Method and apparatus for improved cluster administration

• US 6,311,217 B1
• Filed: 06/04/1998
• Issued: 10/30/2001
• Est. Priority Date: 06/04/1998
• Status: Expired due to Fees

• Current Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Original Assignee: Compaq Computer Corporation (HP Inc.)
• Inventors: Ehlinger, Early David, Fletcher, Mark F.
• Primary Examiner(s): Dinh, Dung C.
• Assistant Examiner(s): TRIM, NKOSI N
• Application Number:
  US09/090,603
• Time in Patent Office: 1,244 Days
• Field of Search: 709/223, 709/224, 709/225, 709/226, 710/200, 710/220, 710/240-244
• US Class Current: 709/226
• CPC Class Codes: H04L 41/0893 Assignment of logical groups to network elements
• CPC Class Codes: H04L 67/01 Protocols
• CPC Class Codes: H04L 67/1034 Reaction to server failures by a load balancer
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]

Abstract

A cluster administration system that is capable of handling a cluster having one or more computing devices. The number of computing devices that may be included in a cluster is limited only by practical considerations rather than software or hardware limitations. A cluster administration system may include a cluster of computing devices, one of the computing devices being an owner. The cluster further includes a resource. Direct access to the resource by the computing devices is controlled by the owner of the cluster. The cluster administration system also includes an arbiter. The arbiter and the cluster are in communication with each other and a network, the cluster providing the network with access to the storage device. The arbiter controls the admission of new computing devices to the cluster when the owner of the cluster is incapable of admitting the new computing device. Having the arbiter outside the cluster provides greater reliability. The arbiter is not affected by failures within the cluster. One or more of the computing devices of the cluster may fail, but the administration of the cluster is not affected. The functions of the arbiter may also be distributed among several independent computing devices which can hand off the primary duties of the arbiter should one or more of the independent computing devices fail to satisfactorily perform the duties of arbitration.

Claims

1. A cluster of computing devices comprising:

a resource; and

a plurality of computing devices in communication with each other, wherein each of the plurality of computing devices are directly coupled to the resource, a one of the plurality of computing devices being an owner of the resource, the owner controlling direct access to the resource by the plurality of computing devices, the cluster of computing devices including one or more computing devices, and the cluster of computing devices providing a network with indirect access to the resource;

wherein an independent computing device, independent of the cluster of computing devices, is in communication with the cluster of computing devices and configured to admit another computing device into the cluster of computing devices if the other computing device is capable of communicating with the independent computing device.

2. The cluster of computing devices of claim 1, wherein the independent computing device determines which of the plurality of computing devices is the owner.

3. The cluster of computing devices of claim 2, wherein the owner is also configured to admit the other computing device if the other computing device is capable of communicating with the owner through the network.

4. The cluster of computing devices of claim 3, wherein the independent computing device is configured to admit the other computing device if the other computing device fails to obtain admission through the owner, and the other computing device is capable of communicating with the independent computing device.

5. A cluster administration system comprising:

a cluster of computing devices including,

a resource, and

a plurality of computing devices in communication with each other, wherein each of the plurality of computing devices are directly coupled to the resource, a one of the plurality of computing devices being an owner of the resource, the owner controlling the direct access by the plurality of computing devices to the resource, the cluster of computing devices providing a network with indirect access to the resource; and

an arbiter, the arbiter being independent of the cluster of computing devices, configured to admit another computing device to the cluster of computing devices if the other computing device is in communication with the arbiter.

6. The cluster administration system of claim 5, wherein the arbiter determines which of the plurality of computing devices is the owner.

7. The cluster administration system of claim 5, wherein the owner is also configured to admit the other computing device if the other computing device is capable of communicating with the owner through the network.

8. The cluster administration system of claim 6, wherein the arbiter is configured to admit the other computing device if the other computing device fails to obtain admission through the owner, the other computing device capable of communicating with the arbiter.

9. The cluster administration system of claim 5 further comprising an independent computing device, wherein the arbiter is a process implemented on the independent computing device in communication with the cluster of computing devices, the independent computing device being independent of the cluster of computing devices.

10. The cluster administration system of claim 5 further comprising a plurality of independent computing devices, wherein the arbiter is a distributed process implemented on the plurality of independent computing devices in communication with the cluster of computing devices, the plurality of independent computing devices being independent of the cluster of computing devices.

11. The cluster administration system of claim 10, wherein the arbiter is primarily implemented on a first independent computing device of the plurality of independent computing devices, the first independent computing device configured to admit another computing device to the cluster of computing devices if the other computing device is in communication with the arbiter.

12. The cluster administration system of claim 11, wherein the owner is also configured to admit the other computing device if the other computing device is capable of communicating with the owner through the network.

13. The cluster administration system of claim 12, wherein the first independent computing device is configured to admit the other computing device if the other computing device fails to obtain admission through the owner, the other computing device capable of communicating with the first independent computing device.

14. The cluster administration system of claim 13, wherein if the first independent computing device loses communication with the cluster of computing devices, the arbiter is primarily implemented on a second independent computing device of the plurality of independent computing devices, the second independent computing device configured to administer the admission of the other computing device.

15. The cluster administration system of claim 10, wherein the plurality of independent computing devices is an independent cluster of computing devices.

16. The cluster administration system of claim 5, wherein a first computing device of the cluster of computing devices performs a function, and if the first computing device is not capable of significantly performing the function the arbiter assigns a second computing device of the computing devices to perform the function.

17. The cluster administration system of claim 16, wherein the first computing device notifies the arbiter that the first computing device cannot significantly perform the function such that the arbiter assigns the second computing device to perform the function.

18. A method of administering a cluster of computing devices, the cluster including a plurality of computing devices and a resource, the plurality of computing devices having direct access to the resource, wherein one of the plurality of computing devices is an owner of the resource, the owner controlling direct access to the resource by the other computing devices of the plurality of computing devices, the cluster of computing devices providing a network with access to the resource, the method comprising:

another computer device requesting admission into the cluster of computing devices from an arbiter that is not included in the cluster of computing devices.

19. The method of claim 18 further comprising:

the other computing device initially requesting admission into the cluster of computing devices from the owner through the network; and

admitting the other computing device to the cluster of computing devices if the other computing device successfully communicates with the owner;

such that the owner does not need to contend for ownership over the resource with the other computing devices.

20. The method of claim 18 further comprising:

admitting the other computing device into the cluster of computing devices if the other computing device successfully requests admission from the arbiter after the other computing device fails to successfully communicate with the owner.

21. The method of claim 18 further comprising:

determining which one of the plurality of computing devices is the owner.

22. The method of claim 18 further comprising:

determining if the owner is active; and

assigning ownership over the resource to a next computing device of the plurality of computing devices if the owner is not active, the next computing device being a new owner.

23. The method of claim 18, wherein the arbiter is implemented on a computing device independent of the cluster of computing devices.

24. The method of claim 18, wherein the arbiter is implemented on a second cluster of computing devices, a first computing device of the second cluster of computing devices acting as the arbiter, the method further comprising:

transferring the duties of the arbiter to a second computing device of the second cluster of computing devices if the first computing device fails or losses communication with the cluster of computing devices.

25. A computer program product for administering a cluster of computing devices, the cluster of computing devices including a plurality of computing devices and a resource, the plurality of computing devices having direct access to the resource, wherein one of the plurality of computing devices is an owner of the resource, the owner controlling direct access to the resource by the other computing devices of the plurality of computing devices, the cluster of computing devices providing a network with access to the resource, the computer program product comprising:

a first computer code that enables a first independent computing device that is not included in the cluster of computing devices to receive requests through the network from another computing device to be admitted to the cluster of computing; and

a computer readable medium that stores the first computer code.

26. The computer program product of claim 25 further comprising:

a second computer code that enables the first independent computing device to admit the other computing device into the cluster of computing devices if the other computing device successfully requests admission from the first independent computing device after the other computing device failed to successfully communicate with the owner for admission;

the computer readable medium further storing the second computer code.

27. The computer program product of claim 25 further comprising:

a third computer code that enables the first independent computing device to determine which one of the plurality of computing devices is the owner;

the computer readable medium further storing the third computer code.

28. The computer program product of claim 26 further comprising:

a fourth computer code that enables the second independent computing device to receive requests for admission and admit the other computing device when the first independent computing device fails or losses communication with the cluster of computing devices; and

a fifth computer code that enables the second independent computing device to determine which one of the plurality of computing devices is the owner;

the computer readable medium further storing the fourth and fifth computer codes.

29. A computer program product for administering a cluster of computing devices, the cluster of computing devices including a plurality of computing devices and a resource, the plurality of computing devices having direct access to the resource, wherein one of the plurality of computing devices is an owner of the resource, the owner controlling the direct access to the resource by the other computing devices of the plurality of computing devices, the cluster of computing devices providing a network with access to the resource, the computer program product comprising:

a first computer code that enables the owner to manage direct access to the resource by the plurality of computing devices when the plurality of computing devices includes more than two computing devices, including the owner; and

a computer readable medium that stores the first computer code.

30. The computer program product of claim 29 further comprising:

a second computer code that enables the owner to receive requests from another computing device to be admitted to the cluster of computing devices through the network, wherein the other computing device is communicatively coupled to the cluster of computing devices and the network;

the computer readable medium further storing the second computer code.

31. The computer program product of claim 30 further comprising:

a third computer code that enables the owner to admit the other computer to the cluster of computing devices through the network;

the computer readable medium further storing the third computer code.

32. The computer program product of claim 29 further comprising:

a fourth computer code that enables the owner to communicate with a first independent computing device, wherein the first independent computing device is communicatively coupled to the cluster of computing devices and the network;

the computer readable medium further storing the fourth computer code.