Method of improving the availability of a computer clustering system through the use of a network medium link state function

A method for improving the availability and reliability of a computer clustering system including a first server and a second server interconnected by a communication link, each of said first server and said second server including at least one associated mass storage device such that each of said first server and second server can receive requests that result in data being written to or read from the associated at least one mass storage device, said second server being assigned right to survive in case of disruption in said computer clustering system, said method comprising the acts of:

• said first server determining that a heartbeat signal associated with said second server is no longer being detected on said communication link;
• said first server determining that first server and said second server cannot mirror at least changes to said at least one associated mass storage device of said first server and said at least one associated mass storage device of said second server;
• said first server analyzing the communication link to determine if any error exists in the integrity of the communication link;
• if said first server determines there is an error in the integrity of the communication link, said first server discontinuing operation and said second server reforming the computer clustering system so that requests that would otherwise be directed to said first server are serviced by said second server;
• and
• if said first server determines that there is no error in the integrity of the communication link, said first server assigning itself the right to survive and said first server reconfiguring the computer clustering system without communicating with said at least one associated mass storage devices of said first server and said second server so that said first server services requests that would otherwise be directed to said second server.

A method for improving the availability and reliability of a computer clustering system that includes a first server and a second server interconnected by a communication link, each of said first server and said second server including at least one associated mass storage device such that each of said first server and second server can receive requests that result in data being written to or read from the associated at least one mass storage device, said second server being assigned right to survive in case of disruption in said computer clustering system, said method comprising the acts of:

• during normal operation of the computer clustering system, mirroring data on the at least one mass storage device associated with the first server and on the at least one mass storage device associated with the second server by transmitting the data between the first server and the second server using the communication link;
• determining that said second server is not functioning properly such that said first server and said second server cannot mirror the data that is written to or read from the least one mass storage device of the first server and the at least one mass storage device of the second server, including the acts of:
• analyzing the communication link at specified time intervals;
• failing to detect communication from said second server on said communication link;
• and
• detecting proper functionality of said communication link based on the act of analyzing said communication link;
• taking control of the computer clustering system by assuming the right to survive without communication with another server or with said at least one associated mass storage devices of said first server and said second server;
• and
• reconfiguring the computer clustering system so said first server receives file server requests that would otherwise be directed to said second server.

A computer program product for implementing, in a first server included in a computer clustering system that also includes a second server and a communication link connecting the first server and the second server, a method for said first server to assume control of the computer clustering system in response to a failure of said second server, the computer program product comprising:

• a computer-readable medium carrying computer-executable instructions for implementing the method, said computer-executable instructions including:
• program code means for determining whether said first server has a right to survive if said first server fails to detect communication from said second server, said first server initially not having the right to survive;
• program code means for detecting loss of communication from said second server that prevents said first server from mirroring data with said second server;
• program code means for determining, in response to the loss of communication, whether the communication link is functioning properly;
• program code means for executing the acts of:
• if the communication link is not functioning properly, discontinuing operation of said first server;
• and
• if the communication link is functioning properly, continuing operation of said first server and assuming the right to survive without communicating with another server or device, notwithstanding said first server initially not having the right to survive.