Hello,
In a failover configuration, all clients send traffic to and receive traffic from server 1. If server 1 goes offline, all clients send traffic to and receive traffic from server 2 until server 1 comes back online.
Load balancing occurs between the servers assigned to Priority 1 in a Management Server list. If more than one server is assigned to Priority 1, the clients randomly choose one of the servers and establish communication with it. If all Priority 1 servers fail, clients connect with the server assigned to Priority 2.
Load balancing servers
Load balancing is used to distribute client management between management servers.
Servers in the Management Server List that have the same priority are load balancing servers. When clients connect to the servers, they are distributed between the available servers with the same priority in order to distribute the load evenly. For example, if there are two servers with priority 1, the clients will be distributed between those two servers.
Only servers at the same site should be configured with the same priority level in the Management Server List. If management servers from different sites have the same priority, they are treated as load balancing servers. This causes clients to switch between different sites, and incurs the risk of data inconsistency.
Combining failover and load balancing
You can configure failover and load balancing by assigning priorities to management servers in Management Server Lists. Load balancing occurs between the servers assigned to the highest priority in a Management Server List. Servers with lower priority are failover servers. If more than one server is assigned to Priority 1, each client randomly chooses one of the servers and establishes communication with it. If all Priority 1 servers are unavailable, clients connect with the failover servers that are assigned to Priority 2.
If you use the Embedded DB instead of Microsoft SQL, only one manager can be added to each site. In this case, only replication partners are available to use as failover and load balancing servers. Note that this does incur the risk of data inconsistency.
Reference: http://www.symantec.com/docs/TECH104519
https://www-secure.symantec.com/connect/forums/failover-concept
Hope that helps!!