I am testing haproxy with keepalive to 3 servers.
Haproxy server configurations work fine (USLTS4/USLTS5), but not keepalived.
The process started on both haproxy's, but the VIP is not assigned to the master (USLTS4), so not reachable from clients.
Here are haproxy and keepalived configs:
Both haproxy's work fine separately and loadbalance traffic
USLTS4:~$ netstat -nlta | grep :80 tcp 0 0 0.0.0.0:80
0.0.0.0:* LISTENUSLTS5:~$ netstat -nlta | grep 80 tcp 0 0 0.0.0.0:80
0.0.0.0:* LISTEN
No VIP assigned to the main haproxy
Tested with unicast and multicast (default) configurations.
unicast_peer {
192.168.20.254
}
Are you deliberately setting both keepalived instances to be MASTER? Usually the passive node's state would be BACKUP.