when glibc upgraded to 2.3... and "old" sarge with glibc 2.2 and kernel around 2.4.18 had worked without that problem.

Quoted 1 line

And when system hung, if you wait some minutes (about 3 min iirc), don't you read SCSI BUS RESET messages at the console?

Nothing like that. It dies completly, heart stops beating (all diodes on the front panel stops blinking). It doesn't respond to anything.

Quoted 1 line

Trust me this logic has no sence in electronic: 3 weeks ago we (I mean an Hp engineer) power off a N system to add a FC card, when we try to restart, ...

Yeah, I know. But somehow I feel, this is not hardware-related problem... There shuould be some symptoms of disk related problems in the dmesg after few weeks of working, but I have nothing like that.

Quoted 1 line

That said, the final step is to 'TOC' your system when it hung (do you know the bleue button behind your box?)

No?

Quoted 1 line

then grab the 'ser piminfo' and analyse what you grab on your system when it reboot with <http://cvs.parisc-linux.org/build-tools/dump_analyser.sh>

I'll try this, but not before October or so. I'm currently rather busy with my MSc final exam.

Thanks,
BO
_______________________________________________
parisc-linux mailing list
[email protected]
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

#5 lore raw
In reply to Bartlomiej Ochman
Quoted 7 lines · Bartlomiej Ochman

Recently I
discovered, it hangs completly, without any message, under heavy disc load (eg. bzip2 -cd kernel-source-2.6.8.tar.bz | tar xf -, or apt-get upgrade cycle).

As I am not familiar with hppa architecture, where should I start to snoop from?

Bartlomiej,
the things we need to know are in this FAQ:
http://www.parisc-linux.org/faq/kernelbug-howto.html

BTW, please don't compress files you attach - the mail archive can't search through them.

thanks,
grant
_______________________________________________
parisc-linux mailing list
[email protected]
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux