[26115] in Athena Bugs
Re: linux 9.3.11: matlab
daemon@ATHENA.MIT.EDU (Camilla R Fox)
Wed Oct 13 17:09:49 2004
Message-Id: <200410132107.i9DL7EqJ012163@red-herring.mit.edu>
To: Alex T Prengel <alexp@mit.edu>
In-Reply-To: Your message of "Wed, 13 Oct 2004 12:14:12 EDT."
<200410131614.i9DGECbN016516@astrophel.mit.edu>
Date: Wed, 13 Oct 2004 17:07:14 -0400
From: Camilla R Fox <cfox@mit.edu>
cc: bugs@mit.edu
Errors-To: bugs-bounces@mit.edu
I've been testing from my w92 workstation by running matlab -tty in
my screen session. It starts within a few seconds, and ping times to
various fileservers all look good.
I moved my afs homedir to a fileserver in w92, just to be sure; I can
still reproduce the problem reliably.
Cleaning out my .sw_messages hasn't helped, either; if I move it aside,
it gets created empty next time I run matlab.
(I'd heard scattered complaints a week or two ago about homedirs on
andromache feeling slow; I'd concluded at the time that a full backup
running was just enough to make some of our more sensitive users notice,
and that the balancer had concentrated the most vigorous afs users on
that fileserver.)
If it's a slowness problem, shouldn't I get the token no matter what,
but perhaps get the message despite already having the token?
(I think Jacob was describing a slow-afs problem, while mine is a
different problem.)
-Camilla
