Method to revive and reconstitute majority node set clusters
View Patent ↗The method of the present invention is useful in detecting the lack of majority cluster nodes within a cluster, and reorganizes the cluster membership to allow for running with less than a majority of nodes connected. As nodes become alive, they are added to the cluster until the original majority is reached, which then allows for the cluster to return to normal operation. Thus, the method of the present invention allows for the revival and reconstitution of a majority node set of clusters.
1. A system where a cluster of servers constitute a multiplicity of cluster nodes which are all connected to each other to service multiple PC users via a public network comprising:
(a) shared storage means virtually providing a disk resource (Quorum) of stored data, said means being seen as a single quorum repository of a Microsoft Cluster Service Program;
(b) local quorum disk resource means which are positioned in each node of said cluster of servers and which operate as a single quorum;
(c) means to automatically synchronize the remaining nodes of the cluster when a node is disconnected or inoperative including:
(c1) means to reorganize said cluster nodes when the operable node count falls below 50% of the total node count;
(d) means to convert said multiplicity of cluster nodes into a single node with its own local quorum disk resource including:
(d1) means to sense a network outage which severs said cluster into two partitions;
(d2) means to enable the remaining group holding the majority of nodes to continue in operable activity;
(d3) means to automatically take off-line the non-majority group of nodes;
(e) means to troubleshoot said single node to find the cause of its inoperability, said means including:
(e1) means to test the activation signal common to each mode;
(e2) means to test the storage capability of each node;
(e3) means to test the remote processing capability of each node.
2. A cluster system of “M” server nodes operating with an (MSCS) Microsoft Cluster Service program and able to revive and reconstitute a majority node set cluster after a node failure comprising:
(a) a majority node set quorum resource servicing said server nodes when a majority of the cluster nodes are operating and are all in communication with each other;
(b) means to automatically bring back the cluster on-line after a failure causing non-utilization of one or more nodes said means including:
(b1) means to detect whether the system is operating as a majority node set, a shared disk quorum or as a local quorum, including:
(b1a) means to revive and put back online non-functioning server nodes;
(b1b) means to reconstitute sufficient server nodes to establish a majority node set cluster, including:
(b1ba) means for establishing an operating cluster where M/2+1 is the number of operational server nodes.
3. In a cluster system of “M” server nodes operating with an (MSCS) Microsoft Cluster Service Program, a method for reviving and reconstituting a majority node set cluster after a node failure comprising the steps of:
(a) executing a detection phase wherein a Cluster Verifier determines the type of clustering mechanism being deployed;
(b) instituting a Revival Phase, when the operating cluster nodes fall below M/2+1, in order to allow a user to manually restart the cluster;
(c) instituting an automatic Reconstitution Phase, when the prior non-operating cluster nodes are operational and ready to rejoin said cluster including the steps of:
(c1) terminating said Revival phase; and
(c2) reconnecting said newly restored operational nodes back into said cluster.