Bug #23

NFS panic

Added by swildner over 8 years ago. Updated over 7 years ago.

Status:ClosedStart date:
Priority:NormalDue date:
Assignee:-% Done:

0%

Category:-
Target version:-

Description

Hi,

I'm having actually two problems. All machines involved are running
1.5.0-DEVELOPMENT.

PROBLEM 1
---------
When I set nfs_server_enable to "yes" on one of my test machines and
export a few directories (which reside on FAT32 partitions) I get
reproducable panics on that machine as soon as I try to write from
another machine. Reading seems to work.

I've taken the following picture: http://yoyodyne.ath.cx/tmp/LPIC0019.JPG

The top line reads: "panic: lockmgr: locking against myself"

I would have liked to provide a crash dump and everything but
unfortunately...

PROBLEM 2
---------
For some strange reason I can't produce crash dumps on that box.

Here's what happens:

db> call dumpsys

dumping to dev #ad/0x20041, offset 2097536
dump ata4: resetting devices .. done
1023

And then the machine just sits there.

Sascha

History

#1 Updated by dillon over 8 years ago

:Hi,
:
:I'm having actually two problems. All machines involved are running
:1.5.0-DEVELOPMENT.
:
:
:PROBLEM 1
:---------
:When I set nfs_server_enable to "yes" on one of my test machines and
:export a few directories (which reside on FAT32 partitions) I get
:reproducable panics on that machine as soon as I try to write from
:another machine. Reading seems to work.
:
:I've taken the following picture: http://yoyodyne.ath.cx/tmp/LPIC0019.JPG
:
:The top line reads: "panic: lockmgr: locking against myself"
:
:I would have liked to provide a crash dump and everything but
:unfortunately...

The problem is virtually guarenteed to be msdosfs not playing nice
with the NFS server. Probably a bug in the msdosfs lookup code not
releasing a vnode lock that it should be releasing. I will try to
recreate it after christmas.

I'm assuming that an NFS export of a UFS filesystem works fine?

:PROBLEM 2
:---------
:For some strange reason I can't produce crash dumps on that box.
:
:Here's what happens:
:
:db> call dumpsys
:
:dumping to dev #ad/0x20041, offset 2097536
:dump ata4: resetting devices .. done
:1023
:
:And then the machine just sits there.
:
:Sascha

Don't know about that one. hmm. Maybe its deadlocked on a vnode,
try setting sysctl kern.sync_on_panic to 0.

-Matt
Matthew Dillon
<>

#2 Updated by swildner over 8 years ago

Matthew Dillon wrote:

> :PROBLEM 1
> :---------
> :When I set nfs_server_enable to "yes" on one of my test machines and
> :export a few directories (which reside on FAT32 partitions) I get
> :reproducable panics on that machine as soon as I try to write from
> :another machine. Reading seems to work.
> :
> :I've taken the following picture: http://yoyodyne.ath.cx/tmp/LPIC0019.JPG
> :
> :The top line reads: "panic: lockmgr: locking against myself"
> :
> :I would have liked to provide a crash dump and everything but
> :unfortunately...
>
> The problem is virtually guarenteed to be msdosfs not playing nice
> with the NFS server. Probably a bug in the msdosfs lookup code not
> releasing a vnode lock that it should be releasing. I will try to
> recreate it after christmas.
>
> I'm assuming that an NFS export of a UFS filesystem works fine?

Yes, UFS works fine.

> :PROBLEM 2
> :---------
> :For some strange reason I can't produce crash dumps on that box.
> :
> :Here's what happens:
> :
> :db> call dumpsys
> :
> :dumping to dev #ad/0x20041, offset 2097536
> :dump ata4: resetting devices .. done
> :1023
> :
> :And then the machine just sits there.
> :
> :Sascha
>
> Don't know about that one. hmm. Maybe its deadlocked on a vnode,
> try setting sysctl kern.sync_on_panic to 0.

No, that didn't help.

Sascha

Also available in: Atom PDF