I am trying to set up a split-site clustering with Pacemaker. Each site is
likely be served by a different ISP and plan is to have N-M config with 20
active nodes and 2 passive nodes. In the event of a failover, data needs to be
replicated over to the new active node. The clients will be configured with a
single IP address(a set of clients per each active node) and will have to stay
that way. Basically, the failover should be transparent for the clients. I am
not clear on how the IP addressing would work and how to configure this setup.
Is this possible with Pacemaker? I have seen documentation about split-site
supported by Pacemaker. However, it is not clear if the nodes need to be part
of the private network connected by fiber links or something or they can be
connected via WAN.
Also, in case nodes being part of a private network is a requirement, can I
overcome this with configuring a VPN.
And have iptables direct the public IP to the floating private IP. That ways,
client will always communicate with
the known public IP.
Any help/pointers to documentation will be greatly appreciated.
_______________________________________________
Pacemaker mailing list: [email protected]
http://oss.clusterlabs.org/mailman/listinfo/pacemaker
Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker
