[TriLUG] DHCP renewal problems

[Christopher L Merrill]

We have a stripped-down Fedora Live 7 distribution that basically has
nothing but networking, sshd and our software on it.  At one point,
our software makes some configuration changes and restarts the network
service.  When that happens, we see the usual messages on the console
about eth0 going down, coming back up and obtaining IP info (DHCP).

One one machine (an old PowerEdge 300), the restarting works fine.
The DHCP request is answered almost immediately. In /var/log/messages
I see:
dhcpclient: DHCPREQUEST on eth0 to 255.255.255.255 port 67
dhcpclient: DHCPACK from 192.168.1.1
dhcpclient: bound to 192.168.1.224

But on several newer machines (PowerEdge SC1420s), I see:
dhcpclient: DHCPREQUEST on eth0 to 255.255.255.255 port 67
(repeated several times)
dhcpclient: DHCPDISCOVER on eth0 to 255.255.255.255 port 67
(repeated several times)
dhcpclient: no DHCPOFFERS received
dhcpclient: trying recorded lease 192.168.1.203

Immediately after, we find that eth0 is not configured and
we have failures in some other scripts as a result.

Both machines are connected to the same switch on the same network...
and presumably talking to the same DHCP server (since we only have 1).
Note that the machines that fail can get a DHCP address when they
first boot.  5 minutes later, they cannot.  This is 100% repeatable
every time.

Any ideas?  I don't even know where to start googling...




-- 
------------------------------------------------------------------------ -
Chris Merrill                           |  Web Performance, Inc.
chris at webperformance.com                |  http://webperformance.com
919-433-1762                            |  919-845-7601

Website Load Testing and Stress Testing Software & Services
------------------------------------------------------------------------ -