Failover not working for ICMP and UDP



  • Hi,

    I'm using the current config:
    ClientA - LAN - pfSense

    pfSense- WAN (NATed) - WANGW/WANEmu1 - ServerA
    pfSense - OPT1 (NATed) - OPT1GW/WANEmu2- ServerA

    WAN and OPT1 are in a single load balancing gateway group trigger = member down.

    http transparent proxying enabled.

    ClientA sends pings to ServerA and various other destinations
    ClientA sends UDP packets to ServerA

    When I set WANEmu1's uplink to 100% packet loss, pfSense eventually notices that WANGW is offline in Status/Gateways.

    HTTP and HTTPS connections are failed over but affected ICMP and UDP connections are not failed over even after many minutes. New ICMP connections will work and use the working gateway.

    Then I set WANEmu1's uplink back to 0% loss. Affected ICMP and UDP connections start working back again.
    pfSense will notice WANGW is online.

    Then I set WANEmu2's uplink to 100% loss. pfSense will notice OPT1GW is offline.

    HTTP and HTTPS connections are failed over but affected ICMP and UDP connections are not failed over even after many minutes.

    If I manually reset the state using Diagnostics/States the ICMP and UDP stuff gets failed over.

    The ICMP and UDP stuff also gets failed over if I delete their states using pfctl. But not if I only delete the states to and from the affected interface's IPs/subnet and leave the states without the NATed IPs.

    Is there a way to automatically reset the states for the WAN interface IPs/IP range when WANGW goes down and reset the states for the OPT1 interface IPs/IP range when OPT1GW goes down?


Log in to reply