[pnfs] wireshark patches

[J. Bruce Fields]

On Mon, Nov 26, 2007 at 08:13:40PM +0200, Benny Halevy wrote:
> http://www.bhalevy.com/pnfs/cthon04-g-nfs41.trace.bz2
> contains traces from a non-pnfs run of the connectathon test general tests
> over a localhost loopback mount.

Thanks!  Any idea what the "Unreassembled Packet" thing wireshark's
giving on frame 24 (a CREATE_SESSION reply) means?  (Or the two
"Continuation Data" frames right before and after?)

--b.

> 
> Benny
> 
> On Nov. 26, 2007, 19:05 +0200, "J. Bruce Fields" <bfields at fieldses.org> wrote:
> > On Mon, Nov 26, 2007 at 05:02:10PM +0000, J. Bruce Fields wrote:
> >> On Mon, Nov 19, 2007 at 10:50:25AM -0500, bfields wrote:
> >>> On Thu, Nov 15, 2007 at 02:43:55PM -0500, J. Bruce Fields wrote:
> >>>> On Wed, Nov 14, 2007 at 06:43:23PM -0500, bfields wrote:
> >>>>> Eh, I don't know if I was nuts that day, or looking at some ancient
> >>>>> trace, or what, but I can't reproduce the setclientid or
> >>>>> setclientid_confirm problems now that I made time for a longer look.
> >>>>>
> >>>>> There are a few other odd problems--I think what happened is that at
> >>>>> some point somebody took a tree with the changes and diff'ed it against
> >>>>> an upstream tree that was slightly newer than the one the changes were
> >>>>> based on, so the initial monster-patch reverts a few things it shouldn't
> >>>>> have.  I've almost got that fixed, so I'll try to have a clean
> >>>>> cumulative patch by tommorow that we can submit....
> >>>> OK, I've fixed up those accidental reverts.  Here's the patch I intend
> >>>> to submit.  I assume it's missing some attributions; let me know if you
> >>>> did some work on it.
> >>> I didn't hear any objections, so I've submitted it as is:
> >>>
> >>> 	http://bugs.wireshark.org/bugzilla/show_bug.cgi?id=1995
> >> I got a few comments back, which I'll deal with, but one was a request
> >> for a sample 4.1 trace for testing purposes; could someone please send
> >> me a good representative trace?
> > 
> > And if the pFNS xdr is still changing too fast, maybe it should be a
> > trace that just demonstrates sessions, or something else that's fairly
> > stable.
> > 
> > --b.
> > _______________________________________________
> > pNFS mailing list
> > pNFS at linux-nfs.org
> > http://linux-nfs.org/cgi-bin/mailman/listinfo/pnfs
> 
> 
> -- 
> Benny Halevy
> Software Architect
> Tel/Fax: +972-3-647-8340
> Mobile: +972-54-802-8340
> US:      +1-412-203-3187
> bhalevy at panasas.com
>  
> Panasas, Inc.
> Leader in Parallel Storage
> www.panasas.com