Hi Everyone, I've just pushed the changes to the DHCPD resource agent to my repository, that makes the following changes:
- Removes the required part of the chrooted_path variable. - Adds a default entry "/var/lib/dhcp" as a reasonable value to the chrooted_path variable. The agent has been able to recover itself under all tests I have thrown at it so far, except for one corner that should be extremely rare. The corner case is as follows (non-clustered): "A DHCP server process is live, the network interface is live, but no DHCP requests are being responded to, even though the server is fully responsive in all other aspects." I have only ever run into this issue once in 10 years of running DHCP servers under different platforms, and as such have not been able to reproduce it. Otherwise, I am happy with the agent itself, it has been running almost 2 months now with out any hiccups on a cluster I have, and has successfully failed over to secondary nodes on each occasion it has needed to do so. I believe it is time to open the floor up for additional testing and review. _______________________________________________________ Linux-HA-Dev: [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev Home Page: http://linux-ha.org/
