Why does a small crisis only ever happen when you have 200 or so machines to image and a looming deadline ?
Had a ‘brown afterburner’ moment today when my production FOG server (Ubuntu 12.04 LTS ‘Alternate’ 64 bit / Fog 1.2 stable release with FOG doing the dhcp and tftp serving) all of a sudden just decided it wasn’t going to hand out any more IP addresses thank you very much. All the services were running happily and a restart didn’t clear the problem.
It turned out to be that the dhcpd.conf file had a very limited scope - just 25 addresses, and a default lease time of around 7 days. The scope was probably my fault because I’d set the IP address of the Fog server to something daft - 192.168.1.75 instead of something more sensible like 192.168.1.10, but the very odd thing was that rebooting the server didn’t seem to release the previously offered dhcp leases ? Wierd. Editing the dhcpd.conf file to more sensible numbers has fixed the issue.
Anyway, just in case this helps somebody in future…