How to boot an LTSP client?

[Eamonn Sullivan]

On 16/10/05, Eamonn Sullivan <eamonn.sullivan at gmail.com> wrote:
> On 16/10/05, Oliver Grawert <ogra at ubuntu.com> wrote:
> > hi,
> > Am Sonntag, den 16.10.2005, 15:34 +0100 schrieb Eamonn Sullivan:
> > > What I'm stuck on is a *really* basic concept -- how to PXE boot a
> > > client. I have the server on my normal desktop. I have another Windows
> > > desktop (the previoulsy mentioned one, directly connected to the
> > > router) and an IBM ThinkPad T21 laptop with a wireless card (Netgear
> > > WG511v2 PC Card). Both have an option to boot from the LAN in their
> > > bioses. But if I choose that, it doesn't work -- both boot up into
> > > Windows as normal.
> > PXE only works via wired networks, i guess the BIOS setting on your
> > laptop applies to your internal NIC... try with a wire connected ;)
> > PXE then looks for a dchp server on boot and requests the kernel via
> > tftp from the source thats configured in the settings of the dhcp
> > server. after booting up the kernel it mounts your rootfs readonly from
> > the server via nfs and starts up a X login manager...
> >
> > ciao
> >         oli
>
> Thanks. It always happens that I struggle with something for a couple
> of days and then figure it out 20 minutes *after* posting to
> Ubuntu-users-list...
>
> Anyway, I now realize I need to choose "Intel Pre-boot environment" on
> the thinkpad. When I try that when wired it finds the server, loads
> the kernel, goes through the hardware recognition process and then,
> when it finishes executing something called /scripts/nfs-preboot, I
> get an endless stream of Read: Connection Refused. If I CTRL-C out of
> those I get a message saying the server doesn't have NFS over TCP and
> I'm left at an ash prompt.
>
> I've checked /etc/exports (it's sharing /opt/ltsp  *(ro,no_root_squash,async)
> , and anyway, it must have loaded the kernel from somewhere), checked
> the firewall settings (even turned it off at one point) and my
> hosts.allow looks like this:
>
> portmap:        192.168.0.0/24
> rpc.mountd:     192.168.0.0/24
> rpc.statd:      192.168.0.0/24
> in.tftpd:       192.168.0.0/24
>
> I still end up stuck in the same place. What else should I check?
>
> (I'm hoping the magic of posting to the list will work here. I should
> discover my error in about 19 minutes....)
>
> -Eamonn
>

See, right on schedule: portmap was being started (as is does by
default) to only work over the loopback interface. I'm posting this
from my new thin client... Too bad it has to be wired.

-Eamonn