On Tue, May 23, 2000 at 03:23:02PM -0700, Linda Walsh wrote:
Is it just 'down', or should there never be an 'unkillable'
user-level process? I.e. -- in SunOS, you could mount NFS as
intr, or not. Basically if you were mounted 'hard' and didn't
specify 'intr', you couldn't interrupt a process with kill -9
that was trying to access a hard mounted filesystem. So even if
your sever went down -- you couldn't kill it. This is consistent
with the unix paradigm of a single fs write being indivisible.
I've seen processes get stuck under high IO load myself, it's a bug
(or rather was, I've not seen it for some time). I had no NFS on
these machines.
Generally, if something gets stuck in 'down' for a very long time,
then something is probably broken (exceptions like NFS do exist
presumably).
Does Linux not have a similar semantic for NFS? Would it
have such a 'special case' anywhere else?
umount2 allows you for forcibly unmount a file-system, would this not
be better?
--cw
-
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/
This archive was generated by hypermail 2b29 : Tue May 23 2000 - 21:00:25 EST