Actually the DHCP client on the box was setup during installation (pretty much the default there). As for the server, that was done by installing and enabling the server (via "yum install dhcpd" and "chkconfig ... dhcpd on" and "/etc/init.d/dhcpd restart") with a hand edited /etc/dhcpd.conf configuration file.

Basically the problem was that I unloaded the kernel module responsible for the network interface on the DHCP server and then loaded a different once and forgot to restart applicable services (DHCP) that bind direct to devices - thus leading to no DHCP on the reloaded network interface for a little over quarter of a day (until I restarted the DHCP server). Obviously the server issue was an oops on my part, but the question is why doesn't the client recover gracefully....

Cheers, MaZe.

On Tue, 17 Jan 2006, Anthony Lordi wrote:

Maciej. No answer to your loss of DHCP. Haven't even found the DHCP setup.

How did you set it up? Gui? Cli? I have CentOS 2.1, but imagine the same setup. Have tried wired and wireless and have had no connection. Tony

On 1/17/06, Maciej Żenczykowski maze@cela.pl wrote:

Hi,

I've got a box with CentOS 4.2 x86_64 which had no DHCP access for a few hours, now I can't ping it (I'm pretty sure the box is still up). I expect the DHCP lease has timed out and the box has lost it's IP, I was expecting it to reacquire it once the DHCP came back up (configuration problem) - but no luck. Is this a known problem? Any solution/workaround? (for now or for the future...)

Cheers, MaZe. _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos

It does if you have a 'retry' value assigned in /etc/dhclient.conf, however this file doesn't exist by default. I'm not sure what it does without this file, but I'd gather from your experience, it's nothing good.

I've read through dhclient.conf and I haven't found anything which would appear to govern what I think is happening - 'retry' value included. But maybe my interpretation is incorrect.

Here's what happened again:

a) client boot b) get DHCP - success c) a lot of time (days) passes with DHCP renews d) --- DHCP server is broken --- e) DHCP renewals fail (?) f) DHCP lease times out (?) g) IP address is deconfigured (?) h) --- DHCP server is repaired --- i) nothing happens... while I would expect: i) after sometime the 'n'th periodic DHCP retry finally reacquires an IP address...

[the retry statement has a default of 5 min, and even 24 hours haven't helped...]

Cheers, MaZe.

On Tuesday 17 January 2006 01:35 pm, Andrew Cotter wrote:

Anyone running on Sun gear? Wondering how the performance has been from your experiences and if there have been any issues.

I doubt you'll get any reasonable response; since you posted to the "DHCP without network access" thread the best you'll probably do is a response or two about SUN and DHCP. And the worst? A few more requests to not hijack threads.

Jeff

Chris Mauritz wrote:

Andrew Cotter wrote:

...

Anyone running on Sun gear? Wondering how the performance has been from your experiences and if there have been any issues.

Are you talking x86 or sparc? On the x86 boxes, it should be a rather uneventful install. Not sure about the current batch of sparcs though...

Cheers,

Not on their x86. More along the lines of 280R/V440/V880 systems. Just curious since my company has a number of these systems around.

Where can I find GFS refer?? Except centos.org

On Tue, 2006-01-17 at 15:33, Maciej Żenczykowski wrote:

Actually the DHCP client on the box was setup during installation (pretty much the default there). As for the server, that was done by installing and enabling the server (via "yum install dhcpd" and "chkconfig ... dhcpd on" and "/etc/init.d/dhcpd restart") with a hand edited /etc/dhcpd.conf configuration file.

Basically the problem was that I unloaded the kernel module responsible for the network interface on the DHCP server and then loaded a different once and forgot to restart applicable services (DHCP) that bind direct to devices - thus leading to no DHCP on the reloaded network interface for a little over quarter of a day (until I restarted the DHCP server). Obviously the server issue was an oops on my part, but the question is why doesn't the client recover gracefully....

Maybe it did but has a different IP address now? Or did you lock IP address assigned to the MAC address?