Package com.opencloud.rhino.facilities.clusterstate

Interface ClusterStateChangeListenerV3

  • public interface ClusterStateChangeListenerV3

    Listener interface implemented by resource adaptors to receive cluster state change callbacks.

    This interface and its callback behaviour is identical to ClusterStateChangeListenerV2 except that it expresses node IDs as String values instead of integer values, for compatibility with the pool clustering mode.

    Since:
    Rhino 3.2

    • Method Detail

      • clusterStateChange

        void clusterStateChange​(String[] clusterMembers,
                        String[] running,
                        String[] stopping)

        Notification that the cluster has changed state with respect to the resource adaptor entity that attached the listener. This callback has slightly different semantics depending on whether the node is running in SAVANNA or POOL clustering mode as described below.

        Rhino Savanna Clustering Mode A running node is a node that satisfies the following conditions:

        • the node is a member of the cluster’s primary component

        • the resource adaptor entity is in the Active state on the node

        • the SLEE has an actual state of Running on the node

        A stopping node is a node that satisfies the following conditions:

        • the node is a member of the cluster’s primary component

        • the resource adaptor entity has an actual state of Stopping on the node, and/or the SLEE has an actual state of Stopping on the node

        A node appears in the running set once it satisfies the conditions specified above. The node moves from the running set to the stopping set if the actual state of the resource adaptor entity or the SLEE transitions to Stopping on that node. Once the actual state of the resource adaptor entity becomes Inactive or the actual state of the SLEE becomes Stopped on that node, the node no longer appears in either of the running or stopping sets.

        Rhino Pool Clustering Mode A running node is a node that satisfies the following conditions:

        • the node is an OPERATIONAL live (or assumed live) member of the pool

        • the SLEE has an actual state of Running on the node

        A stopping node is a node that satisfies the following conditions:

        • the node is an OPERATIONAL live (or assumed live) member of the pool

        • the SLEE has an actual state of Stopping on the node

        A node appears in the running set once it satisfies the conditions specified above. The node moves from the running set to the stopping set if the actual state of the SLEE transitions to Stopping on that node. Once the actual state of the SLEE becomes Stopped on that node, the node no longer appears in either of the running or stopping sets. The actual state of the resource adaptor entity itself is ignored in pool mode.

        Parameters:
        clusterMembers - the current set of cluster members. Only event router nodes that have completed bootup initialisation tasks are included in this set.
        running - the current set of cluster nodes where the resource adaptor entity is running.
        stopping - the current set of cluster nodes where the resource adaptor entity is stopping.
        Since:
        Rhino 3.2