MASTER DHCP w/ CARP not issuing IPs while SECONDARY is powered off

  • I have a two node CARP cluster, with three interfaces (PSYNC/CARP, WAN, and LAN). DHCP is configured for the LAN interface only. Everything was working for years until I took down the SECONDARY due to the pending arrival of Irma. As soon as I took it down I started getting DHCPDISCOVER reporting "…via hn0: not responding (startup)" in the logs. Then....

    Sep 6 21:16:12 dhcpd failover peer dhcp_lan: I move from startup to communications-interrupted

    On the MASTER I had the Failover Peer IP set to the static IP address of the SECONDARY LAN interface (and vice versa), and as soon as I removed that entry, the MASTER started issues IP addresses again. The desired state is that this is automatically failed over if the SECONDARY goes offline/unresponsive, but I guess I never accomplished that and no one noticed as both routers have been rock solid.

    Am I missing something or is this working as designed?

    Thanks in advance.


  • Rebel Alliance Developer Netgate

    If both are up and in a 'normal' status and then one fails, the other will keep handing out addresses.

    If DHCP is restarted on the primary while the secondary is offline, it can never reach a failover peering state where it believes it should be able to hand out addresses.

  • That makes sense. So there is no way to have the failover peer added to work during "normal" operations AND have it work during a long-term peer outage?


  • Rebel Alliance Developer Netgate

    No, because failover is not designed to work with a long-term peer outage. Multiple things will fail (xmlrpc sync, for example)

    If you will have a long outage you'll have to manually remove some of the sync settings until the peer is repaired.

Log in to reply