Well.... Disabling iptables temporarily has resolved the issue. But this is not a solution, just confirmation that iptables is the cause. Not sure how I can be more permissive than opening tcp and udp ports 1024-65535
I've wiresharked the connection and I see them talking to one another on port 9999. Why the time out?
Confirmed - disabling iptables resolves the issue. This is not acceptable tho.
What access is required to join a slave into the master domain?
For the domain mode, port 9999 will be necessary, so you will need to allow outgoing and allow incoming on that port on both nodes. Are all the other necessary ports in
BTW For clustering itself you will typically need to allow multicast. So something like:
[root@localhost ~]$ iptables -t filter -A INPUT -i eth0 -d 18.104.22.168/4 -j ACCEPT