REBUILDING DATA ACROSS STORAGE NODES

• US 20150355974A1
• Filed: 06/04/2014
• Published: 12/10/2015
• Est. Priority Date: 06/04/2014
• Status: Active Grant

• Current Assignee: Pure Storage, Inc.
• Original Assignee: Pure Storage, Inc.
• Inventors: Colgrove, John, Lee, Robert, Robinson, Joshua, Ostrovsky, Igor, HAYES, John
• Granted Patent: US 9,218,244 B1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/1076 Parity data used in redundant arrays of independent storages, e.g. in RAID systems
• CPC Class Codes: G06F 11/108 Parity data distribution in semiconductor storages, e.g. in SSD
• CPC Class Codes: G06F 11/1088 Reconstruction on already foreseen single or plurality of spare disks
• CPC Class Codes: G06F 11/1096 Parity calculation or recalculation after configuration or reconfiguration of the system
• CPC Class Codes: G06F 11/1415 at system level

Abstract

A method for proactively rebuilding user data in a plurality of storage nodes of a storage cluster is provided. The method includes distributing user data and metadata throughout the plurality of storage nodes such that the plurality of storage nodes can read the user data, using erasure coding, despite loss of two of the storage nodes. The method includes determining that one of the storage nodes is unreachable and determining to rebuild the user data for the one of the storage nodes that is unreachable. The method includes reading the user data across a remainder of the plurality of storage nodes, using the erasure coding and writing the user data across the remainder of the plurality of storage nodes, using the erasure coding. A plurality of storage nodes within a single chassis that can proactively rebuild the user data stored within the storage nodes is also provided.

Claims

1. A plurality of storage nodes in a single chassis, comprising:

the plurality of storage nodes configured to communicate together as a storage cluster;

each of the plurality of storage nodes having nonvolatile solid-state memory for user data storage;

one or more controller of each of the plurality of storage nodes configured to distribute the user data and metadata throughout the plurality of storage nodes according to a first erasure coding scheme; and

the one or more controller of each of the plurality of storage nodes further configured to decide to utilize one of the first erasure coding scheme or a differing erasure coding scheme to rebuild the user data across stripes in a first set of the plurality of storage nodes from recovered user data striped across a second set of the plurality of storage nodes, the second set of the plurality of storage nodes having an unreachable storage node.

2. The storage cluster of claim 1, wherein the user data is rebuilt independent of a request from outside the storage cluster to read the user data, and wherein the plurality of storage nodes is configured to detect the unreachable storage node.

3. The storage cluster of claim 1, further comprising:

each of the plurality of storage nodes configured to emit a heartbeat, the heartbeat comprising a periodically issued message; and

each of the plurality of storage nodes configured to detect the heartbeat of another of the plurality of storage nodes.

4. The plurality of storage nodes of claim 1, wherein the single chassis is an enclosure with internal power distribution and an internal communication bus coupling the plurality of storage nodes, and wherein the single chassis couples the plurality of storage nodes to an external communication bus.

5. The storage cluster of claim 1, wherein:

the user data is striped across the first set of the plurality of storage nodes according to a RAID level, and the user data is rebuilt by reading and recovering the user data from the first set of the plurality of storage nodes according to the RAID level.

6. The storage cluster of claim 1, further comprising:

each of the plurality of storage nodes configured to consult others of the plurality of storage nodes to determine an erasure coding scheme for the user data that is striped across the first set of the plurality of storage nodes.

7. (canceled)

8. A plurality of storage nodes in a single chassis, comprising:

the plurality of storage nodes configured to communicate together as a storage cluster;

each of the plurality of storage nodes having nonvolatile solid-state memory for user data storage;

the plurality of storage nodes configured to distribute the user data and metadata throughout the plurality of storage nodes;

the plurality of storage nodes configured to recover user data written according to a first erasure coding scheme across the plurality of storage nodes; and

the plurality of storage nodes configured to write the recovered user data according to a second erasure coding scheme across the plurality of storage nodes, wherein each storage node is configured to decide whether the second erasure coding scheme is one of a same erasure coding scheme or a differing erasure coding scheme from the first erasure coding scheme.

9. The storage cluster of claim 8, wherein the solid-state memory includes flash.

10. The storage cluster of claim 8, further comprising:

each of the plurality of storage nodes configured to determine whether the one of the plurality of storage nodes is unreachable.

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

the plurality of storage nodes configured to initiate user data recovery based on lack of a heartbeat of the one of the plurality of storage nodes that is unreachable.

13. The storage cluster of claim 8, wherein user data recovery is initiated independent of an attempt to read the user data.

14. The storage cluster of claim 8, wherein to recover the user data comprises:

one of reading a redundant copy of the user data, or a stripe-wide read with parity correction of the user data, in accordance with the first erasure coding scheme.

11. (canceled)

15. A method for proactively rebuilding user data in a plurality of storage nodes of a storage cluster in a single chassis, comprising:

distributing user data and metadata throughout the plurality of storage nodes;

determining that one of the plurality of storage nodes is unreachable;

determining to rebuild the user data for the one of the plurality of storage nodes that is unreachable;

reading the user data across a remainder of the plurality of storage nodes, using erasure coding; and

writing the user data, as recovered user data, across the remainder of the plurality of storage nodes, using the erasure coding, wherein the reading is performed according to a first erasure coding scheme and the writing is performed according to a differing second erasure coding scheme, and wherein each storage node is configured to decide whether the second erasure coding scheme is one of a same erasure coding scheme or a differing erasure coding scheme from the first erasure coding scheme.

16. The method of claim 15, wherein reading the user data across the plurality of storage nodes using the erasure coding comprises:

recovering the user data in accordance with a data striping scheme governed by the first erasure coding.

17. The method of claim 15, wherein determining that the one of the plurality of storage nodes is unreachable is based on one of a lack of a heartbeat of the storage node, a lack of a response to a query of the storage node, or a timeout of the storage node.

18. The method of claim 15, wherein determining that the storage node is unreachable is independent of a request from outside the storage cluster to read the user data.

20. The method of claim 15, further comprising:

reassigning, to one of the remainder of the plurality of storage nodes, entities owned by the unreachable storage node, wherein the entities include units of data.

19. (canceled)