[1426] in linux-net channel archive
Re: Strange behaviour with NFS (fwd)
daemon@ATHENA.MIT.EDU (Robert Glamm)
Sat Nov 25 07:16:15 1995
From: Robert Glamm <glamm@mountains.ee.umn.edu>
To: linux-net@vger.rutgers.edu
Date: Fri, 24 Nov 1995 23:46:33 -0600 (CST)
> Um, this isn't limited to Linux - Solaris also exhibits this
> behavior. When one of our NFS servers dumps the big one, we often
> have to reboot the clients. Someone suggested that the mounts be
> changed from 'hard' to 'soft', but I don't think this would make
> any difference.
>
> Soft NFS mounts are dangerous because I/O operations will fail on any
> timeout. This can cause file corruption (if programs don't check
> return status), unnecessary program failures, and possibly nasty
> program aborts if paging in a binary fails.
>
> Mounting hard and interruptable is generally the best bet. Rebooting
> the NFS server also generally brings things back to life.
I guess I can't see how a dead soft NFS mount can lead to file corruption,
unless the server dies and comes back to life in a fairly short time
& the program doesn't check the return status of a sequence of writes.
At the supercomputer institute I used to work at we mounted everything
soft because dead hard NFS mounts would tend to kill all users on the
machines, not just a few of them (e.g., having multiple home directories
mounted via NFS). I think data loss only occurred maybe once during the
period of time I was there (18 months) over multiple unscheduled hangs
of the server machines. ;)
--
Bob Glamm | "You can't do a `goto' to a block
Email: glamm@mountains.ee.umn.edu | that has been optimized away.
URL: http://www.cs.umn.edu/users/glamm | Darn."
Home: (612)623-9437 Work: (612)625-7876 | - from the perltrap(1) manpage
