| 摘要 |
A system and method for recovering from a server failure in a computer network, wherein the network contains several stand-alone, non-clustered, servers, and a cluster, wherein a clustered server also serves as the spare server, is disclosed. This cluster will have one standby recovery group for each non-clustered server in the computer network. Each recovery group contains the IP address and network name of the associated stand-alone server. The cluster monitors the health of the stand-alone servers, preferably through the use of a heartbeat mechanism. If the cluster detects a failure, it will reassign the LUNs owned by the failing server to the cluster. After the cluster has reassigned the LUNs, it will activate the recovery group containing the IP address and network name of the failing server. Subsequently, the cluster will assume the identity of the failing server and serve its users, until the failing server is repaired or replaced.
|
