Coda File System

Re: venus death/codasrv problem?

From: Tom Tarka <tommy_at_mp3.com>
Date: Thu, 03 Feb 2000 21:52:05 -0800
Jan Harkes wrote:

> On Thu, Feb 03, 2000 at 07:55:13PM -0800, Tom Tarka wrote:
> > I'm running a codaserver (v. 5.3.4) on a PowerPC MkLinux machine.
> >
> > When I start venus on an intel RHL machine pointing at this server,
> > it pretends to connect, then just dies.  However, when I point at the
> > testserver, I have no problems -- so I assume it's a server issue.
> >
> > Is there anything I can do to better find out what's happening?
> >
> >     -t
> ...
> > 17:40:40 Initial LRDB allocation
> > 17:40:40 Getting Root Volume information...
> > 17:40:42 Venus starting...
> >
> > Program terminated with signal SIGKILL, Killed.
> > The program no longer exists.
>
> That's the child process that tries to mount /coda killing its parent
> because the mount failed.
>
> - Can you insmod /lib/modules/.../coda.o without problems?

ayup.


> - Did you unmount /coda after shutting down venus. Sometimes the umount
>   fails because some processes are still holding on to dentries/inodes
>   in the kernel.

[root_at_agrippa tommy]# umount /coda
umount: /coda: not mounted
[root_at_agrippa tommy]# /sbin/insmod -v
/lib/modules/2.2.14-1.1.0swan/fs/coda.o
[root_at_agrippa tommy]# ps au|grep venus
root      2774  0.0  0.1  1240  492 pts/0    S    19:57   0:00 grep venus
[root_at_agrippa tommy]# /usr/sbin/venus -init &


ayup.

                -t
Received on 2000-02-03 22:52:52