>| ran fine until I did: [root_at_crush /root]# venus
>|
>| Date: Sat 01/30/99
>|
>| 19:35:50 Probably another Venus is running! open(/dev/cfs0) failed
>| (1), exiting
>
>Unmount /coda, then try again.  It is not possible to reattach to an

[root_at_crush i386]# umount /coda
umount: /coda: not mounted

Coda doesn't appear to be mounted:
[root_at_crush i386]# cat /etc/mtab
/dev/hdd2 / ext2 rw 0 0
none /proc proc rw 0 0
/dev/hdd3 /home ext2 rw 0 0
/dev/hdd4 /usr ext2 rw 0 0
/dev/hda1 /backup ext2 rw 0 0
/dev/hdb2 /space ext2 rw 0 0

>already mounted coda fs, and although it is automatically mounted when
>venus starts, it is not unmounted when venus exits.
>
>I believe that the reason for this is that it is impossible to unmount 
>while some program still has something in the filesystem opened. So 
>when the unmount fails, it is likely that there some shell that has its
>current working directory in /coda.

This is not the case - no venus was running so I couldn't have a shell
in the coda namespace.

The relevant portion of /sbin/lsmod is:
Module                  Size  Used by
coda                   50488   1  (autoclean)

Which suggests there is really something attached.  When I kill every
coda related process on this machine 'used' still doesn't drop to 0.

-John Langford