摘要 |
A computing system method for high-availability operation is disclosed. Implementations of the system include a plurality of host computers connected together in a network, each host computer running an HA process to monitor activity on the host computer. The system further includes a master HA process, designated from all other HA processes running on the plurality of host computers. The master HA process includes a heartbeat monitor to monitor activity of the network to detect a failed host computer, and a failover handler to transfer at least one of the one or more tasks from the failed host computer to an active host computer. All other HA processes other than the master HA process are used as cache for read requests to topology data.
|