摘要:
Methods and systems are directed to managing a protocol network failure in a network clustering system. A cluster master and each client member in the cluster are configured to determine a primary protocol network, a secondary protocol network, and the like, for exchanging a cluster protocol message. Each client member in the cluster determines its ‘active protocol network’ from at least the primary protocol network, and the secondary protocol network when it joins the cluster. Each client member tests its connectivity to the cluster master through the primary protocol network. If a client member detects a failure on the primary protocol network, it dynamically fails over to the secondary protocol network, or the like, without leaving the cluster. If connectivity to the primary protocol network is restored, the client member dynamically resumes the protocol message exchange on the primary protocol network without leaving the cluster.
摘要:
A system, apparatus, and method are directed to managing multicast routing using a network cluster. A cluster master actively participates in multicast control protocol communications with the cluster neighbors, while non-master member nodes, do not participate in the protocols' operation outside of the cluster. However, each cluster member maintains virtually the same protocol state, such that should the master become unavailable, another cluster member may assume the master role with minimum delay. New members to the cluster are synchronized by receiving initial protocol state information from the master. After the initial protocol state synchronization, cluster members actively monitor the network to acquire new protocol state information. In addition, cluster members exchange protocol state information between themselves that may not be obtainable off the network. Filtering of multicast data packets is performed by each cluster member, to load balance multicast data traffic across the cluster members.