Re: proc stalls

From: Denis Vlasenko
Date: Wed Sep 08 2004 - 13:56:20 EST


On Wednesday 08 September 2004 18:56, Tommy Reynolds wrote:
> Uttered Yaroslav Halchenko <yoh@xxxxxxxxxxxxxxxxxxxxxx>, spake thus:
> > that problem was linked to the fact that nfs-mounted directory became
> > unavailable...
> > Any ideas on how to further debug this situation to avoid future
> > problems?
>
> This is the required behavior for "hard" NFS mounts. NFS doesn't
> deal with servers that drop off-line very well.
>
> Perhaps you should use the "soft" and/or the "timeo=N" value. A
> "soft" mount will not cause your client to hang if the server goes
> away. Unfortunately, this also has implications for application
> program's ideas about file integrity, but there you go.

Think about some very important work being run on NFS-mounted file,
and server is brought down while you're at lunch. I much prefer
client to hang forever (i.e., no 'soft' option!), waiting for admin
to take action.

I use 'hard,intr' so admin can kill 'hung' processed if (s)he wants to.
--
vda

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/