Hi! I need to set up a high availability system. I was thinking of using two SANs and two regular high performance servers. Each of the high performance servers would run two virtual machines: one for HTTP services and another for MySQL data. Load would be balanced between them. Now, about the SANs I was thinking about using DRDB to create a network RAID1 (Master/Slave style) and connect them to my servers using ISCSI. My questions about this are: Where is the heartbeat service going to run? On each SAN? Or is it going to run on the servers? HOW do the servers know where to connect? If they are setup to connect to the master and the master simply crash, will they know how to connect to the secondary server? Thanks PS: to summarize what I need: I have two identical servers that need to load balance between them. They will connect to a SAN and in case that SAN fails, they need to automatically connect to the secondary SAN.
This might be what you need: http://www.howtoforge.com/openfiler...-heartbeat-drbd-with-offsite-replication-node You have a virtual IP (failover IP) that is managed by heartbeat and that is automatically switched to the slave if the master fails.
