[clue-tech] nfs timeouts

[mike havlicek]

--- On Sat, 4/11/09, Angelo Bertolli <angelo.bertolli at gmail.com> wrote:

> From: Angelo Bertolli <angelo.bertolli at gmail.com>
> Subject: Re: [clue-tech] nfs timeouts
> To: "CLUE technical discussion" <clue-tech at cluedenver.org>
> Date: Saturday, April 11, 2009, 5:16 PM
> Michael J. Hammel wrote:
> > On Tue, 2009-04-07 at 17:28 -0400, Angelo Bertolli
> wrote:
> >   
> >> Michael J. Hammel wrote:
> >>     
> >>> I think you can do this by setting the "hard"
> and "intr" options on the
> >>> mount command.
> >>>         
> >> I was looking at hard and soft.  The wording
> make me think that maybe soft is the one I wanted (since
> hard is the default).
> >>     
> > 
> > No, you want hard in conjunction with intr so that you
> can interrupt the
> > operation but not risk data corruption.  Soft has
> the potential of
> > having data corruption.  The man page describes
> this.
> >   
> 
> Hmmm, after reading this, I don't think either hard or soft
> actually unmounts the fs.  It looks like all these do
> is send messages to the software that the resource is
> unavailable.  Which actually might be just as
> good.  But then I would like the system to report a
> problematic mount.  I was going to do this by testing
> to see what was mounted, but it won't work if it doesn't
> automatically unmount.
> 
> I didn't see the part about  risking data corruption
> with soft.  In fact, it is not clear what the
> difference between soft and intr are:  they both report
> errors to the calling program.  I guess intr lets the
> program decide to stop trying?
> 
> It's also unclear as to whether or not the option retrans
> refers to cumulative timeouts, or successive timeouts. 
> I think successive minor timeouts makes more sense.  It
> seems I might want to set this one also.
> 
> Angelo
> 

You might also look at the retry,fg,bg options mentioned in man nfs(5).

-Mike
> _______________________________________________
> clue-tech mailing list
> clue-tech at cluedenver.org
> http://www.cluedenver.org/mailman/listinfo/clue-tech
>