Description of problem: First of all, why don't you have dhclient in the list of components ? The system: Dell Latitude CPx with PIII700, 128MB RAM, ATI Mach 64, Pegasus-based USB ethernet The kernel: Vanilla 2.4.20 compiled so RedHat likes it (modular USB, sound, etc.) The software: RedHat 8.0 The story: For the longest time, and across 2.4.18, 2.4.19, and 2.4.20, I would leave my laptop on, and suspend it instead of powering it down. Then I would resume at another location, hotplug would fire up, set up eth0 (pegasus), DHCP an IP, and be fine. Then I got Internet service via a cable modem, service that includes 2 free IPs. So I bought a small hub, hooked up all the cables to get the connectivity right, and resumed my laptop. Sure enough, it DHCP-ed an IP via the hub, via the cable modem, from my ISP, and was fine (I could ping all the sites I wanted). But then, I did the unthinkable: I rebooted my laptop. When it came up, it hung on "Bringing up interface eth0". Of course, it couldn't grab an IP. After frelling around with the cable modem, hub, et. al., I tried running dhclient eth0 Much to my surprise, it worked. So, I thought all was well again, and rebooted. Of course, it hung on eth0 again. So I thought there was still something wrong with my cable modem/hub setup and I went onto another network (at the University) where the laptop used to grab IPs just fine. Well, it didn't work there either (at boot time). I tried going into runlevel 1 and doing ifconfig eth0 up dhclient eth0 That hung, so I hit C-c. I did dhclient eth0 again, and this time it worked. Conclusion: For some unknown reason, ever since my reboot at the time when I started using the hub with the cable modem, even on other networks, I have to run dhclient, C-c it, then run it again, if I want to grab an IP. I scripted this workaround into /etc/sysconfig/network-scripts/ifup like so: --- ifup 2003-01-15 14:57:36.000000000 -0500 +++ new.ifup 2003-01-15 14:57:15.000000000 -0500 @@ -183,6 +183,13 @@ FWHACK=1 fi + [ -x /sbin/dhclient ] && + { + /sbin/dhclient ${DEVICE} & + sleep 2 + /bin/kill -9 $! + } + if [ -x /sbin/dhclient ] && /sbin/dhclient ${DHCLIENTARGS} ${DEVICE} ; then echo $" done." elif [ -z "`pidof -x dhclient`" ] && [ -x /sbin/dhcpcd ] && /sbin/dhcpcd ${DHCPCDARGS} ${DEVICE} ; then Well, that's not exactly C-c, I know, but you can always replace kill -9 with kill -INT. I realize this is a dirty, ugly, rotten workaround, but it's the only one that seems to work. Version-Release number of selected component (if applicable): dhclient-3.0pl1-9 How reproducible: Every time Steps to Reproduce: 1. Boot up 2. Watch it hang at "Bringing up interface eth0", even though it had worked a million times before, on a myriad of different networks where DHCP was the means of grabbing an IP. 3. Reboot Actual results: rc hangs on 'network start' for a time exceeding 3 minutes (that's the longest I've waited for it to move on), namely on bringing up eth0 (pegasus based USB eth NIC) Expected results: Bringing up interface eth0: [ OK ] within about 5-7 seconds Additional info: None
Bugs are assigned by source package, not by binary subpackage. Reassigning to dhcp...
*** Bug 81973 has been marked as a duplicate of this bug. ***
I do not believe there is a bug here, but a setup problem. Please contact Red Hat support for help on diagnosing this problem. If you don't have a Red Hat support contract you might want to look at ethereal to see what is going on between the client and the server. to help diagnose the problem. Dan