Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 20 Dec 2010 16:59:54 -0500
From:      Adam McDougall <mcdouga9@egr.msu.edu>
To:        freebsd-stable@freebsd.org
Subject:   Re: Stale NFS file handles on 8.x amd64
Message-ID:  <4D0FD1DA.8090209@egr.msu.edu>
In-Reply-To: <20101204233906.GY82436@egr.msu.edu>
References:  <4CF5DEC4.3070901@egr.msu.edu>	<1918825226.1032577.1291246386604.JavaMail.root@erie.cs.uoguelph.ca> <20101204233906.GY82436@egr.msu.edu>

next in thread | previous in thread | raw e-mail | index | archive | help
On 12/04/10 18:39, Adam McDougall wrote:
> On Wed, Dec 01, 2010 at 06:33:06PM -0500, Rick Macklem (and others) wrote: (various suggestions)
>

I had to call off the experimentation with live users and resorted to 
running all IMAP connections from a single NFS client for now.  I was 
causing a little too much of a stir.  Here are the things I can remember 
trying:

negnametimeo=0 (no perceived difference)

acregmin=0,acregmax=0,acdirmin=0,acdirmax=0 (slowed down imap but 
otherwise no perceived difference)

stopped statd and lockd after changing dovecot to use dotlock (no 
perceived difference)

vfs.lookup_shared=0 (no perceived difference)

dovecot 1.1 (no better than 1.2 regarding this issue)


Did not try (yet?):

udp, nfsv4, freebsd 7

I feel compelled to keep the users happy for now while I look into the 
nfs director included with dovecot 2.0 designed to keep individual IMAP 
users on their own server, so I'm taking a stab at the recommended usage 
with newer software rather than more debugging with previous versions. 
I may return to other tactics if the future path slows down,
I'm still curious why things changed.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4D0FD1DA.8090209>