booting thin clients and firestarter
gmccullagh at gmail.com
Sun Feb 3 22:59:04 GMT 2008
On Sun, 03 Feb 2008, DB Clinton wrote:
> > A dhcp boot normally looks like this for us:
> > Feb 1 08:18:57 brooks dhcpd: DHCPDISCOVER from 00:02:a5:77:46:6a via eth0
> > Feb 1 08:18:58 brooks dhcpd: DHCPOFFER on 22.214.171.124 to 00:02:a5:77:46:6a via eth0
> > Feb 1 08:18:58 brooks dhcpd: DHCPREQUEST for 126.96.36.199 (188.8.131.52) from 00:02:a5:77:46:6a via eth0
> > Feb 1 08:18:58 brooks dhcpd: DHCPACK on 184.108.40.206 to 00:02:a5:77:46:6a via eth0
> > Feb 1 08:18:58 brooks in.tftpd: tftp: client does not accept options
> > do you see all five of those lines in /var/log/syslog?
> Nope. The DHCPREQUEST, DHCPACK AND tftp: client does not... lines don't
> appear at all.
To me, the lack of these lines suggests something odd about your client, ie
you're not getting the DHCPREQUEST which should come from the client after
which the server should respond with a DHCPACK. Then when the tftp request
happens the little tftp warning goes in.
To do the network booting, are you using a PXE booting network card, a
floppy or what?
Can you do the same test again, with the "tail -f" monitoring the logfile
and also run this command:
sudo /usr/sbin/tcpdump -n -i eth0 udp port 67 or port 68
which will show us what dhcp packets get sent and received. As an example,
here's one I've just done:
gavin at hector:~$ sudo /usr/sbin/tcpdump -n -i ath0 udp port 67 or port 68
tcpdump: WARNING: ath0: no IPv4 address assigned
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on ath0, link-type EN10MB (Ethernet), capture size 96 bytes
22:50:34.004414 IP 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request from 00:09:5b:d2:f0:2f, length 300
22:50:34.236718 IP 192.168.0.1.67 > 255.255.255.255.68: BOOTP/DHCP, Reply, length 548
22:50:34.253600 IP 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request from 00:09:5b:d2:f0:2f, length 300
22:50:34.477935 IP 192.168.0.1.67 > 255.255.255.255.68: BOOTP/DHCP, Reply, length 548
> > > pxelinux.0 and nbi.img were actually not in that directory either. I
> > > copied them there but still no results.
> > Any thoughts on how that happened? I would have thought you can't have
> > had booting working ever without at least one of those being in place.
> I have no idea. It certainly wasn't anything I did manually as I wasn't
> even aware there was such a directory until today. Sometime before the
> "crash", I did install a few packages (in my effort to set up some
> controls over the internet access of my clients). I installed (and
> subsequently, post-crash, uninstalled) Firehol, Dansguardian and Squid. I
> think that was all.
I can't imagine those would make a difference. Ogra would know better what
creates them but I'm pretty sure you couldn't have working network booting
More information about the edubuntu-users