Subject: Re: net boot problems
To: David Laight <email@example.com>
From: Andrey Petrov <firstname.lastname@example.org>
Date: 02/01/2003 13:02:35
On Sat, Feb 01, 2003 at 08:35:06PM +0000, David Laight wrote:
> On Sat, Feb 01, 2003 at 12:20:50PM -0800, Andrey Petrov wrote:
> > Sounds as nfs server isn't actually started, make sure you
> > have rpcbind, mountd and nfsd running.
> > Andrey
> > > After the 'Using BOOTPARMS protocol' message, there is a bit of nfs traffic,
> > > lookup fh 2048,1/2 "netbsd" this is GENERIC
> > > read fh 2048,1/3 1024 bytes @ 0
> > > read fh 2048,1/3 1024 bytes @ 1024
> > > read fh 2048,1/3 1024 bytes @ 2048
> > > read fh 2048,1/3 1024 bytes @ 3072
> > > read fh 2048,1/3 64 bytes @ 52
> > > read fh 2048,1/3 440 bytes @ 3071820
> > > All the responses are 'ok'.
> NFS reads of the kernel...
I had very similar log not long time ago when i mistyped some in
rc.conf, rebooted server, and found myself in troubleshooting why
my client doesn't boot anymore. nfsd wasn't started.
The easiest test you can do, boot solaris, and mount your remote root
directory from there. If you have other machines, no need to boot
> I've since made more/less (maybe different) progress.
> The timeout of the bootp requests can be removed by running a dhcp server
> (I should know that, but 'man bootp' really ought to link a dhcp page).
> However the netbsd dhcp server seems unwilling to report the 'server path'
> option - which I think needs setting to the base of the exported filesystem.
> (I don't fancy running nfsd chrooted....)
> I've searched the source (parts are trully horrid) but the only reference
> to the server path option is in a header file. Maybe it can be specified
> by number?
Don't know, I use bootparamd for my sun clients.