I have implemented a MaaS based cluster for deploying OpenStack via Juju on some servers but also serving as DHCP for computers in a computer lab.
I managed to get OpenStack running scaled up to 8 servers and MaaS serving IPs in the same subnet without issue until very recently when the DHCP in MaaS stopped working. Now none of the computers have an IP address anymore and asking for renewal or release doesn’t seem to fix it. The OpenStack deployment is still working though but any other deployment fails because IP-config gives up.
I started inspecting packets in one of the computers using Wireshark and I’m getting NAK messages right before ACK from the DHCP. The NAK messages say IP address not available but I can see it in the DNS of MaaS so I guess the computer receives the NAK and ignores the following ACK. Why is my DHCP sending duplicate conflicting messages?
Both computers and servers are in the same subnet because we are limited by our router assigned IP range inside the university network.
Any help would be very appreciated! Thanks!