I have set up a DHCP server on CentOS 7. It runs beautifully, however due to business demands, we are expanding to another site far away. This gives us the opportunity to host a failover DHCP server at this seperate location. The two locations will be connected via a site-to-site VPN.
In the above image is a summary of our network. We will have the Perth and Melbourne sites, connected over the Internet via a VPN connection. Each site is given a different IP range to operate off. Each site will have their own DHCP server.
I would like to configure the DHCP servers as following:
DHCP 1 should serve as the primary DHCP server for Perth. That is, PC1, PC2 and PC3 will get their IPs off DHCP 1.
DHCP 2 should serve as the primary DHCP server for Melbourne. PC4, PC5 and PC6 will get their IPs off DHCP 2.
In a case where DHCP1 goes down, DHCP2 would take over the Perth site's IP allocation until DHCP1 comes back online, and vice versa. So each DHCP server shall serve as failover for the alternate location.
The only issue I can see here is that the DHCP server can not differentiate between clients in the Perth site, versus clients in the Melbourne site.
Imagining DHCP2 were not responding for some time, what is stopping DHCP 1 from assigning a Melbourne system with a Perth IP? And once this DHCP configuration is in place, how can I avoid the remote DHCP server acknowledging local requests unless it is in this failover mode?