Systems and methods for maintaining cache coherency

• US 10,452,545 B1
• Filed: 10/06/2017
• Issued: 10/22/2019
• Est. Priority Date: 10/06/2017
• Status: Active Grant

• Current Assignee: Veritas Technologies, LLC (Whitehouse Group Ltd.)
• Original Assignee: Veritas Technologies, LLC (Whitehouse Group Ltd.)
• Inventors: Patidar, Jitendra, Banerjee, Anindya, Kumar, Kundan
• Primary Examiner(s): Khan, Masud K
• Application Number:
  US15/727,244
• Time in Patent Office: 746 Days
• CPC Class Codes: G06F 12/0815 Cache consistency protocols
• CPC Class Codes: G06F 12/084 with a shared cache
• CPC Class Codes: G06F 12/0866 for peripheral storage systems, e.g. disk cache
• CPC Class Codes: G06F 12/0868 Data transfer between cache memory and other subsystems, e.g. storage devices or host systems
• CPC Class Codes: G06F 12/0875 with dedicated cache, e.g. instruction or stack
• CPC Class Codes: G06F 2212/1016 Performance improvement
• CPC Class Codes: G06F 2212/154 Networked environment
• CPC Class Codes: G06F 2212/283 Plural cache memories
• CPC Class Codes: G06F 9/526 Mutual exclusion algorithms

Abstract

The disclosed computer-implemented method for maintaining cache coherency may include (1) receiving an indication of a revocation of a grant to an owner node of a shared lock for a data object owned by the owner node; (2) invalidating, in response to the indication of the revocation, a copy of the object in a local cache of the owner node; (3) receiving, while the requesting node holds a grant of an exclusive lock, a request from the requesting node to update the object in the local cache; (4) transmitting, in response to the request to update the object, a request for a shared lock for the object; (5) receiving an indication of a grant of the shared lock; and (6) updating, in response to the grant of the shared lock, the object in the local cache. Various other methods, systems, and computer-readable media are also disclosed.

Claims

1. A computer-implemented method for maintaining cache coherency, at least a portion of the method being performed by a computing device comprising at least one processor, the method comprising:

receiving, by an owner node of a plurality of nodes of a clustered storage system, an indication of a revocation of a grant to the owner node of a shared lock for a data object owned by the owner node;

invalidating, by the owner node in response to the indication of the revocation of the grant to the owner node of the shared lock, a copy of the data object in a local cache of the owner node;

receiving, by the owner node while a first requesting node of the plurality of nodes holds a grant of an exclusive lock, a request from the first requesting node to update the data object in the local cache and a first validation cookie for the request from the first requesting node;

receiving, by the owner node while a second requesting node of the plurality of nodes holds a grant of the exclusive lock, a request from the second requesting node to update the data object in the local cache and a second validation cookie for the request from the second requesting node;

transmitting, by the owner node in response to one of the requests to update the data object, a non-preferred request for a shared lock for the data object;

receiving, by the owner node, an indication of a grant of the shared lock to the owner node and a third validation cookie, wherein the grant of the shared lock occurs after grants of preceding requests for a lock for the data object;

determining, by the owner node, whether the third validation cookie matches the first validation cookie or the second validation cookie; and

updating, by the owner node in response to the grant of the shared lock, the data object in the local cache based on determining whether the third validation cookie matches the first validation cookie or the second validation cookie.

2. The computer-implemented method of claim 1, wherein the updating of the data object in the local cache comprises populating the data object in the local cache with data received in one of the requests to update the data object.

3. The computer-implemented method of claim 1, further comprising:

receiving, by the owner node from a third requesting node of the plurality of nodes, a request to read the data object; and

transmitting, by the owner node to the third requesting node in response to the request to read the data object, the data object from the local cache of the owner node to the third requesting node.

4. The computer-implemented method of claim 1, wherein the local cache of the owner node is located on a solid-state drive of the owner node.

5. The computer-implemented method of claim 1, wherein updating the data object comprises:

populating, by the owner node in response to the grant of the shared lock to the owner node, the data object in the local cache of the owner node with data received in the request from the first requesting node based on the third validation cookie matching the first validation cookie.

6. The computer-implemented method of claim 1, wherein updating the data object comprises:

populating, by the owner node in response to the grant of the shared lock to the owner node, the data object in the local cache of the owner node with data received in the request from the second requesting node based on the third validation cookie matching the second validation cookie.

7. The computer-implemented method of claim 1, wherein invalidating the copy of the data object in the local cache of the owner node occurs concurrently with invalidating a copy of the data object in a local cache of at least one other node of the plurality of nodes.

8. The computer-implemented method of claim 1, wherein receiving the request from the first requesting node to update the data object in the local cache comprises storing, by the owner node, data in the request from the first requesting node to update the data object to an in-memory cache of the owner node that is separate from the local cache of the owner node.

9. The computer-implemented method of claim 1, wherein the grant to the first requesting node of the exclusive lock for the data object is responsive to a request by the first requesting node for the grant of the exclusive lock after receiving a write request for the data object from a client computer system.

10. The computer-implemented method of claim 1, wherein the grant to the first requesting node of the exclusive lock for the data object causes an initiation of an update operation of the data object by the first requesting node to a backend storage system accessible by the owner node.

11. A system for maintaining cache coherency, the system comprising:

a lock communication module, stored in memory, that receives an indication of a revocation of a grant to an owner node of a plurality of nodes of a clustered storage system of a shared lock for a data object owned by the owner node;

a local cache control module, stored in memory, that invalidates, in response to the indication of the revocation of the grant to the owner node of the shared lock, a copy of the data object in a local cache of the owner node; and

a node communication module, stored in memory, that;

receives, while a first requesting node of the plurality of nodes holds a grant of an exclusive lock, a request from the first requesting node to update the data object in the local cache and a first validation cookie for the request from the first requesting node; and

receiving, while a second requesting node of the plurality of nodes holds a grant of the exclusive lock, a request from the second requesting node to update the data object in the local cache and a second validation cookie for the request from the second requesting node;

wherein the lock communication module transmits, in response to one of the requests to update the data object, a non-preferred request for a shared lock for the data object;

wherein the lock communication module receives an indication of a grant of the shared lock to the owner node and a third validation cookie, wherein the grant of the shared lock occurs after grants of preceding requests for a lock for the data object; and

wherein the local cache control module;

determines whether the third validation cookie matches the first validation cookie or the second validation cookie; and

updates, in response to the grant of the shared lock, the data object in the local cache based on determining whether the third validation cookies matches the first validation cookie or the second validation cookie; and

at least one physical processor that executes the lock communication module, the local cache control module, and the node communication module.

12. The system of claim 11, wherein the local cache control module updates the data object in the local cache by populating the data object in the local cache with data received in one of the requests to update the data object.

13. The system of claim 11, wherein the node communication module:

receives, from a third requesting node of the plurality of nodes, a request to read the data object; and

transmits, to the third requesting node in response to the request to read the data object, the data object from the local cache of the owner node to the third requesting node.

14. The system of claim 11, wherein the local cache control module updates the data object by populating, in response to the grant of the shared lock to the owner node, the data object in the local cache of the owner node with data received in the request from the first requesting node based on the third validation cookie matching the first validation cookie.

15. The system of claim 11, wherein the local cache control module updates the data object by populating, in response to the grant of the shared lock to the owner node, the data object in the local cache of the owner node with data received in the request from the second requesting node based on the third validation cookie matching the second validation cookie.

16. A non-transitory computer-readable medium comprising one or more computer-executable instructions that, when executed by at least one processor of a computing device, cause the computing device to:

receive an indication of a revocation of a grant to an owner node of a plurality of nodes of a clustered storage system of a shared lock for a data object owned by the owner node;

invalidate, in response to the indication of the revocation of the grant to the owner node of the shared lock, a copy of the data object in a local cache of the owner node;

receive, while a first requesting node of the plurality of nodes holds a grant of an exclusive lock, a request from the first requesting node to update the data object in the local cache and a first validation cookie for the request from the first requesting node;

receive, while a second requesting node of the plurality of nodes holds a grant of the exclusive lock, a request from the second requesting node to update the data object in the local cache and a second validation cookie for the request from the second requesting node;

transmit, in response to one of the requests to update the data object, a non-preferred request for a shared lock for the data object;

receive an indication of a grant of the shared lock to the owner node and a third validation cookie, wherein the grant of the shared lock occurs after grants of preceding requests for a lock for the data object;

determine whether the third validation cookie matches the first validation cookie or the second validation cookie; and

update, in response to the grant of the shared lock, the data object in the local cache based on determining whether the third validation cookie matches the first validation cookie or the second validation cookie.

17. The non-transitory computer-readable medium of claim 16, wherein updating the data object in the local cache comprises populating the data object in the local cache with data received in one of the requests to update the data object.

18. The non-transitory computer-readable medium of claim 16, wherein the one or more computer-executable instructions further cause the computing device to:

receive, from a third requesting node of the plurality of nodes, a request to read the data object; and

transmitting, to the third requesting node in response to the request to read the data object, the data object from the local cache of the owner node to the third requesting node.

19. The non-transitory computer-readable medium of claim 16, wherein updating the data object comprises:

populating, in response to the grant of the shared lock to the owner node, the data object in the local cache of the owner node with data received in the request from the first requesting node based on the third validation cookie matching the first validation cookie.

20. The non-transitory computer-readable medium of claim 16, wherein updating the data object comprises:

populating, in response to the grant of the shared lock to the owner node, the data object in the local cache of the owner node with data received in the request from the second requesting node based on the third validation cookie matching the second validation cookie.