[K12OSN] workstation boot issues - not typical

["Kris A. Wotipka" <kris wotipka com>]

First, kudos to Paul and the gang as well as the Red Hat crew for the 
explosion that this project has seemed to have.  I have been out of the loop 
for a few months but by the size of my inbox, I would say that more people 
are definitely working on this.


We are having a voodoo boot issue here with our new installation.  We had a 
similar issue with the 1.0 install and I really do not think it is a K12 
issue, but something we have not done correctly given our setup.

here it goes:

K12LTSP installed typical fashion.

eth1: external public IP 199.80.xx.xx
eth0: internal private IP's

now to allow some other win based machines to act as terminals at times, we 
have "bridged" the two eth cards.  (they are plugged into two ports on the 
SAME switch). 

Now I can hear the moaning and groaning from here...........   but this use 
to work for us.  I believe that we added a 127.0.0.1 as the third DNS and it 
made booting from the workstations almost 100%

When we replaced 1.0 with 2.0, after the usual setup, we could not get any 
workstations to boot.  Our test workstation which is connected via a switch 
to the ltsp eth0 would finally boot if we took down eth1.  If eth1 was up, it 
would hang on "mounting the root file system" after the dhcpc line came up.  

Server log looks like:

Mar ?1 15:59:03 ltsp-server dhcpd: DHCPDISCOVER from 00:04:5a:57:b3:9b via 
eth0
Mar ?1 15:59:03 ltsp-server dhcpd: DHCPOFFER on 192.168.0.1 to 
00:04:5a:57:b3:9b via eth0
Mar ?1 15:59:05 ltsp-server dhcpd: DHCPREQUEST for 192.168.0.1 
(192.168.0.254) from 00:04:5a:57:b3:9b via eth0
Mar ?1 15:59:05 ltsp-server dhcpd: DHCPACK on 192.168.0.1 to 
00:04:5a:57:b3:9b via eth0
Mar ?1 15:59:05 ltsp-server rpc.mountd: authenticated mount request from 
192.168.0.1:687 for /opt/ltsp/i386 (/opt/ltsp/i386) 
Mar ?1 15:59:36 ltsp-server last message repeated 15860 times
Mar ?1 16:00:37 ltsp-server last message repeated 35142 times
Mar ?1 16:01:38 ltsp-server last message repeated 35506 times
Mar ?1 16:02:39 ltsp-server last message repeated 35547 times
Mar ?1 16:03:39 ltsp-server last message repeated 35506 times
Mar ?1 16:04:41 ltsp-server last message repeated 35772 times
Mar ?1 16:05:42 ltsp-server last message repeated 35839 times
Mar ?1 16:06:43 ltsp-server last message repeated 35898 times

The other test workstation which is 3 switches away from the server sees the 
DHCP server, gets an ip request 192.168.0.2 and then sleeps after the loading 
kernel line.

We have gotten the second terminal to boot several times with either eth1 
down, ipchains stopped, etc.  

All of this smells of a loop somewhere or resolving issues..........  
everything appears to be correct in the usual DNS, gateway, etc.

Here are some other specs:

Server name:  ltsp-server
domain:  ltsp

Sometimes when the workstations hang, it appears that they are trying to send 
traffic towards the router / internet / eth1 rather than eth0 (looking at the 
nice blinkie lights on the switches).

thanks in advance

kw