OK. This is a very old problem but with a relatively new OS. I'm running CentOS 6.10 on one server with KeepAliveD and HAPROXY and the same on two other CentOS 7.x servers. I've created an HA cluster between them. All servers have these set:
net.ipv4.ip_forward = 1 net.ipv4.ip_nonlocal_bind = 1
The CentOS 7.x machines will allow HAPROXY to start without the IP address assigned to them; while the CentOS 6.x server will not.
Starting haproxy: [ALERT] 351/155831 (9970) : Starting frontend primary_port_80: cannot bind socket [:80] [ALERT] 351/155831 (9970) : Starting frontend primary_port_443: cannot bind socket [:443]
I can't be the only one that has seen this, so I am hoping that someone can help me out.
Thank-you!