Application failover in a cluster environment
AssigneeEMC Corporation
InventorsAparna Bharthulwar, Odell Walker, Jr.
Life of this patent
Estimating the term…
Abstract
Cluster unaware application failover in a clustered environment. A cluster typically includes a shared disk and at least two nodes referred to as an “active node” where the application is processed, and a “passive node” capable of processing the application in the event of an “active node” failure. The method includes identifying active node failures using a cluster manager, activating application resources from the shared disk, transferring the application resource services to the passive node, and processing the application from the passive node.
Claims22
In a clustered environment that processes an application, the clustered environment including a shared disk connected with at least two nodes, a method of failover for a cluster unaware application in the clustered environment in the event of an node failure, the method comprising:
- identifying a failure of a first node using a cluster manager, wherein the first node provides application resources to a cluster unaware application and wherein a configuration of the first node includes a key registry that identifies each process of the cluster unaware application to the cluster manager;
- identifying the key registry for each process of the cluster unaware application;
- activating application resources for the cluster unaware application from the shared disk;
- protecting the shared disk from corruption during failover;
- transferring the application resources to a second node;
- and
- processing the application from the second node.
A method of failover for a cluster unaware application in a clustered environment comprised of a shared disk and at least a first node where the application is preferably processed, and a second node capable of processing the application in the event of a first node failure, the method comprising:
- configuring a first node and a second node in a clustered environment for a cluster unaware application to enable failover of the cluster unaware application, wherein a configuration of the first node and the second node includes a key registry;
- identifying a failure of the first node using a cluster manager, wherein application resources are associated with a cluster unaware application operating on the first node;
- and
- failing over the cluster application to the second node by performing:
- identifying the key registry for each process of the cluster unaware application;
- activating application resources from a shared disk connected to the first node and to the second node using the key registry;
- protecting the shared disk from corruption during failover;
- transferring the application resources to the second node;
- and
- processing the application from the second node.
A method of failover for a cluster unaware application in a clustered environment comprised of a shared disk and at least two nodes referred to as an “active node” where the application is processed, and a “passive node” capable of processing the application in the event of a “active node” failure, the method comprising:
- identifying active node failures using a cluster manager, wherein a cluster unaware application is processing through the active node and wherein the cluster unaware application is associated with application resources that have been previously configured on each node in the cluster for the cluster unaware application, wherein a configuration of each node includes a key registry;
- activating the application resources from the shared disk, wherein the passive node is able to recognize the cluster unaware application based on the previous configuration of each node by identifying the key registry for each process of the cluster unaware application;
- protecting the shared disk from corruption during failover;
- transferring the application resources to the passive node;
- processing the cluster unaware application from the passive node;
- detecting availability in the active node;
- activating the application resources from the shared disk;
- transferring the application resources back to the active node;
- and
- processing the cluster aware application in the active node.
Dependencies are inferred from the claim text.