I'm observing a puzzling problem. It seems that the DHCP request sent
out by the thin client is subtly different the first and second time
(two DHCP requests are sent during a single boot). The second DHCP
request contains more information, and the DHCPD server sees this as a
different request. Even if the MAC address is the same, they have
different uids and so a different IP is handed out. These two stanzas
from a leases file shows this behaviour:
lease 192.168.199.153 {
starts 3 2005/05/25 09:56:32;
ends 4 2005/05/26 09:56:32;
tstp 4 2005/05/26 09:56:32;
binding state active;
next binding state free;
hardware ethernet 00:50:8b:50:c1:85;
}
lease 192.168.199.245 {
starts 3 2005/05/25 09:56:43;
ends 4 2005/05/26 09:56:43;
tstp 4 2005/05/26 09:56:43;
binding state active;
next binding state free;
hardware ethernet 00:50:8b:50:c1:85;
uid "\001\000P\213P\301\205";
An extract from syslog shows that the server hands out different IPs to
the same client during boot:
May 25 11:56:25 ltspserver dhcpd: DHCPDISCOVER from 00:50:8b:50:c1:85
via eth2
May 25 11:56:26 ltspserver dhcpd: DHCPOFFER on 192.168.199.153 to
00:50:8b:50:c1:85 via eth2
May 25 11:56:26 ltspserver dhcpd: DHCPDISCOVER from 00:50:8b:50:c1:85
via eth2
May 25 11:56:26 ltspserver dhcpd: DHCPOFFER on 192.168.199.153 to
00:50:8b:50:c1:85 via eth2
May 25 11:56:28 ltspserver dhcpd: DHCPDISCOVER from 00:50:8b:50:c1:85
via eth2
May 25 11:56:28 ltspserver dhcpd: DHCPOFFER on 192.168.199.153 to
00:50:8b:50:c1:85 via eth2
<some seconds delay while kernel is transferred and client boots it>
May 25 11:56:42 ltspserver dhcpd: DHCPDISCOVER from 00:50:8b:50:c1:85
via eth2
May 25 11:56:43 ltspserver dhcpd: DHCPOFFER on 192.168.199.245 to
00:50:8b:50:c1:85 via eth2
May 25 11:56:43 ltspserver dhcpd: DHCPREQUEST for 192.168.199.245
(192.168.199.254) from 00:50:8b:50:c1:85 via eth2
May 25 11:56:43 ltspserver dhcpd: DHCPACK on 192.168.199.245 to
00:50:8b:50:c1:85 via eth2
This client was booted several times during the morning, and only these
two IP's were used.
For installations that use dynamic IPs for the thin clients, there will
be a shortage of IPs when the number of clients pass half the number of
dynamic IPs.
I am trying out the statement "deny duplicates;" in the dhcpd.conf file
to see if that cures the problem.
I know one way of solving this is to use static IP's. But that is not
always convenient. There must be a reason why the client sends out
differently formatted DHCP requests, perhaps that could be changed.
The installations I have seen this particular problem use Debian Sarge
with the DHCPD server package
dhcp3-server 3.0.1-2
Both LTSP 4.1 and 4.1.1 show the same thing.
--
Med vennlig hilsen
Ragnar Wisl�ff
-------------
life is a reach. then you gybe.
-------------------------------------------------------
SF.Net email is sponsored by: GoToMeeting - the easiest way to collaborate
online with coworkers and clients while avoiding the high cost of travel and
communications. There is no equipment to buy and you can meet as often as
you want. Try it free.http://ads.osdn.com/?ad_idt02&alloc_id135&op=click
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net
