SUMMARY: Jumpstart and TFTP

Thanks to Darren Dunham (gee that was quick!)

My comments inline as to what I actually changed.


> > Hi,
> >
> > I have set up a Solaris 9 Jumpstart server running on Solaris 9 that I 
> have
> > used once before a couple of months back (for host fw). This machine does
> > nothing else (it is my workstation that I don't really use) so I am 
> certain
> > that nothing has been modified.
> >
> > I tried to add another jumpstart client by typing "add_install_client -d
> > elearn sun4u". The machine is a SunFire V120 so I am sure sun4u is 
> right. I
> > have added a dhcp entry to my dhcp server and restart that. I have added
> > entries to both the NIS maps file and /etc files on the jumpstart server
> > containing the following entries:
> > ethers:
> > 0:3:ba:35:c8:96 fw.domain.com
> > 0:3:ba:35:c0:8c elearn.domain.com
>
>Are you doing DHCP jumpstart or traditional jumpstart?  If you're doing
>DHCP jumpstart, then the command should have added everything to the
>DHCP server running on the jumpstart server itself.  Also, the
>/etc/ethers file would be unused.

Apparently I was trying to use both :) I disabled the DHCP entry on the 
server tho and simply entered "add_install_client elearn sun4u".

>If you're using traditional jumpstart, why are you messing with the -d
>flag and DHCP servers?
>
> > I have tried removing the nis entries and working off local files only and
> > tried removing the local entries and working off nis. My nsswitch.conf has
> > "nis files" for all of these three files.
> > When I type "boot net - install" on "elearn" it says "Timeout waiting for
> > ARP/RARP packet" twice and then does nothing.
>
>You're specifing a traditional boot (not DHCP).  I think you don't want
>to use the -d flag on your add_install_client.
>
> > I have checked that both tftpd and rarpd are running on the jumpstart 
> server
> > root 29496     1  0 10:39:05 ?        0:00 /usr/sbin/in.rarpd -a
> > root 29243   154  0 10:24:20 ?        0:00 in.tftpd -s /tftpboot
>
>Why is tftpd running?  Normally thats started only on demand by inetd.

I didn't notice it adding the tftpd entry into inetd.conf (by default I 
usually disabled everything in there), so I commented out the tftpd stuff 
in /etc/init.d/nfs.server and stopped and started it. Yes as Darren 
mentioned I was running tftpd by hand and through inetd.

> > I have tried stopping and starting nfs.server script (which includes rarpd
> > and tftpd) many times.
> >
> > I saw a post to this list a while back about this issue but didn't see a
> > conclusion or followup.
>
>I would ...
>
>1) Verify the hardware address in /etc/ethers is not in there twice
>    (with two different names).
>2) run in.rarpd in debug mode (kill the one running and then start
>    /usr/sbin/in.rarpd -d -a by hand.  See what output appears when the
>    client is booted.
>3) Possibly run 'snoop 0:3:ba:35:c0:8c' on the server to see if the
>    client's requests are reaching the server.

Kelly Ormsby BSc(Curtin)
Systems Administrator
WATRI\ATCRC
University of Western Australia
http://www.watri.org.au
Room G03, 39 Fairway
Nedlands

Phone : +61 8 9380 8110
Email : kelly@watri.org.au 
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers