Distributed hot-spare storage in a storage cluster

• US 8,479,037 B1
• Filed: 08/29/2011
• Issued: 07/02/2013
• Est. Priority Date: 10/20/2005
• Status: Active Grant

• Current Assignee: AmZetta Technologies, LLC (American Megatrends International, LLC)
• Original Assignee: American Megatrends Incorporated
• Inventors: Chatterjee, Paresh, Narayanan, Ajit, Mahalingam, Anandh, Subramanian, Srikumar
• Primary Examiner(s): Wilson, Yolanda L
• Application Number:
  US13/220,305
• Time in Patent Office: 673 Days
• Field of Search: None
• US Class Current: 714/4.11
• CPC Class Codes: G06F 11/2061 combined with de-clustering of data
• CPC Class Codes: G06F 11/2082 Data synchronisation
• CPC Class Codes: G06F 11/2094 Redundant storage or storage space G06F11/2056 takes precedence
• CPC Class Codes: G06F 2211/1028 Distributed, i.e. distributed RAID systems with parity

Abstract

A method, system, apparatus, and computer-readable medium are described for providing distributed hot-spare storage in a redundant storage cluster. According to one method, a portion of the unutilized space on the storage cluster is utilized as a distributed hot-spare storage node. Through this mechanism, a redundant storage cluster with N storage nodes may be contracted to a redundant array with N−1 nodes. Thin provisioning and intelligent data placement may also be utilized to implement the distributed hot-spare storage node. Through repeated application of such methods and systems, the failure of any storage node or the sequential failure of multiple storage nodes within a redundant storage cluster results in the recreation of the cluster as a redundant storage array with one fewer node, but with the same redundancy.

Claims

1. A method for providing distributed hot-spare storage in a storage cluster, the method comprising:

allocating unutilized space on all nodes in the storage cluster, the storage cluster having N storage nodes, wherein N is greater than or equal to three;

detecting a failure of a storage node in the storage cluster; and

in response to detecting the failure, reforming the storage cluster as a storage cluster having N−

1 storage nodes using at least a portion of the unutilized space on the remaining storage nodes,

wherein reforming the storage cluster comprises copying data that was mirrored on the failed storage node from the N−

1 storage nodes to the at least a portion of the unutilized space on the remaining storage nodes, and

wherein the distributed hot-spare storage node comprises the at least a portion of the unutilized storage space on the remaining storage nodes of the storage cluster.

2. The method of claim 1, wherein the data is stored on the storage cluster such that when the data that was mirrored on the failed storage node is copied from the N−

1 storage nodes to the distributed hot-spare storage node an equal amount of data is copied from each of the N−

1 storage nodes to the distributed hot-spare storage node.

3. The method of claim 2, wherein capacity on the storage cluster is allocated utilizing thin provisioning.

4. The method of claim 3, wherein copying the data that was mirrored on the failed storage node from the N−

1 storage nodes to the distributed hot-spare storage node comprises;

creating an exception table for the N−

1 storage nodes of the storage cluster, the exception table identifying the areas of the N−

1 storage nodes that need to be copied to the distributed hot-spare storage node; and

executing a background process operative to copy the data from the N−

1 storage nodes to the distributed hot-spare storage node according to the contents of the exception table.

5. The method of claim 1, wherein the storage cluster further comprises sufficient free storage space to store all of the data stored on the cluster, and wherein the method further comprises:

repeatedly performing the reforming operation without user intervention in response to detecting the failure of a storage node in the storage cluster until the storage cluster has contracted from N nodes to two nodes.

6. The method of claim 1, wherein the reforming is performed without any quality of service loss at the storage cluster.

7. A non-transitory computer-readable medium having computer-executable instructions stored thereon which, when executed by a computer, cause the computer to:

allocate unutilized space on all nodes in the storage cluster, the storage cluster having N storage nodes, wherein N is greater than or equal to three;

detect a failure of a storage node in the storage cluster; and

in response to a detection of the failure, reform the storage cluster as a storage cluster having N−

1 storage nodes using at least a portion of the unutilized space on the remaining storage nodes,

wherein reforming the storage cluster comprises copying data that was mirrored on the failed storage node from the N−

1 storage nodes to the at least a portion of the unutilized space on the remaining storage nodes, and

wherein the distributed hot-spare storage node comprises the at least a portion of the unutilized storage space on the remaining storage nodes of the storage cluster.

8. The non-transitory computer-readable medium of claim 7, wherein the data is stored on the storage cluster such that when the data that was mirrored on the failed storage node is copied from the N−

1 storage nodes to the distributed hot-spare storage node an equal amount of data is copied from each of the N−

1 storage nodes to the distributed hot-spare storage node.

9. The non-transitory computer-readable medium of claim 8, wherein capacity on the storage cluster is allocated utilizing thin provisioning.

10. The non-transitory computer-readable medium of claim 9, wherein copying the data that was mirrored on the failed storage node from the N−

1 storage nodes to the distributed hot-spare storage node comprises;

creating an exception table for the N−

1 storage nodes of the storage cluster, the exception table identifying the areas of the N−

1 storage nodes that need to be copied to the distributed hot-spare storage node; and

executing a background process operative to migrate data from the N−

1 storage nodes to the distributed hot-spare storage node according to the contents of the exception table.

11. The non-transitory computer-readable medium of claim 7, wherein the operation to reform the storage cluster is performed without any quality of service loss at the storage cluster.

12. A system for providing a distributed hot-spare storage node in a storage cluster, the system comprising:

two or more storage nodes that together comprise a redundant storage cluster, the two or more storage nodes each having allocated therein unutilized space; and

a distributed hot-spare storage node comprising at least a portion of the unutilized space from the two or more storage nodes,

wherein the storage cluster is operative to detect a failure of one of the storage nodes and to reform itself following the failure as a redundant storage cluster having one less storage node utilizing the distributed hot-spare storage node, wherein data stored on the storage cluster is mirrored on two of the storage nodes, and wherein reforming the storage cluster comprises copying data that was mirrored on the failed storage node to the distributed hot-spare storage node.

13. The system of claim 12, wherein the data is stored on the storage cluster such that when the data that was mirrored on the failed storage node is copied from the storage nodes to the distributed hot-spare storage node an equal amount of data is copied from each of the storage nodes remaining in the cluster following the failure to the distributed hot-spare storage node.

14. The system of claim 13, wherein the storage cluster is operative to allocate requests to store data to the storage nodes utilizing thin provisioning.

15. The system of claim 14, wherein copying the data that was mirrored on the failed storage node from the storage nodes to the distributed hot-spare storage node comprises:

creating an exception table for the storage nodes of the storage cluster, the exception table identifying the areas of the storage nodes that need to be copied to the distributed hot-spare storage node; and

executing a background process operative to migrate data from the storage nodes to the distributed hot-spare storage node according to the contents of the exception table.

16. The system of claim 12, wherein the operation to reform the storage cluster is performed without any quality of service loss at the storage cluster.