We have a couple of Linux machines here at the lab. They mount some
directories from our RS/6000. The machines are running Linux 1.0.9,
but I would like to upgrade to 1.1 to get the newest libraries and GCC
2.6.2 to work.
Strangely NFS-mounts of the volumes on the RS/6000 tend to fail when
running 1.1.72 or 1.1.73 (possibly with earlier alpha kernels as well,
I haven't tried). The mounts succeed only randomly, apparently with a
probability of about 50%.
If I use options hard (default) or intr, this failure causes mount
to hang indefinitely. The hung mount process cannot be interrupted
(even if intr was used); one can only kill -9 it.
With the soft option, mount quits if it fails. In this case I get the
following messages (dopey is the RS/6000):
NFS server dopey not responding, timed out
NFS: notice message: result=-11
nfs_fhget: getattr error = 11
nfs_read_super: get root inode failed
With 1.0.9 NFS worked without problems. With the exception of NFS,
networking seems to work fine with 1.1.73. I would be grateful for
any clues about the reason for the NFS difficulties.