Decoupled reliability groups

• US 9,367,394 B2
• Filed: 12/07/2012
• Issued: 06/14/2016
• Est. Priority Date: 12/07/2012
• Status: Active Grant

• Current Assignee: NetApp, Inc.
• Original Assignee: NetApp, Inc.
• Inventors: Storer, Mark Walter
• Primary Examiner(s): Rones, Charles
• Assistant Examiner(s): DOAN, HAN V
• Application Number:
  US13/708,784
• Publication Number:
  US 20140164694A1
• Time in Patent Office: 1,285 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/1004 to protect a block of data words, e.g. CRC or checksum G06F11/1076 takes precedence; security arrangements for protecting computers or computer systems against unauthorized activity G06F21/00
• CPC Class Codes: G06F 11/1092 Rebuilding, e.g. when physically replacing a failing disk
• CPC Class Codes: G06F 11/2094 Redundant storage or storage space G06F11/2056 takes precedence
• CPC Class Codes: G06F 2201/805 Real-time
• CPC Class Codes: G06F 2211/1061 Parity-single bit-RAID4, i.e. RAID 4 implementations

Abstract

Methods and apparatuses for updating members of a data storage reliability group are provided. In one exemplary method, a reliability group includes a data zone in a first storage node and a checksum zone in a second data storage node. The method includes updating a version counter associated with the data zone in response to destaging a data object from a staging area of the data zone to a store area of the data zone without synchronizing the destaging with the state of the checksum zone. The method further includes transmitting, from the data zone to the checksum zone, an update message indicating completion of the destaging of the data object, wherein the update message includes a current value of the version counter.

Claims

1. A method comprising:

updating a version indicator associated with a data zone of a first data storage node in response to destaging a data object from a staging area of the data zone to a store area of the data zone, wherein the destaging is not synchronized with a state of a checksum zone in a second data storage node; and

transmitting, from the data zone to the checksum zone, an update message indicating completion of the destaging of the data object, wherein the update message includes a current value of the version indicator, wherein the second data storage node compares the value of the version indicator to one of one or more elements of a contribution vector to determine whether all updates associated with the data zone have already been processed by the checksum zone;

updating the one element of the contribution vector based on the current value of the version indicator in response to destaging the data object from the staging area of the data zone to the store area of the data zone;

wherein the data zone and the checksum zone are members of a reliability group, each of the one or more elements is associated with one data zone of the reliability group, and destaging the data object further comprises;

writing the data object to a location of the store area of the data zone without determining that the checksum zone is ready to perform a related action associated with the data object;

storing the location and the version indicator to a metadata area of the data zone; and

removing the data object from the staging area of the data zone.

2. The method of claim 1 further comprising, prior to the destaging, transmitting a replica of the data object to a staging area of the checksum zone.

3. The method of claim 1 further comprising:

receiving a request from the checksum zone for a prior update to the data zone, the request including a prior version number;

locating the prior update based on the prior version number and information retrieved from the metadata area; and

transmitting the prior update to the checksum zone.

4. The method of claim 1 wherein the writing to the store area is performed in an append-only log format.

5. The method of claim 1 wherein updating the version indicator includes setting the version indicator to a predetermined value if the data zone has not been previously updated.

6. The method of claim 1 wherein the first data storage node and the second data storage node are included in a data storage cluster.

7. The method of claim 1 further comprising transmitting the update message to a second checksum zone, wherein the second checksum zone is in a third data storage node.

8. The method of claim 1 wherein the update message is transmitted to the checksum zone for performing erasure coding of the data object.

9. A method comprising:

receiving, at a parity zone of a first data storage node from a data zone of a second data storage node, an update message indicating transfer of a data object from a staging area of the data zone to a store area of the data zone, the update message including a revision value associated with the data zone; and

storing checksum information, generated based on a replica of the data object stored in a staging area of the parity zone, in a store area of the parity zone if a comparison of the revision value to one of one or more elements of a contribution vector associated with the data zone indicates that all prior updates associated with the data zone have already been processed by the parity zone, wherein the storing of the checksum information is performed out of lockstep with the transfer of the data object from the staging area of the data zone to the store area of the data zone;

updating the one element of the contribution vector based on the revision value in response to destaging the data object from the staging area of the data zone to the store area of the data zone;

wherein the data zone and the parity zone are members of a reliability group, each of the one or more elements is associated with one data zone of the reliability group, the reliability group further includes another data zone having another data object, and the checksum information is generated further based on a replica of the other data object also stored in the staging area of the parity zone.

10. The method of claim 9 wherein the contribution vector further includes an element associated with the other data zone and is updated based on a revision value received in an update message from the other data zone.

11. The method of claim 9 wherein the first data storage node and the second data storage node comprise a data storage cluster.

12. The method of claim 9 further comprising processing, in the parity zone before storing the checksum information, the prior updates associated with the data zone in an order indicated by revision values.

13. The method of claim 12 wherein processing the prior updates includes generating and storing checksum information in the parity zone for the prior updates.

14. The method of claim 9 wherein storing the checksum information includes updating the checksum information in-place in the store area of the parity zone.

15. The method of claim 9 further comprising, before storing the checksum information:

requesting from the data zone information pertaining to a prior update to the data zone if the comparison of the revision value to the element of the contribution vector associated with the data zone indicates that a prior update associated with the data zone has not been processed and a data object associated with the prior update is not available in the staging area of the parity zone.

16. The method of claim 9 further comprising:

transmitting an acknowledgement to the data zone in response to receiving the update message from the data zone; and

removing the data object from the staging area of the parity zone after storing the checksum information.

17. The method of claim 9 wherein the checksum information includes an erasure coded checksum.

18. A data storage system comprising:

a memory containing computer-executable instructions; and

a computer processor operably connected to the memory and configured to execute the instructions to;

store a data object in a staging area of a data storage zone;

transfer the data object to a store area of the data storage zone without establishing a commit protocol with a parity storage zone;

update a revision value associated with the data storage zone in response to the transfer;

identify prior unprocessed updates associated with the data storage zone by comparing the revision value to one of one or more elements of a contribution vector associated with the data storage zone;

sequentially generate and store, in a store area of the parity storage zone, checksum information associated with the prior unprocessed updates in an order determined based on the revision value and revision values of the unprocessed updates; and

update the element of the contribution vector based on the revision value associated with the data storage zone in response to destaging the data object from the staging area of the data storage zone to a store area of the data storage zone;

wherein each of the one or more elements is associated with one data storage zone of a reliability group, the reliability group includes the data storage zone and the parity storage zones as members, and the checksum information is generated using erasure coding.

19. The system of claim 18 further comprising:

additional data storage zones; and

additional parity storage zones; and

wherein the checksum information is generated based on the data storage zones and the parity zones using Reed Solomon encoding.

20. The system of claim 19 implemented in a redundant array of independent nodes (RAIN) configuration wherein the parity storage zones and the data storage zones are each located in a different node of the independent nodes.

21. The system of claim 18 wherein the data object is stored in the store area of the data storage zone in append-only log format.

22. The system of claim 18 wherein the checksum information is stored in the store area of the parity storage zone in-place relative to a prior version of the checksum information.

23. The system of claim 18 wherein the parity storage zone is further configured to:

determine if information associated with one of the prior unprocessed updates is not available in the parity storage zone; and

request the information from the data storage zone.

24. The system of claim 18 wherein the parity storage zone is further configured to transmit an acknowledgement to the data storage zone in response to receiving the update from the data storage zone.

25. The system of claim 18 wherein the storage system is operated in a Network Attached Storage (NAS) environment or in a Storage Area Network (SAN).

26. A method comprising:

operating a data storage system that includes a plurality of data elements and at least one checksum element; and

decoupling updating operations in the at least one checksum element from update operations in the data elements by staging data in the checksum element and allowing the checksum element to update error correction data with a timing that is independent of the update operations in the data elements through use of update counters and a contribution vector, wherein the contribution vector includes one or more vector elements each of which is associated with one data element of the plurality of data elements;

updating the element of the contribution vector based on the update counter in response to destaging the data element from a staging area to a store area;

wherein destaging the data element from the staging area to the store area comprises;

writing the data element to a location of the store area without determining that the checksum element associated with the data element being ready to update;

storing the location and a version indicator to a metadata area of the data area; and

removing the data element from the staging area of the data area.