We are running DHCP Failover on 2 Debian/Jessie servers. Version 4.3.1 of ISC DHCPD is running on both servers. The gateway router of our access network relays the DHCP communication to both DHCP-servers.
Some clients don't come online when both servers are running. In the /var/log/syslog I see a normal DISCOVER coming in on both servers and one server replying with an OFFER. So far so good. After that I see the REQUEST coming in on both servers but then both servers replying with an ACK of the same IP. It sames that some clients don't like receiving an ACK from 2 servers at the same time and therefor send the REQUEST again. In that cases the REQUEST and dual ACK keeps repeating and the client never comes online. When I stop the server that didn't OFFER the IP the client is coming online after the first REQUEST to an ACK from the remaining server. The problem is in a normal synchronised operation. Not all clients seems to bother about receiving 2 ACK's, but some do. Can anybody tell me what's going wrong?
_______________________________________________
dhcp-users mailing list
[hidden email]
https://lists.isc.org/mailman/listinfo/dhcp-users
Locked
