Updating a list of quorum disks

• US 8,443,231 B2
• Filed: 04/12/2010
• Issued: 05/14/2013
• Est. Priority Date: 04/12/2010
• Status: Active Grant

• Current Assignee: Veritas Technologies, LLC (Whitehouse Group Ltd.)
• Original Assignee: Symantec Corporation (NortonLifeLock Inc.)
• Inventors: Abraham, Sara, Harmer, Craig, Dash, Prasanta, Kher, Vishal
• Primary Examiner(s): Baderman, Scott
• Assistant Examiner(s): BRYAN, JASON B
• Application Number:
  US12/758,201
• Publication Number:
  US 20110252270A1
• Time in Patent Office: 1,128 Days
• Field of Search: None
• US Class Current: 714/12
• CPC Class Codes: G06F 11/1425 by reconfiguration of node membership
• CPC Class Codes: G06F 11/2094 Redundant storage or storage space G06F11/2056 takes precedence

Abstract

A node in a server cluster is designated as a quorum disk. The node stores a list of other nodes in the server cluster also designated as quorum disks. The node can replace the first list with a second and more recent list of quorum disks only if the second list is updated on at least a simple majority of quorum disks on the first list.

Claims

1. A non-transitory computer-readable storage medium having computer-executable instructions for causing a computer system to perform a method comprising:

identifying a first number of nodes that constitute a server cluster;

designating a subset comprising not more than a second number of said nodes as quorum disks for said server cluster, said second number less than said first number, wherein said quorum disks are accessed in response to a split-brain event that partitions said server cluster into different sub-clusters of nodes;

storing, on each of said quorum disks, a first list identifying said nodes that constitute said server cluster;

storing a second list identifying said subset of nodes designated as quorum disks on at least a simple majority of nodes on said second list; and

replacing said second list with a third and more recent list of quorum disks only if said third list is updated on at least a simple majority of quorum disks on said second list.

2. The non-transitory computer-readable storage medium of claim 1 wherein said identifying further comprises choosing quorum disks from different failure domains that are defined for said server cluster.

3. The non-transitory computer-readable storage medium of claim 1 wherein said method further comprises:

determining that a node on said second list is unavailable; and

updating said second list to include another node in said server cluster in place of said node that is unavailable.

4. The non-transitory computer-readable storage medium of claim 1 wherein said method further comprises automatically changing the number of quorum disks in response to a change in the number of nodes in said server cluster.

5. The non-transitory computer-readable storage medium of claim 1 wherein said method further comprises

monitoring activity of said nodes in said server cluster to identify candidates for quorum disks;

monitoring activity of said quorum disks on said second list;

automatically identifying a quorum disk on said second list that is unavailable; and

automatically replacing the unavailable quorum disk with a node selected from said candidates.

6. The non-transitory computer-readable storage medium of claim 1 wherein said storing further comprises associating version information with said second list.

7. The non-transitory computer-readable storage medium of claim 1 wherein said method further comprises storing on each of said quorum disks information comprising an activity count associated with said nodes in said server cluster.

8. A computer-implemented method comprising:

accessing a server cluster comprising a first number of nodes;

identifying a first subset of said nodes during a first phase of a search, said search to identify nodes designated as quorum disks for said server cluster;

storing, on each of said quorum disks, a first list identifying said nodes that constitute said server cluster;

after said first phase, propagating said search through succeeding subsets of said nodes, wherein each succeeding subset of said nodes is selected based upon lists of nodes stored on nodes identified in a preceding phase of said search;

identifying a latest list of said nodes, wherein said latest list is selected from lists of nodes identified during said search and comprises a second subset comprising not more than a second number of said nodes in said server cluster, said second number less than said first number, wherein said second subset of nodes comprises said quorum disks; and

storing said latest list identifying said second subset of nodes designated as quorum disks on at least a simple majority of nodes on said latest list, wherein further nodes in said latest list are accessed in response to a split-brain event that partitions said server cluster into different sub-clusters of nodes.

9. The computer-implemented method of claim 8 further comprising replacing a second list of quorum disks with a third and more recent list of quorum disks only if said third list is updated on at least a simple majority of quorum disks on said second list.

10. The computer-implemented method of claim 8 further comprising choosing quorum disks from different failure domains defined for said server cluster.

11. The computer-implemented method of claim 8 further comprising:

monitoring activity of said nodes in said server cluster to identify candidates for quorum disks;

monitoring activity of said nodes designated as said quorum disks;

automatically determining that a node designated as a quorum disk is unavailable; and

automatically designating another node selected from said candidates as a quorum disk in place of said node that is unavailable.

12. The computer-implemented method of claim 8 further comprising selecting a number of said nodes in said server cluster as quorum disks based on a factor selected from the group of:

the number of said nodes in said server cluster;

a desired availability level;

a desired redundancy level.

13. The computer-implemented method of claim 8 further comprising associating version information with each of said lists including said latest list.

14. A data server in a server cluster, said data server comprising:

a processor; and

memory coupled to said processor and having stored therein instructions that, if executed by said data server, cause said data server to execute a method comprising;

selecting a subset of a first number of data servers that are members of said server cluster and designating data servers in said subset as quorum disks for said server cluster, wherein said subset includes said data server and comprises not more than a second number of said data servers in said server cluster, said second number less than said first number;

storing, on each of said data servers designated as quorum disks, a first list identifying said data servers that constitute said server cluster;

storing a second list identifying said subset of data servers designated as quorum disks on at least a simple majority of data servers on said second list; and

replacing said second list with a third and more recent list of quorum disks only if said second list is received by at least a simple majority of quorum disks on said second list, wherein said data server is accessed in response to a split-brain event that partitions said server cluster into different sub-clusters of data servers.

15. The data server of claim 14 wherein said quorum disks are chosen from different failure domains that are defined for said server cluster.

16. The data server of claim 14 wherein said method further comprises updating said second list to include another data server in said server cluster in place of a data server on said second list that is unavailable.

17. The data server of claim 14 wherein a number of said data servers designated as said quorum disks is a function of a factor selected from the group consisting of:

the number of said data servers in said server cluster;

a desired redundancy level;

a desired availability level.

18. The data server of claim 14 wherein said data servers in said server cluster are monitored to identify active data servers that are candidates for said quorum disks, wherein said quorum disks on said second list are also monitored to identify a quorum disk on said second list that is unavailable, wherein further an unavailable quorum disk on said second list is automatically replaced with a data server selected from said candidates to generate said third list.

19. The data server of claim 14 wherein first version information is associated with said second list and second version information is associated with said third list.