Granted Patent B2

US 12,204,413 · App. 17/900,596 · Granted Jan 21, 2025

Inventors: Ronald Karr (Palo Alto, CA); Robert Lee (San Carlos, CA); Igor Ostrovsky (Mountain View, CA)

Assignee: PURE STORAGE, INC.

G06F11/1451G06F3/0619G06F3/065G06F3/067G06F11/1435G06F11/1464G06F2201/84

Abstract

Claims (51)

1. A method, comprising:

recording, by a coordinating agent in a data structure stored in memory of one or more storage nodes of a storage system, an intent to create a snapshot of contents of the storage system associated with an existing snapshot copy identifier, wherein the data structure stored in the memory is monitored by distributed agents of the storage system to identify the intent to create the snapshot;

switching to a new snapshot copy identifier;

receiving, by the coordinating agent, corresponding requests for a snapshot copy identifier from the distributed agents of the storage system performing input/output (I/O) operations after recording the intent to create the snapshot;

responding to the corresponding requests with the new snapshot copy identifier; and

committing the snapshot and exiting a snapshot creation mode upon receiving, by the coordinating agent, responses from each distributed agent of the storage system that service is complete for all I/O operations that started before recording the intent to create the snapshot.

2. The method of claim 1 , wherein the one or more storage nodes have the distributed agents distributed among the one or more storage nodes.

3. The method of claim 1 , wherein the recording of the intent to create the snapshot is in response to receiving a remote procedure call.

4. The method of claim 1 , further comprising:

communicating to each of the distributed agents that the storage system is entering the snapshot creation mode.

5. The method of claim 1 , further comprising:

directing each of the distributed agents to complete service, using the existing snapshot copy identifier, to all of the I/O operations that started before recording the intent to create the snapshot.

6. The method of claim 1 , further comprising:

directing each of the distributed agents to complete service, using the existing snapshot copy identifier, to pending I/O operations, responsive to starting up and finding the recorded intent to create the snapshot.

7. The method of claim 6 , further comprising:

resuming creation of the snapshot, wherein the committing the snapshot is responsive to receiving responses from the distributed agents that service is complete for the pending I/O operations.

8. A non-transitory computer-readable storage medium having instructions thereupon which, when executed by a processor, cause the processor to:

record, by a coordinating agent in a data structure stored in memory of one or more storage nodes of a storage system, intent to create a snapshot of contents of the storage system associated with an existing snapshot copy identifier, wherein the data structure stored in the memory is monitored by distributed agents of the storage system to identify the intent to create the snapshot;

switch to a new snapshot copy identifier;

receive, by the coordinating agent, corresponding requests for a snapshot copy identifier from the distributed agents of the storage system performing input/output (I/O) operations after recording the intent to create the snapshot;

respond to the corresponding requests with the new snapshot copy identifier; and

commit the snapshot and exiting a snapshot creation mode, responsive to receiving, by the coordinating agent, confirmation from each distributed agent of the storage system that service is complete for all I/O operations that started before recording the intent to create the snapshot.

9. The non-transitory computer-readable storage medium of claim 8 , wherein the processor is further to:

communicate to the distributed agents that the snapshot is visible, wherein each of the distributed agents recognizes the snapshot and grants read-only access to the snapshot.

10. The non-transitory computer-readable storage medium of claim 8 , wherein to record the intent to create the snapshot the processor is further to:

write metadata into a data structure stored in the memory of the one or more storage nodes, the metadata indicating the intent to create the snapshot.

11. The non-transitory computer-readable storage medium of claim 8 , wherein to record the intent to create the snapshot, the processor is further to:

receive a remote procedure call, wherein recording the intent to create the snapshot is in response to receiving the remote procedure call.

12. The non-transitory computer-readable storage medium of claim 8 , wherein the processor is further to:

communicate an error response that indicates no snapshot creation is in progress, responsive to starting up and finding one or more of the distributed agents is in the snapshot creation mode but finding no record of the intent to create the snapshot, as a failover before precommitment of the snapshot.

13. The non-transitory computer-readable storage medium of claim 8 , wherein the processor is further to:

direct each of the distributed agents to complete service, using the existing snapshot copy identifier, to pending I/O operations, responsive to starting up and finding the recorded intent to create the snapshot; and

resume creation of the snapshot, responsive to receiving responses from the distributed agents that service is complete for the pending I/O operations, as a failover after precommitment of the snapshot.

14. A storage system, comprising:

a memory; and

a processor, operatively coupled to the memory, configured to:

record, by a coordinating agent in a data structure stored in memory of one or more storage nodes of a storage system, intent to create a snapshot of contents of the storage system associated with an existing snapshot copy identifier, wherein the data structure stored in the memory is monitored by distributed agents of the storage system to identify the intent to create the snapshot;

switch to a new snapshot copy identifier;

receive, by the coordinating agent, corresponding requests for a snapshot copy identifier from the distributed agents of the storage system performing input/output (I/O) operations after recording the intent to create the snapshot;

respond to the corresponding requests with the new snapshot copy identifier; and

commit the snapshot and exiting a snapshot creation mode, responsive to receiving, by the coordinating agent, confirmation from each distributed agent of the storage system that service is complete for all I/O operations that started before recording the intent to create the snapshot.

15. The storage system of claim 14 , wherein the processor is further configured to:

communicate to the distributed agents that the snapshot is visible, wherein each of the distributed agents recognizes the snapshot and grants read-only access to the snapshot.

16. The storage system of claim 14 , to record the intent to create the snapshot the processor is further configured to:

write metadata into a data structure stored in the memory of the one or more storage nodes, the metadata indicating the intent to create the snapshot.

17. The storage system of claim 14 , wherein the recording of the intent to create the snapshot is in response to receiving a remote procedure call.

18. The storage system of claim 14 , wherein the processor is further configured to:

communicate an error response that indicates no snapshot creation is in progress, responsive to starting up and finding one or more of the distributed agents is in the snapshot create mode but finding no record of the intent to create the snapshot, as a failover before precommitment of the snapshot.

19. The storage system of claim 14 , wherein the processor is further configured to:

direct each of the distributed agents to complete service, using the existing snapshot copy identifier, to pending I/O operations, responsive to starting up and finding the recorded intent to create the snapshot; and

resume creation of the snapshot, responsive to receiving responses from the distributed agents that service is complete for the pending I/O operations, as a failover after precommitment of the snapshot.

Continuity (3)

References Cited (400)