CARP failover events triggered for no obvious



  • Hi,

    I have a problem with a SG-8860 1U in a CARP-Cluster with another SG8860 1U.

    The CARP-Status of the Interfaces changes from one moment to another resulting in a brief network timeout of 1 to 3 pings for no obvious reason. Both Pfsense or up and working, but they change the role of master and member (Carp events are visible in the Logfile) on random interfaces. The change of master and member does not include all interfaces at the same time but only some, resulting in a "split brain" situation where the active Pfsense has some CARP-Interfaces as master and the passiv Pfsense has also some of the CARP-Interfaces as master.

    The state of changing from master to member would happen in less than 5 seconds and then the changes are "rolled back" so the active one would get the CARP-Interfaces back as "master". This event will happen round about two or three times a minute.

    If you have fourther questions please let me know.


  • LAYER 8 Netgate

    Almost certainly an interruption at Layer 2. Packet capture for CARP on both nodes on an interface that malfunctions.

    Compare the heartbeats sent by the MASTER node against those received by the BACKUP node right before the BACKUP assumes MASTER.



  • @Derelict said in CARP failover events triggered for no obvious:

    those
    Hallo,

    thanks for your reply.

    I have some layer 2 errors on the switch (spanning-tree). I will try to fix the errors and provide feedback as soon as possible, but I only have "downtime" at Friday to test my configs.

    Thanks for your help.


Log in to reply