I'm installing a HA multiserver setup. I want to have servers (not real IPs): web1 192.168.0.1 - web2 192.168.0.2 (Cluster IP 192.168.0.10, drbd mounted /dev/drbd0 /var/www); mysql1 192.168.0.3 - mysql2 192.168.0.4 (Cluster IP 192.168.0.20, drbd mounted /dev/drbd0 /where mysql is stored ); mail1 192.168.0.5 - mail2 192.168.0.6 (Cluster IP 192.168.0.30, drbd mounted /dev/drbd0 /var/vmail); I'm using pacemaker to change the drbd block device and the cluster IP between nodes (web1 - web2,...). The router will send requests to the cluster IPs. I know how to configure the master-mirror server setup that web2 is mirror of web1. But now I need the ISPConfig interface on both web1 and web2 servers and that it mirrors the data to the sescondary server that can be web1 or web2. Pacemaker defines what server is primary or secondary. Should I install 2 sets of master web servers(multiserver setups) and connect the other servers by cluster IPs and replicate on every server the mysql files so that the web1 and web2 have the same dbispconfig table? CREATE USER 'root'@'CLUSTER_IP' IDENTIFIED BY 'myrootpassword'; GRANT ALL PRIVILEGES ON * . * TO 'root'@'CLUSTER_IP' IDENTIFIED BY 'myrootpassword' WITH GRANT OPTION MAX_QUERIES_PER_HOUR 0 MAX_CONNECTIONS_PER_HOUR 0 MAX_UPDATES_PER_HOUR 0 MAX_USER_CONNECTIONS 0 ; But if I do like that what to do with the user that connects with the hostname... sholud I make 2 users for every 1 cluster IP. Is it even possible to do such a thing?