Overlapping write detection and processing for sync replication

• US 10,248,341 B2
• Filed: 03/23/2018
• Issued: 04/02/2019
• Est. Priority Date: 08/29/2014
• Status: Active Grant

• Current Assignee: NetApp, Inc.
• Original Assignee: NetApp, Inc.
• Inventors: Sundararajan, Manoj V., Ngan, Ching-Yuk Paul, Mu, Yuedong, Coatney, Susan M.
• Primary Examiner(s): Nguyen, Hiep T
• Application Number:
  US15/933,556
• Publication Number:
  US 20180217771A1
• Time in Patent Office: 375 Days
• 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

A primary write request that is to modify a primary portion of primary data stored in a primary storage node is received. The primary write request is to be replicated to create a current secondary write request. The current secondary write request is to modify a current secondary portion of secondary data that is stored in a secondary storage node. A current data range of the current secondary portion is determined. A determination is made of whether a previous secondary write request is in process of modifying a previous data range that at least partially overlaps with a current data range of the current secondary portion. Execution of the primary write request is suspended, until the previous secondary write request has completed updating the secondary storage node.

Claims

1. A method comprising:

replicating, by a first node, a write request to modify data stored by the first node as a replicated write request targeting second data stored by a second node based upon a sync mapping indicating that the write request is to be replicated to the second node;

determining, by the first node, that the replicated write request overlaps a pending replicated write request pending at the second node based upon the replicated write request and the pending replicated write request targeting a same portion of a data range of the second data; and

suspending execution of the write request at the first node upon the data stored by the first node until completion of the pending replicated write request at the second node upon the second data.

2. The method of claim 1, comprising:

maintaining the sync mapping indicating synchronization configuration among logical storage objects of the first node and the second node.

3. The method of claim 2, comprising:

specifying, within the sync mapping, a first mapping between a first logical object identifier of a first logical object stored by the first node to a second logical object identifier of a second logical object stored by the second node.

4. The method of claim 2, comprising:

specifying, within the sync mapping, a first mapping between first filesystem location information of the first data stored by the first node to a second filesystem location information of the second data stored by the second node.

5. The method of claim 1, the determining that the replicated write request overlaps comprising:

identifying i) a first data range, between a first offset and a second offset, targeted by the write request and ii) a second data range, between a third offset and a fourth offset, targeted by a first write request, targeting data of the first node, from which the pending replicated write request was replicated, wherein the replicated write request is determined as overlapping the pending replicated write request based upon the third offset being between the first offset and the second offset.

6. The method of claim 5, comprising:

determining that the replicated write request overlaps the pending replicated write request based upon the fourth offset being between the first offset and the second offset.

7. The method of claim 5, comprising:

determining that the replicated write request overlaps the pending replicated write request based upon the first offset being between the third offset and the fourth offset.

8. The method of claim 5, comprising:

determining that the replicated write request overlaps the pending replicated write request based upon the second offset being between the third offset and the fourth offset.

9. The method of claim 1, comprising:

maintaining a range lock tree comprising a hierarchical tree structure with a root node and substrees of children nodes comprising a set of linked nodes, wherein a first node comprises a first data structure that stores a first data range that is currently locked and stores references to children nodes of the first node.

10. The method of claim 9, wherein the first data range is locked based upon a first pending replicated write request, targeting the first data range at the second node, that has not yet completed.

11. The method of claim 9, the determining that the replicated write request overlaps comprising:

traversing the range lock tree using a targeted data range of the write request to identify the pending replicated write request that has not completed at the second node based upon the range lock tree comprising a node corresponding to the targeted data range.

12. The method of claim 9, comprising:

traversing the range lock tree using a targeted data range of the write request to determine that there is no overlap of the replicated write request with pending replicated write requests at the second node based upon the range lock tree not comprising a node corresponding to the targeted data range.

13. A non-transitory machine readable medium comprising instructions for performing a method, which when executed by a machine, causes the machine to:

replicate, by a first node, a write request to modify data stored by the first node as a replicated write request targeting second data stored by a second node based upon a sync mapping indicating that the write request is to be replicated to the second node;

determine, by the first node, that the replicated write request overlaps a pending replicated write request pending at the second node based upon the replicated write request and the pending replicated write request targeting a same portion of a data range of the second data; and

suspend execution of the write request at the first node upon the data stored by the first node until completion of the pending replicated write request at the second node upon the second data.

14. The non-transitory machine readable medium of claim 13, wherein the instructions cause the machine to:

maintain a range lock tree comprising a hierarchical tree structure with a root node and substrees of children nodes comprising a set of linked nodes, wherein a first node comprises a first data structure that stores a first data range that is currently locked and stores references to children nodes of the first node.

15. The non-transitory machine readable medium of claim 13, wherein the instructions cause the machine to:

maintain the sync mapping indicating synchronization configuration among logical storage objects of the first node and the second node.

16. The non-transitory machine readable medium of claim 15, wherein the instructions cause the machine to:

specify, within the sync mapping, a first mapping between a first logical object identifier of a first logical object stored by the first node to a second logical object identifier of a second logical object stored by the second node.

17. The non-transitory machine readable medium of claim 15, wherein the instructions cause the machine to:

specify, within the sync mapping, a first mapping between first filesystem location information of the data stored by the first node to a second filesystem location information of the second data stored by the second node.

18. A computing device comprising:

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

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

replicate, by a first node, a write request to modify data stored by the first node as a replicated write request targeting second data stored by a second node based upon a sync mapping indicating that the write request is to be replicated to the second node;

determine, by the first node, that the replicated write request overlaps a pending replicated write request pending at the second node based upon the replicated write request and the pending replicated write request targeting a same portion of a data range of the second data; and

suspend execution of the write request at the first node upon the data stored by the first node until completion of the pending replicated write request at the second node upon the second data.

19. The computing device of claim 18, wherein the machine executable code causes the processor to:

maintain the sync mapping indicating synchronization configuration among logical storage objects of the first node and the second node.

20. The computing device of claim 19, wherein the machine executable code causes the processor to:

specify, within the sync mapping, a first mapping between a first logical object identifier of a first logical object stored by the first node to a second logical object identifier of a second logical object stored by the second node.