RMI-IIOP request failover mechanism

An application server cluster system, comprising:

• two or more application server instances in a cluster;
• and
• an IIOP or RMI-IIOP client of one of the application server instances, wherein the client is configured to:
• obtain a cluster-aware remote reference to an object on the application server instance from the application server instance, wherein the cluster-aware remote reference indicates a primary endpoint for accessing an instance of the object on the application server instance and one or more alternate endpoints for accessing other instances of the object on the other application server instances;
• detect a communication failure for a request to the primary endpoint for accessing the instance of the object on the application server instance;
• and
• failover the request to one of the alternate endpoints on one of the other application server instances in response to said communication failure.

A system, comprising:

• a processor;
• and
• a memory comprising program instructions, wherein the program instructions are executable by the processor to implement an IIOP or an RMI-IIOP client application for accessing a cluster comprising a plurality of application server instances, wherein the client application is configured to:
• obtain a cluster-aware remote reference to an object on the application server instances from the cluster, wherein the cluster-aware remote reference indicates a primary endpoint for accessing an instance of the object on one of the application server instances and one or more alternate endpoints for accessing other instances of the object on other application server instances in the cluster;
• detect a communication failure for a request to the primary endpoint for accessing the instance of the object on the application server instance;
• and
• failover the request to one of the alternate endpoints on one of the other application server instances in response to said communication failure.

A system, comprising:

• a processor;
• and
• a memory comprising program instructions, wherein the program instructions are executable by the processor to implement an application server instance in a cluster of application server instances, wherein the application server instance is configured to:
• receive a failed-over request to an object on a failed application server instance in the cluster from a client of the failed application server instance;
• and
• restore a state of the object on the failed application server instance to the application server instance in response to said failed-over request received from the client.

A method, comprising:

• providing a cluster-aware remote reference to an object on an application server instance in a cluster of two or more application server instances to an IIOP or RMI-IIOP client of the application server instance, wherein the cluster-aware remote reference indicates a primary endpoint on the application server instance and one or more alternate endpoints on other application server instances in the cluster;
• the client issuing a request to the object via the cluster-aware remote reference;
• detecting a communication failure for the request, wherein the communication failure indicates failure of the application server instance;
• selecting an alternate endpoint on one of the other application server instances from the one or more alternate endpoints indicated by the cluster-aware remote reference;
• and
• failing over the cluster-aware remote reference to the alternate endpoint on the other application server instance.

A computer-readable storage medium, storing program instructions, wherein the program instructions are computer-executable to implement:

• providing a cluster-aware remote reference to an object on an application server instance in a cluster of two or more application server instances to an IIOP or RMI-IIOP client of the application server instance, wherein the cluster-aware remote reference indicates a primary endpoint on the application server instance and one or more alternate endpoints on other application server instances in the cluster;
• the client issuing a request to the object via the cluster-aware remote reference;
• detecting a communication failure for the request, wherein the communication failure indicates failure of the application server instance;
• selecting an alternate endpoint on one of the other application server instances from the one or more alternate endpoints indicated by the cluster-aware remote reference;
• and
• failing over the cluster-aware remote reference to the alternate endpoint on the other application server instance.