Subject: Re: nfs-rooted machines
To: Chris Jones <email@example.com>
From: Guenther Grau <Guenther.Grau@bk.bosch.de>
Date: 07/13/1998 20:57:59
Chris Jones wrote:
> This is slightly off-topic, but I hope everybody will bear with me.
> I have a 1.3.2 i386 which is booting off a floppy and mounting its root
> via NFS from a Solaris 2.6 machine. At least, it *was* doing that, until
> I rebooted the Solaris machine last night. Now, when it boots, I get an
> error 13 (EACCES) when it tries to mount its root. I've verified that IP
> addresses are correct on both ends for server and client. When I run
> showmount -e on the server, the client shows up as able to mount that
> directory. I've got it exported as "rw=client,root=client", so I'd think
> it would be able to do anything it needed to on the filesystem.
> Strangely enough, I've got another 1.3.2 i386 box that, if I add it to the
> dfstab on the Solaris machine, can mount the same directory with no
> trouble. So what's the difference between these two machines? They're in
> the same subnet, and they're using the same (correct) subnet mask... I'm
> at a loss. What else can I do to debug this?
Well, you could run tcpdump on the Solaris machine and show us the
If you've got another machine, why not try to `rename' it's
to act as the first one. See if it acts the same.
Depending on the outcome, you might take further steps like dropping
into the kernel debugger, if it looks like it's a NetBSD problem.
Or enable exhaustive logging for NFS on the Sun machine, ...
Hope sth. sounds appealing to you :-)