Carp is Switching serval times a day



  • Hi at all,

    i have a small Setup with 2 Pfsense 2.3.1 (actual version)
    From the beginnen on i have a lot of trouble with CARP Failover.

    I see a lot of thinks like that:

    carp: VHID 5@igb2: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: VHID 1@igb0: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: demoted by 240 to 480 (interface down)
    carp: VHID 4@igb1: BACKUP -> MASTER (master down)
    carp: VHID 2@igb1: BACKUP -> MASTER (master down)
    carp: VHID 6@igb2: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: demoted by 240 to 480 (interface down)
    carp: VHID 4@igb1: BACKUP -> MASTER (master down)
    carp: VHID 2@igb1: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: VHID 1@igb0: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: demoted by 240 to 480 (interface down)
    carp: VHID 4@igb1: BACKUP -> MASTER (master down)
    carp: VHID 2@igb1: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: VHID 1@igb0: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: demoted by 240 to 480 (interface down)
    carp: demoted by 240 to 720 (interface down)
    carp: VHID 6@igb2: BACKUP -> MASTER (master down)
    carp: VHID 5@igb2: BACKUP -> MASTER (master down)
    carp: VHID 3@igb2: BACKUP -> MASTER (master down)
    carp: VHID 5@igb2: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: VHID 1@igb0: BACKUP -> MASTER (master down)
    carp: VHID 6@igb2: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: demoted by 240 to 480 (interface down)
    carp: VHID 4@igb1: BACKUP -> MASTER (master down)
    carp: VHID 2@igb1: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: demoted by 240 to 240 (interface down)
    carp: demoted by 240 to 240 (interface down)
    carp: demoted by 240 to 480 (interface down)
    carp: demoted by 240 to 720 (interface down)
    carp: demoted by 240 to 960 (interface down)
    carp: demoted by 240 to 1200 (interface down)
    carp: demoted by 240 to 1440 (interface down)
    carp: demoted by 240 to 240 (interface down)
    carp: demoted by 240 to 480 (interface down)
    carp: demoted by 240 to 720 (interface down)
    carp: VHID 6@igb2: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: demoted by 240 to 480 (interface down)
    carp: demoted by 240 to 720 (interface down)
    carp: VHID 6@igb2: BACKUP -> MASTER (master down)
    carp: VHID 6@igb2: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: demoted by 240 to 480 (interface down)
    carp: demoted by 240 to 720 (interface down)
    carp: VHID 6@igb2: BACKUP -> MASTER (master down)
    carp: VHID 5@igb2: BACKUP -> MASTER (master down)
    carp: VHID 3@igb2: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: VHID 1@igb0: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: VHID 1@igb0: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: VHID 1@igb0: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: VHID 1@igb0: BACKUP -> MASTER (master down)
    carp: demoted by 240 to 240 (interface down)
    carp: VHID 1@igb0: BACKUP -> MASTER (master down)

    After 1-2 Seconds it switched backup to Master. It seems there i some problems with igmp or somethink like that.

    Any ideas how i can debug that issue?
    Maybe its to much broadcast because i have a big Proxmox Cluster which sends also a lot of broadcast for cluster messages.

    I Am Thinking to rebuild my network with 2 bgp routers and just do norma routing with 2 serperated networks. All clients use LAGG interfaces then active/pasive

    Somethink like that because CARP seems not working.

    Cheers

    Daniel


  • Netgate

    It looks like (as expected) carp: interface down is an actual carrier loss on the interface. Here are two CARP events on a backup node. The first is a unplug/plug of the patch cable. The second is removing the interface from the same VLAN as the master node, which leaves the interface up but stops CARP advertisement reception.

    Jun 16 07:08:03 pfSenseB kernel: carp: demoted by 240 to 240 (interface down)
    Jun 16 07:08:17 pfSenseB kernel: carp: VHID 66@igb1: INIT -> BACKUP
    Jun 16 07:08:17 pfSenseB kernel: carp: demoted by -240 to 0 (interface up)

    Jun 16 07:13:02 pfSenseB kernel: carp: VHID 66@igb1: BACKUP -> MASTER (master down)
    Jun 16 07:13:36 pfSenseB kernel: carp: VHID 66@igb1: MASTER -> BACKUP (more frequent advertisement received)

    What is your switch logging on the switchport in question? I'd take a good look at your switching and cabling.