Failover explanation

marius

Hello everyone,
I am deeply interesting by how exactly pfsense works and since i have a failover for my two firewall, i was wondering what happen when my primary node (master) comes back from his tomb.

I've already read this https://forum.netgate.com/topic/119864/carp-ha-sync-and-xmlrpc-sync-explained/2 and from what i understood i can say this :

First message from the secondary node log, when i stop the primary one :
carp: 1@em3: BACKUP -> MASTER (master timed out)
Carp master event

Finally, the secondary node assumes the CARP MASTER status
/rc.carpmaster: HA cluster member "(192.168.168.254@em1): (CARP)" has resumed CARP state "MASTER" for vhid 3

This happened because the secondary node didn't receive the "heartbeat" from the one, if i'm not wrong.
But then, when the primary one come back, is it taking by force the master status ? or because it's finally sending advertisement to the secondary node, the secondary will understand he no longer has to be the master ?

I have these messages (still on the backup pfsense log):
carp: 1@em3: MASTER -> BACKUP (more frequent advertisement received)
Carp backup event
/rc.carpbackup: HA cluster member "(192.168.34.254@em3): (LAN)" has resumed CARP state "BACKUP" for vhid 1

I guess it works that way but I am not sure.
Plus and what confusing me a bit, is the log on the primary node when it comes back :

kernel: carp: 3@em1: INIT -> BACKUP (initialization complete)
first, it start as the backup, alright

kernel: carp: 3@em1: BACKUP -> MASTER (preempting a slower master)
it becomes the master, because the made the changes

kernel: carp: 3@em1: MASTER -> BACKUP (more frequent advertisement received)
but why this ?

kernel: carp: 3@em1: BACKUP -> MASTER (master timed out)
for finally take back MASTER status?

I don't know if it is normal or something wrong, if someone could explain to me what happened at the end I would really appreciate.
Thank you!

(sorry if I misexplained something, english isn't my main language)