Machine freezes frequently (nfs issue?)

Asked by Rolf Leggewie

my LAN server (edgy) is crashing all the time. First time I saw this is (after more than 100 days uptime!) when I was copying a large file over NFS from one of the clients. I discovered bug 71212, went back to kernel 2.6.15.28 from dapper and now I can copy large files over NFS again.

But the server still crashes frequently unless it is left completely alone. What else could I be looking for? The logs have nothing as for everybody else in bug 71212. I am at my wits end. Help very much appreciated.

Question information

Language:
English Edit question
Status:
Solved
For:
Ubuntu linux-source-2.6.17 Edit question
Assignee:
No assignee Edit question
Solved by:
Rolf Leggewie
Solved:
Last query:
Last reply:
Revision history for this message
Rolf Leggewie (r0lf) said :
#1

It does seem though, that nfs activity is more "successful" in triggering this problem. But If I can remember correctly, it is not always necessary.

Revision history for this message
williamts99 (williamts99) said :
#2

Have you done a memory test? Also can you watch the system monitor to see if any process is going to 100%?

Best Regards,
Williamts99

Revision history for this message
Rolf Leggewie (r0lf) said :
#3

Will, thank you very much for picking this up (Dankeschön!).

Trying to pinpoint the problem, I mounted the directory usually exported via NFS by sshfs now and unfortunately ran into the same trouble. To rule out hardware problems, I booted into a Knoppix Live CD. And indeed, I copied large amounts of files without any hang-ups ever. So I guess that this is after all a bug or possibly an issue with my installation.

I have not yet done a memory check but it looks like the hardware is OK since Knoppix runs fine. I will do a memory test, nonetheless to be sure. Thanks for pointing out that (quite obvious, duh) possibility of analysis.

Revision history for this message
Rolf Leggewie (r0lf) said :
#4

memory test ran for 2,5 hours without any problems. It passed tests five times.

Revision history for this message
williamts99 (williamts99) said :
#5

Would you be interested in testing this out on Feisty to see if you have the same issue? It's still in testing, but is planned for release in a couple of weeks.

http://www.ubuntu.com/testing

Best Regards,
Williamts99

Revision history for this message
Rolf Leggewie (r0lf) said :
#6

I think it is not the best idea to update when you ran into trouble from updating. I cannot get to my computer right now since it is about 10.000km away and I cannot log into it because of course it crashed shortly after I left it.

Revision history for this message
williamts99 (williamts99) said :
#7

That of course is your decision. Could you describe the method that you used to upgrade? What all applications did you install and method of installation? Also, could you post a copy of your sources.list file.

Best Regards,
Williamts99

Revision history for this message
Rolf Leggewie (r0lf) said :
#8

No particular method. Just aptitude. I can answer the rest of your questions only after getting back home.

Revision history for this message
Launchpad Janitor (janitor) said :
#9

This question was expired because it remained in the 'Open' state without activity for the last 15 days.

Revision history for this message
Rolf Leggewie (r0lf) said :
#10

I found out by chance that this was caused by the same buggy dash as another hard-to-track problem I was having on my laptop. Making sure that /bin/sh points to bash and not dash apparently made the problems go away. The machine has been running and performing tasks that previously crashed it for sure during the last 36 hours or so. Before this "dpkg-reconfigure dash" it would not last 15 minutes.

Revision history for this message
Rolf Leggewie (r0lf) said :
#11

sources.list only contains official ubuntu repos. I cannot completely rule out having ever installed a 3rd party package on this (or any other) machine.

Revision history for this message
Best Rolf Leggewie (r0lf) said :
#12

the machine is running stable now. I suggest reconsideration of dash as default sh. It does not seem to be ready quite yet.