Description of problem: After installing a second NIC, DHCP is run on startup when there's no signal on the cable. This adds 20 seconds or so to boot time. Version-Release number of selected component (if applicable): Fedora 7 test 3, updated as of March 30 Expected result: The DHCP daemon should see that there's no signal on the cable and terminate directly.
This is happening because of something in initscripts. The settings in /etc/sysconfig/network-scripts/ifcfg-* are being ignored for some reason and the network service is running DHCP for everything. Reassigning. See also bug #234764.
What's your ifcfg file?
There's a chance that this bug doesn't exist after all. The (secondary, eth1) NIC on which this error occurred is dead now, so maybe it was dying back then. Sorry for the wasted time in that case. If there's a problem, it's probably this: running "dhclient eth0" by hand waits forever when there's no signal, while running DHCP through Network Confguration and /etc/init.d/networking correctly terminates with the message that there's no signal on the cable. I don't have the original configuration files anymore alas.
dhclient doesn't do a link check, so it may wait forever (modulo whatever normal dhcp timeouts it has). The ifup script that calls dhclient is what has the link test. However, this particular issue (networking going strange on first bootup) has been reported enough that it's obviously a real issue.
*** This bug has been marked as a duplicate of 234764 ***