Re: Working around problems when an NFS server dies / is unavailable.

On Sat, Jan 04, 2003 at 04:52:48PM -0800, richard childers / kg6hac wrote:
> I seem to recall a 'nomount' option in fstab(5), the manual page that
> describes the contents of the /etc/fstab ('filesystem table') file.
> 
> That plus amd(8) should, in theory, get you a relatively stateless NFS
> connection.

I must admit, I've never worked with amd but it may be worth a try -- it
can't be any worse than the situation I'm in at the moment.  Thanks for
bringing it to my attention ;)

> I've had some success using 'umount -f' to forcibly umount NFS filesystems,
> from the client side, in the past.

This surprised me.  I had assumed a forced umount would actually force
the umount, yet, it seems not to have done.  I am confuzed as to what is
and what isn't a bug (often I'm screwing things up myself, I'm sure).
Do you think this would count as a bug?  If other people have had
problems too, it almost suggests it might be.

> However, NFS, while stateless in intent, is really not a good
> infrastructural element for a plug-and-play network. Plugging in is easy;
> unplugging may require an explicit shutdown to properly deallocate
> resources.

Yeah, I'm starting to learn this now.  I don't really know what else to
use.  I've heard of CODA and even AFS but I don't know their suitability
to my purpose.  I think I might do a bit more research then ask again on
this list.

  Thanks for your information, I'll go read about amd ;)

-lewiz.

-- 
All true wisdom is found on T-shirts.
------------------------------------------------------------------------
--|| url: http://lewiz.info/ | http://www.westwood.karoo.net/pgpkey ||--

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (FreeBSD)

iD8DBQE+GBgrItq0KFQv7T8RAqf8AJ4tER/9ipTa7Juecsv1jn8woqK3WwCgmwKb
zDavIMAJYcr/2m9eAH5YaGM=
=Acyb
-----END PGP SIGNATURE-----