Method and apparatus for partitioning a computer cluster through coordination point devices

• US 8,024,432 B1
• Filed: 06/27/2008
• Issued: 09/20/2011
• Est. Priority Date: 06/27/2008
• Status: Active Grant

• Current Assignee: Veritas Technologies, LLC (Whitehouse Group Ltd.)
• Original Assignee: Symantec Corporation (NortonLifeLock Inc.)
• Inventors: Sharma, Sumit, Katkar, Amol Shivram
• Primary Examiner(s): Tran; Philip B
• Application Number:
  US12/147,764
• Time in Patent Office: 1,180 Days
• Field of Search: 709223-224, 709/220, 709/203, 709/209, 714/4, 714/4.1, 714/4.11
• US Class Current: 709/220
• CPC Class Codes: G06F 9/5061 Partitioning or combining of resources

Abstract

A method and apparatus for partitioning a computer cluster through coordination point devices that provide enhanced split brain arbitration during I/O fencing. In one embodiment, a method for using coordination point devices to support Input/Output fencing in a computer cluster comprises processing at least one arbitration request from at least one client computer upon an occurrence of a split brain within a computer cluster, wherein each arbitration request indicates at least one established connection between at least one coordination point device and each client computer of the at least one client computer and each client sub-cluster membership, wherein the at least one coordination point device form a coordination point cluster membership and selecting a client computer of the at least one client computer to continue to be a portion of the computer cluster based on the arbitration request.

Claims

1. A method comprising:

receiving an arbitration request from a client computer upon an occurrence of a cluster partition, wherein the cluster partition partitions a computer cluster into a plurality of sub-clusters, wherein a first sub-cluster of the plurality of sub-clusters comprises the client computer wherein the arbitration request indicates which of a plurality of coordination point devices the client computer is connected to and which of a plurality of client computers are included in the first sub-cluster;

comparing a coordination point cluster membership of the first sub-cluster of the plurality of sub-clusters and a second coordination point cluster membership of a second sub-cluster of the plurality of sub-clusters;

selecting, based on the arbitration request, to vote for the client computer to continue to be a member of the computer cluster;

identifying a failed coordination point device of the plurality of coordination point devices; and

communicating, to the client computer, the vote and a proxy vote for the failed coordination point device.

2. The method of claim 1, wherein the vote is associated with an Input/Output fencing decision.

3. The method of claim 1, wherein the vote identifies a client sub-cluster that comprises the selected client computer to survive an Input/Output fencing decision.

4. The method of claim 3, wherein the client sub-cluster receives a number of votes that exceeds one-half of a number of configured coordination point devices in the computer cluster.

5. The method of claim 1, further comprising:

determining at least one coordination point server that is not communicable with the client computer.

6. The method of claim 1, wherein the vote further comprises the proxy vote for at least one failed coordination point device.

7. The method of claim 1, wherein the selecting is performed in accordance with a policy, wherein the policy is used to determine a best client computer to provide the vote.

8. The method of claim 1, further comprising communicating a plurality of votes to the client computer from an odd number of coordination point devices, wherein the client computer is to remain a member of the computer cluster if the plurality of votes exceeds at least one-half of a number of coordination point devices.

9. The method of claim 1, further comprising:

selecting a second client computer that is to be fenced from various cluster resources during the cluster partition.

10. The method of claim 1, further comprising separating a second client computer from the computer cluster, wherein the second client computer is not connected with the client computer.

11. An apparatus comprising:

an arbitration module for

receiving an arbitration request from a client computer upon an occurrence of a cluster partition, wherein the cluster partition partitions a computer cluster into a plurality of sub-clusters, wherein a first sub-cluster of the plurality of sub-clusters comprises the client computer, wherein the arbitration request comprises status information indicating which of a plurality of coordination point devices the client computer is connected to and which of a plurality of client computers are included in the first sub-cluster,

comparing a coordination point cluster membership of the first sub-cluster of the plurality of sub-clusters and a second coordination point cluster membership of a second sub-cluster of the plurality of sub-clusters,

selecting, based on the arbitration request, to vote for the client computer to continue to be a member of the computer cluster, and

identifying a failed coordination point device of the plurality of failed coordination point devices;

a memory comprising the status information; and

a communicating module for communicating to the client computer a vote and a proxy vote for the failed coordination point device.

12. The apparatus of claim 11, wherein the arbitration module selects the client computer in accordance with a policy, wherein the policy is used to determine a best client computer to provide a vote to remain a member of the computer cluster.

13. A system comprising:

a client computer;

a memory comprising status information which indicates which of a plurality of coordination point devices the client computer is connected to and which of a plurality of client computers are included in a first sub-cluster of a plurality of sub-clusters, wherein the first sub-cluster of the plurality of sub-clusters comprises the client computer;

a fencing driver for communicating an arbitration request and processing a response for the arbitration request to determine an Input/Output fencing decision;

an odd number of coordination point devices;

an arbitration module for

receiving the arbitration request from the client computer upon an occurrence of a cluster partition, wherein the cluster partition partitions a computer cluster into the plurality of sub-clusters,

comparing a coordination point cluster membership of the first sub-cluster of the plurality of sub-clusters and a second coordination point cluster membership of a second sub-cluster of the plurality of sub-clusters,

selecting, based on the arbitration request, to vote for the client computer to continue to be a member of the computer cluster, and

identifying a failed coordination point device of the plurality of failed coordination point devices; and

a communicating module for communicating to the client computer a vote and a proxy vote for a failed coordination point device.

14. The system of claim 13, wherein the odd number of coordination point devices communicate a plurality of votes to the client computer, wherein the client computer is to remain a member of the computer cluster if the plurality of votes exceeds at least one-half of the odd number of coordination point devices.

15. The system of claim 13, wherein a policy is used to determine a best client computer to provide the vote.