pfSense

When a primary syncs a CARP VIP that is IPv6 to a secondary, it will trip the secondary's IPv6 connectivity by forcing it to go to INIT status and then back to BACKUP. This happens every single time an XMLRPC sync event happens, but does not happen to IPv4 VIPs. This seems like XMLRPC is being too "enthusiastic" about it's syncing behavior for CARP VIPs in IPv6 and always ripping and rebuilding the VIPs, even when nothing has changed, forcing momentary connectivity interruptions on the secondary. This should likely be changed to mimic the IPv4 behavior where it only makes these kinds of changes when VIPs are added or removed.

Relevant log data:
Oct 25 18:46:22 php-fpm 602 /rc.carpbackup: HA cluster member "(2600:1700:5750:fa22:92ec:77ff:fe34:0004@vtnet0): (WAN)" has resumed CARP state "BACKUP" for vhid 2
Oct 25 18:46:22 php-fpm 48528 /rc.carpbackup: HA cluster member "(2600:1700:5750:fa22:92ec:77ff:fe34:0004@vtnet0): (WAN)" has resumed CARP state "BACKUP" for vhid 2
Oct 25 18:46:21 kernel carp: demoted by 0 to 0 (pfsync bulk done)
Oct 25 18:46:21 kernel carp: demoted by 0 to 0 (pfsync bulk start)
Oct 25 18:46:21 kernel carp: 2@vtnet0: INIT -> BACKUP (initialization complete)
Oct 25 18:46:21 kernel carp: 2@vtnet0: BACKUP -> INIT (hardware interface up)
Oct 25 18:46:21 php-fpm 601 /xmlrpc.php: waiting for pfsync...
Oct 25 18:46:21 check_reload_status 654 Carp backup event
Oct 25 18:46:21 check_reload_status 654 Carp backup event
Oct 25 18:46:21 check_reload_status 654 Syncing firewall