On Thu, Nov 13, 1997 at 11:43:15PM -0500, Gordon W. Ross wrote:
> > Date: Thu, 13 Nov 1997 21:56:40 -0500 (EST)
> > From: Mike Frisch <mfrisch@saturn.tlug.org>
> > 
> [...]
> > * Timeouts in the middle of the boot process:
> > 
> > nfs_boot: trying RARP (and RPC/bootparam)
> > nfs_boot: client_addr=0xc0a80004
> > nfs_boot: server_addr=0xc0a80001
> > nfs_boot: hostname=sparc.saturn.tlug.org
> > nfs_boot: timeout...
> > nfs_boot: timeout...
> > nfs_boot: timeout... 
> 
> Oh yeah, I meant to bring that up.  I don't remember seeing
> those timeouts until the most recent changes to the NFS boot
> support that reorganized things to reduce code duplication.

Those 3 timeouts (it's always three) are because the code asks
the bootparamd for a "gateway". Normally, this is not configured,
so what probably happens is that the bootparamd doesn't send a request,
which the client sees as a timeout, and gives up after 3 attempts.

That's what I think anyway, I just bumped into this myself 2 days ago.

- Frank