Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 8 Nov 1996 10:13:17 +0000 (GMT)
From:      Doug Rabson <dfr@render.com>
To:        Karl Denninger <karl@Mcs.Net>
Cc:        hackers@FreeBSD.org
Subject:   Re: Funny behavior in NFS
Message-ID:  <Pine.BSF.3.95.961108100234.26390B-100000@minnow.render.com>
In-Reply-To: <199611051605.KAA05286@Venus.mcs.net>

next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 5 Nov 1996, Karl Denninger wrote:

> > Don't you need the "intr" flag to be able to ^C out of an operation?  The
> > "soft" flag gives up the operation after a number of retries but isn't
> > necessarily interruptable
> 
> Yes, but you can sit with a "soft" mount for hours and it never times out..

I just tried to reproduce this and the soft, intr stuff works fine
for me.  What you may be seeing is that the code exponentially increases
the timeout each time any request times out.  This means that for a
moderately busy machine which loses its connection, the timeout will go
right up to the maximum (256 * timeout specified for mount_nfs) as the
first eight processes attempt NFS requests.

With the default timeout and retry count, this would mean that the
requests will take about 45 minutes to time out.  At any time during this
45 minutes, you can ^C the request (for an interruptable mount) and get
control back at most one second later.  For a single user machine, with
only one NFS request outstanding, the total timeout is about 15 minutes.
If this is not right for your site, you could try reducing the retrans
count down to 5 from the default value of 10.  For my testing, I had an
fstab line like this:

foo:/home	/home	nfs	rw,soft,intr,-x5 0 0

--
Doug Rabson, Microsoft RenderMorphics Ltd.	Mail:  dfr@render.com
						Phone: +44 171 734 3761
						FAX:   +44 171 734 6426




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.95.961108100234.26390B-100000>