SYNCHRONOUS REPLICATION

• US 20170185306A1
• Filed: 12/28/2015
• Published: 06/29/2017
• Est. Priority Date: 12/28/2015
• Status: Active Grant

• Current Assignee: NetApp, Inc.
• Original Assignee: NetApp, Inc.
• Inventors: Eisler, Michael Robin, Kaushik, Akhil, Rao, Santosh Ananth, Mu, Yuedong
• Granted Patent: US 10,496,320 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 11/2058 using more than 2 mirrored copies
• CPC Class Codes: G06F 11/2064 while ensuring consistency
• CPC Class Codes: G06F 11/2069 Management of state, configuration or failover
• CPC Class Codes: G06F 11/2071 using a plurality of controllers
• CPC Class Codes: G06F 11/2076 Synchronous techniques
• CPC Class Codes: G06F 11/2082 Data synchronisation
• 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: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]

Abstract

One or more techniques and/or computing devices are provided for synchronous replication. For example, synchronous replication relationships are established between a first storage object (e.g., a file, a logical unit number (LUN), a consistency group, etc.), hosted by a first storage controller, and a plurality of replication storage objects hosted by other storage controllers. In this way, a write operation to the first storage object is implemented in parallel upon the first storage object and the replication storage objects in a synchronous manner, such as using a zero-copy operation to reduce overhead otherwise introduced by performing copy operations. Reconciliation is performed in response to a failure so that the first storage object and the replication storage objects comprise consistent data. Failed write operations and replication write operations are retried, while enforcing a single write semantic. Dependent write order consistency is enforced for dependent write operations, such as overlapping write operations.

Claims

1. A method comprising:

receiving, by a storage server, a write operation targeting a first storage object hosted by a first storage controller, the first storage object having a synchronous replication relationship with a first replication storage object hosted by a second storage controller and a second replication storage object hosted by a third storage controller;

splitting the write operation into a first replication write operation targeting the first replication storage object and a second replication write operation targeting the second replication storage object;

implementing the write operation upon the first storage object, the first replication write operation upon the first replication storage object, and the second replication write operation upon the second replication storage object in parallel; and

withholding a client acknowledgement that the write operation is complete until the write operation, the first replication write operation, and the second replication write operation are complete.

2. The method of claim 1, comprising:

responsive to detecting a failure associated with the first storage controller, performing a reconciliation between the first storage object, the first replication storage object, and the second replication storage object, wherein the performing a reconciliation comprises;

withholding a current client acknowledgment for an operation associated with the failure;

reading content, targeted by the operation, from the first storage object;

writing the content to the first replication storage object and the second replication storage object; and

releasing the current client acknowledgment.

3. The method of claim 2, comprising:

responsive to successfully writing the content so that data within the first storage object, the first replication storage object, and the second replication storage object have data symmetry, utilizing one or more consistency points to perform a data integrity validation to flag data corruption or data loss issues.

4. The method of claim 1, comprising:

utilizing an inflight time tracking module to track a lifecycle of the write operation.

5. The method of claim 4, comprising:

responsive to the lifecycle indicating that performance of at least one of the first replication write operation or the second replication write operation violates a protocol timeout guarantee, releasing the client acknowledgement.

6. The method of claim 4, comprising:

responsive to the lifecycle indicating that performance of the first replication write operation has encountered an issue, retrying the first replication write operation as a retry first replication write operation.

7. The method of claim 6, comprising:

enforcing a single write sematic for the first replication storage object, the single write semantic specifying;

responsive to the first replication write operation being successfully performed, disregarding the retry first replication write operation; and

responsive to the retry first replication write operation being successfully performed, disregarding the first replication write operation.

8. The method of claim 1, comprising:

receiving a second write operation; and

responsive to the second write operation and the write operation having a dependency, enforcing a dependent write order consistency for the write operation and the second write operation.

9. The method of claim 8, comprising:

determining that the second write operation the write operation have the dependency based upon the write operation and the second write operation overlapping a same region that is to be modified.

10. The method of claim 8, wherein the enforcing a dependent write order consistency comprises:

serially implementing the write operation and the second write operation.

11. The method of claim 1, wherein the implementing comprises:

performing a zero-copy operation upon the first storage controller, the second storage controller, and the third storage controller, the zero-copy operation using a data reference in place of copying data associated with the write operation by using an original copy of data within the write operation for at least one of an initial replication attempt or a subsequent replication attempt.

12. The method of claim 1, comprising:

receiving a metadata operation for the first storage controller; and

implementing the metadata operation upon the first storage controller and replicating the metadata operation to the second storage controller and the third storage controller.

13. The method of claim 12, wherein the metadata operation comprises at least one of an application consistency point creation operation, a resync point creation operation, a clone operation, an attribute modification operation, a create operation, a volume operation, a logical unit number (LUN) operation, or a delete operation.

14. The method of claim 1, wherein the first storage object is exposed with read and write access, and the first replication storage object and the second replication storage object are exposed with read only access with a write access option when a failover is performed.

15. The method of claim 1, wherein the first storage controller and the second storage controller are maintained within a same data center.

16. The method of claim 1, wherein the first storage controller and the second storage controller are maintained within different data centers.

17. The method of claim 1, comprising:

performing load balancing between the first replication storage object and the second replication storage object.

18. A non-transitory machine readable medium having stored thereon instructions for performing a method comprising machine executable code which when executed by at least one machine, causes the machine to:

receive a first replication write operation, corresponding to a replication of a write operation received by a first storage controller for implementation upon a first storage object hosted by the first storage controller, for parallel implementation upon a first replication storage object hosted by a second storage controller;

receive a retry first replication write operation corresponding to a retry of the first replication write operation triggered in response to a lifecycle of the write operation indicating that the first replication write operation previously encountered an issue; and

responsive to the retry first replication write operation being received before the first replication write operation;

implement the retry first replication write operation upon the first replication storage object; and

disregard the first replication write operation based upon an enforcement of a single write semantic for the first replication storage object.

19. The non-transitory machine readable medium of claim 18, wherein the machine executable code causes the machine to:

responsive to the first replication write operation being received before the retry first replication write operation;

implement the first replication write operation upon the first replication storage object; and

disregard the retry first replication write operation based upon enforcement of the single write semantic for the first replication storage object.

20. A computing device comprising:

a memory containing machine readable medium comprising machine executable code having stored thereon instructions for performing a method of synchronous replication; and

a processor coupled to the memory, the processor configured to execute the machine executable code to cause the processor to;

receive a write operation targeting a first storage object hosted by a first storage controller, the first storage object having a synchronous replication relationship with a first replication storage object hosted by a second storage controller and a second replication storage object hosted by a third storage controller;

split the write operation into a first replication write operation targeting the first replication storage object and a second replication write operation targeting the second replication storage object;

utilize a zero-copy operation to implement the write operation upon the first storage object, the first replication write operation upon the first replication storage object, and the second replication write operation upon the second replication storage object in parallel, the zero-copy operation using a data reference in place of copying data associated with the write operation; and

withhold a client acknowledgement that the write operation is complete until the write operation, the first replication write operation, and the second replication write operation are complete.