[snip]
> > getcwd() runs up the directory tree, if its search algorithm crosses
> > the disconnected NFS mount point it tries to access the remote system.
> > Not good when a remote system drops off, even users who are not using
> > ocs3 directories are hit. Lots of tasks invoke getcwd().
>
> I can see that this would be massively annoying, and in this case the
> attempts at revalidation don't make a lot of sense, as there's not much
> the client can do if something gets scrambled on the server.
The solution that I've introduced years ago (non-kernel related) is:
Mount the filesystems one level deeper. So instead of (simplified fstab):
server1:/localdisk/u1 /users/u1
server2:/localdisk/u2 /users/u2
server3:/localdisk/u3 /users/u3
server4:/localdisk/u4 /users/u4
you use:
server1:/localdisk/u1 /users/server1/u1
server2:/localdisk/u2 /users/server2/u2
server3:/localdisk/u3 /users/server3/u3
server4:/localdisk/u4 /users/server4/u4
and
cd /users
ln -s server1/u1 u1
ln -s server2/u2 u2
ln -s server3/u3 u3
ln -s server4/u4 u4
for "backwards" compatibility.
Roger.
-- | Most people would die sooner than think.... | R.E.Wolff@BitWizard.nl | in fact, most do. -- Bertrand Russsell | phone: +31-15-2137555 We write Linux device drivers for any device you may have! fax: ..-2138217- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/faq.html