[3161] in SIPB_Linux_Development
Re: nfs installer trouble
daemon@ATHENA.MIT.EDU (Theodore Y. Ts'o)
Tue Oct 31 13:12:12 2000
Date: Tue, 31 Oct 2000 13:12:10 -0500
Message-Id: <200010311812.NAA20084@tsx-prime.MIT.EDU>
From: "Theodore Y. Ts'o" <tytso@MIT.EDU>
To: Derek Atkins <warlord@MIT.EDU>
CC: "Theodore Y. Ts'o" <tytso@MIT.EDU>, Angie Kelic <sly@MIT.EDU>,
linux-dev@MIT.EDU
In-reply-to: Derek Atkins's message of 31 Oct 2000 11:34:26 -0500,
<sjmu29trmul.fsf@rcn.ihtfp.org>
From: Derek Atkins <warlord@MIT.EDU>
Date: 31 Oct 2000 11:34:26 -0500
Ted, the particular problem we see is that the NFS client hangs and
when we look at the client there is a log message about not being able
to get a slot.
Hmm.. The only time I've seen that message is when the server is down.
So I'd guess this probably has something to do with the fact that the
sipb subnet is rather busy, and perhaps the NFS client isn't dealing
with packet drops. I'll pass the information on to Dave and see what he
thinks.
As for trying out those new RPMS, upgrading the kernel on a redhat
boot disk is a challenging prospect, mostly because it requires
changes to the bootdisk as well as the install server.
Well, it would be useful to try out that kernel on a machine on MIT net,
and then try accessing the SIPB install server. Is there a reason why
we can't test out the NFS client without getting it onto the bootdisk?
Assuming that the newer kernels fixes the problem, then there's the
issue of changing the bootdisk, but let's confirm that supposition
first. :-)
- Ted
