Disaster recovery at high reliability in a storage cluster

• US 10,152,397 B2
• Filed: 05/14/2015
• Issued: 12/11/2018
• Est. Priority Date: 06/04/2014
• Status: Active Grant

• Current Assignee: Pure Storage, Inc.
• Original Assignee: Pure Storage, Inc.
• Inventors: Hayes, John, Botes, Par
• Primary Examiner(s): Chu, Gabriel
• Application Number:
  US14/712,761
• Publication Number:
  US 20150355984A1
• Time in Patent Office: 1,307 Days
• CPC Class Codes: G06F 11/1458 Management of the backup or restore process
• CPC Class Codes: G06F 11/2094 Redundant storage or storage space G06F11/2056 takes precedence
• CPC Class Codes: G06F 21/602 Providing cryptographic facilities or services
• CPC Class Codes: G06F 21/6272 by registering files or documents with a third party
• CPC Class Codes: G06F 2201/84 Using snapshots, i.e. a logical point-in-time copy of the data
• CPC Class Codes: G06F 3/0619 in relation to data integrity, e.g. data losses, bit errors
• CPC Class Codes: G06F 3/065 Replication mechanisms
• CPC Class Codes: H04L 67/1095 Replication or mirroring of data, e.g. scheduling or transport for data synchronisation between network nodes
• CPC Class Codes: H04L 9/085 Secret sharing or secret splitting, e.g. threshold schemes

Abstract

A storage grid is provided. The storage grid includes a first cluster, a second cluster, and a third cluster. Each of the first cluster, the second cluster and the third cluster is configured to store an amount of data ranging from a portion of a copy of the data to a full copy of the data. The first cluster has a full copy of data written to the first cluster and at least a partial copy of data written to the second and third cluster. The second cluster has a full copy of data written to the second cluster, and at least a partial copy of the data written to the first and third cluster. The third cluster has a full copy of data written to the third cluster and at least a partial copy of the data written to the first and second cluster. A method of storing data is also provided.

Claims

1. A method of storing data in a storage grid, comprising:

receiving from a client system, by a storage controller, a first data at the storage grid with a request from the client system to store the first data at the storage grid, the storage grid comprising multiple storage clusters;

storing a full copy of the first data at a first storage cluster of the storage grid;

receiving and storing, by the storage controller, a second data at the storage grid, such that no one storage cluster of the storage grid has a full copy of the first data and a full copy of the second data, and any user data received and stored at the storage grid is recoverable from each pair of storage clusters in the storage grid; and

recovering, by the storage controller, a portion of the first data from a second storage cluster, wherein the portion of the first data is not a full copy of the first data.

2. The method of claim 1, further comprising:

changing data destinations among the storage clusters of the storage grid, wherein each combination of two storage clusters of the storage grid has at least the full copy of the first data and the full copy of the second data, and wherein the multiple storage clusters are disposed within at least two different chassis.

3. The method of claim 1, further comprising:

storing the first data at the first storage cluster;

storing the portion of the first data at the second storage cluster of the storage grid;

storing a remaining portion of the first data at a third storage cluster of the storage grid.

4. The method of claim 3, further comprising:

splitting the first data into the portion and the remaining portion, wherein the splitting is performed by the storage controller, wherein the data includes a first segment of data and a second segment of data, the portion of the data includes the first segment of data, and the remaining portion of the data includes the second segment of data.

5. The method of claim 3, wherein storing a portion of the data at the second storage cluster and storing a remaining portion of the data at the third storage cluster includes writing alternating segments of data to the second storage cluster and the third storage cluster.

6. The method of claim 3, further comprising:

interleaving the portion of the data and the remaining portion of the data so as to recover a copy of the data, responsive to a failure of the first storage cluster of the storage grid.

7. The method of claim 1, further comprising:

encrypting contents of each of the multiple storage clusters via a shared secret; and

storing a first share of the shared secret at a first storage cluster, a second share of the shared secret at a second storage cluster, and a third share of the shared secret at a third storage cluster, such that any two of the first share, the second share, and the third share suffice for decrypting the contents of any of the first storage cluster, the second storage cluster, and the third storage cluster.

8. A storage grid, comprising:

a plurality of storage clusters, each storage cluster of the storage grid configurable, by a storage controller, to store an amount of each of a first data and a second data ranging from a portion of a copy of the first data to a full copy of the first data and from a portion of a copy of the second data to a full copy of the second data, in response to receiving the first data and the second data from one or more client systems, such that;

no one storage cluster of the storage grid has a full copy of the first data and a full copy of the second data, and any user data received and stored at the storage grid is recoverable from each pair of storage clusters in the storage grid.

9. The storage grid of claim 8, wherein:

each two storage clusters has at least the full copy of each of the first data, the full copy of the second data or the full copy of the third data, and wherein at least two of the each storage cluster is disposed within different chassis.

10. The storage grid of claim 8, further configurable so that in an event of a failure of a first storage cluster, the first data can be recovered by interleaving a partial copy of the first data, from a second storage cluster and a difference between the full copy of the first data and the partial copy of the first data, from a third storage cluster, the interleaving at a granularity of a segment.

11. The storage grid of claim 8, further comprising:

each storage cluster including flash memory.

12. The storage grid of claim 8, further comprising:

a first storage cluster, a second storage cluster, and a third storage cluster configurable to couple to each other via a network and wherein each of the first storage cluster, the second storage cluster, and the third storage cluster have different storage capacities.

13. The storage grid of claim 8, further comprising:

a secret generation unit, configurable to generate a shared secret applicable to encrypting keys;

a share generation unit, configurable to generate shares of the shared secret; and

each storage cluster configurable to store one of the shares of the shared secret.

14. The storage grid of claim 8, further comprising:

each storage cluster includes a plurality of storage units with solid-state memory.

15. A storage grid comprising:

multiple storage clusters, configurable to receive a first data and a second data;

each storage cluster configurable, by a storage controller, to store an amount of the first data and the second data ranging from a portion of a copy of the first data to a full copy of the first data and from a portion of a copy of the second data to a full copy of the second data in one of the multiple storage clusters, responsive to receiving the first data, the second data and one or more requests from one or more client systems to store the first data and the second data at the storage grid, such that;

a full copy of the first data is stored on only one storage cluster of the storage grid;

a full copy of the second data is stored on only one storage cluster of the storage grid; and

any user data received and stored at the storage grid is recoverable from each pair of storage clusters in the storage grid.

16. The storage grid of claim 15, wherein:

the multiple storage clusters comprise a first storage cluster, a second storage cluster and a third storage cluster, and wherein the multiple storage clusters are disposed within at least two different chassis; and

any combination of two clusters of the first storage cluster, the second storage cluster, and the third storage cluster can recover the first data, the second data and a third data.

17. The storage grid of claim 15, wherein a first storage cluster is configurable to send the portion of the copy of the first data and a remaining portion of the copy of the first data as alternating segments of data to a third storage cluster and the second storage cluster.

18. The storage grid of claim 15, further comprising:

at least one processor configurable to generate a shared secret and shares of the shared secret;

each of the multiple storage clusters configurable to store at least one of the shares of the shared secret, wherein shares of the shared secret from two of the multiple storage clusters are applicable to decrypt data on any one of multiple storage clusters.

19. The storage grid of claim 15, wherein the first data and the second data are portioned by one of the multiple storage clusters of the storage grid or a network coupled to each of the multiple storage clusters of the storage grid.

20. The storage grid of claim 15, further comprising:

a first storage cluster and a second storage cluster configurable to recover a third data by combining a portion of a copy of a third data from the first storage cluster and the remaining portion of the copy of the third data from the second storage cluster;

the second storage cluster and the third storage cluster configurable to recover the first data by combining the portion of the copy of the first data from the second storage cluster and a remaining portion of the copy of the first data from the third storage cluster; and

the third storage cluster and the first storage cluster configurable to recover the second data by combining the portion of the copy of the second data from the third storage cluster and a remaining portion of the copy of the second data from the first storage cluster.