Non-shared write-back caches in a cluster environment
View Patent ↗Various systems and methods for performing write-back caching in a cluster. For example, one method can involve a first node detecting that no failover nodes are available. A determination is made whether the first node should use write-back caching or not. If the first node is to continue using write-back caching, a first local cache identifier and a global cache identifier are both updated.
1. A method comprising:
a first node detecting that no failover nodes are available, wherein
the first node comprises a local cache and a first local cache identifier,
the first node is coupled to a shared storage device, and
the shared storage device comprises a global cache identifier;
determining whether to allow write-back caching on the first node, wherein
the determining is performed in response to the detecting; and
if write-back caching is allowed on the first node, incrementing the first local cache identifier and the global cache identifier.
2. The method of claim 1 , further comprising:
performing a cache recovery operation in response to detecting that write-back caching is disallowed on the first node.
3. The method of claim 2 , wherein the performing comprises:
copying data from the local cache to the shared storage device;
preventing further data modifications from being stored in the local cache; and
updating a tracking structure to include information that identifies data modified during the recovery operation.
4. The method of claim 2 , further comprising:
updating cache status information to indicate that write-back caching is disallowed for the shared storage device, wherein
the updating is performed in response to determining that all data in the local cache is copied to the shared storage device; and
joining a second node to a cluster, wherein
the second node comprises a second local cache and a second local cache identifier.
5. The method of claim 4 , further comprising:
invalidating contents of the second local cache, wherein
the invalidating is performed in response to detecting that write-back caching is disallowed for the shared storage device.
6. The method of claim 4 , further comprising:
comparing the second local cache identifier with the global cache identifier; and
invalidating contents of the second local cache, wherein
the invalidating is performed in response to detecting that there is a mismatch between the second local cache identifier and the global cache identifier.
7. The method of claim 4 , further comprising:
determining whether write-back caching is allowed on the second node;
incrementing the second local cache identifier, wherein
the incrementing is performed in response to determining that write-back caching is allowed on the second node.
8. The method of claim 2 , further comprising:
detecting that the first node has failed, wherein
the first node failed prior to completion of the recovery operation; and
joining a second node to a cluster, wherein
the second node comprises a second local cache and a second local cache identifier.
9. The method of claim 8 , further comprising:
detecting that contents of the second local cache are valid, except for data modified during the recovery operation, wherein
the detecting that contents of the second local cache are valid comprises accessing a tracking structure that includes information that identifies the data modified during the recovery operation.
10. A non-transitory computer readable storage medium storing program instructions executable to:
detect that no failover nodes are available, wherein
detecting that no failover nodes are available is performed by a first node,
the first node comprises a local cache and a first local cache identifier,
the first node is coupled to a shared storage device, and
the shared storage device comprises a global cache identifier;
determine whether to allow write-back caching on the first node, wherein
determining whether to allow write-back caching on the first node is performed in response to detecting that no failover nodes are available; and
if write-back caching is allowed on the first node, increment the first local cache identifier and the global cache identifier.
11. The non-transitory computer readable storage medium of claim 10 , wherein the instructions are further executable to:
perform a cache recovery operation in response to detecting that write-back caching is disallowed on the first node.
12. The non-transitory computer readable storage medium of claim 11 , wherein the instructions are further executable to:
update cache status information to indicate that write-back caching is disallowed for the shared storage device, wherein
updating the cache status information is performed in response to determining that all data in the local cache is copied to the shared storage device; and
join a second node to a cluster, wherein
the second node comprises a second local cache and a second local cache identifier.
13. The non-transitory computer readable storage medium of claim 12 , wherein the instructions are further executable to:
determine whether write-back caching is allowed on the second node;
increment the second local cache identifier in response to determining that write-back caching is allowed on the second node.
14. The non-transitory computer readable storage medium of claim 12 , wherein the instructions are further executable to:
compare the second local cache identifier with the global cache identifier; and
invalidate contents of the second local cache, wherein
invalidating contents of the second local cache is performed in response to detecting that there is a mismatch between the second local cache identifier and the global cache identifier.
15. The non-transitory computer readable storage medium of claim 12 , wherein the instructions are further executable to:
detect that contents of the second local cache are valid, except for data modified during the recovery operation, wherein
detecting that contents of the second local cache are valid comprises accessing a tracking structure that includes information that identifies the data modified during the recovery operation.
16. A system comprising:
one or more processors; and
a memory coupled to the one or more processors, wherein the memory stores program instructions executable by the one or more processors to:
detect that no failover nodes are available, wherein
detecting that no failover nodes are available is performed by a first node,
the first node comprises a local cache and a first local cache identifier,
the first node is coupled to a shared storage device, and
the shared storage device comprises a global cache identifier;
determine whether to allow write-back caching on the first node, wherein
determining whether to allow write-back caching on the first node is performed in response to detecting that no failover nodes are available; and
if write-back caching is allowed on the first node, increment the first local cache identifier and the global cache identifier.
17. The system of claim 16 , wherein the instructions are further executable to:
perform a cache recovery operation in response to detecting that write-back caching is disallowed on the first node;
update cache status information to indicate that write-back caching is disallowed for the shared storage device, wherein
updating the cache status information is performed in response to determining that all data in the local cache is copied to the shared storage device; and
join a second node to a cluster, wherein
the second node comprises a second local cache and a second local cache identifier.
18. The system of claim 17 , wherein the instructions are further executable to:
determine whether write-back caching is allowed on the second node;
increment the second local cache identifier in response to determining that write-back caching is allowed on the second node.
19. The system of claim 17 , wherein the instructions are further executable to:
compare the second local cache identifier with the global cache identifier; and
invalidate contents of the second local cache, wherein
invalidating contents of the second local cache is performed in response to detecting that there is a mismatch between the second local cache identifier and the global cache identifier.
20. The system of claim 16 , wherein the instructions are further executable to:
detect that contents of the second local cache are valid, except for data modified during the recovery operation, wherein
detecting that contents of the second local cache are valid comprises accessing a tracking structure that includes information that identifies the data modified during the recovery operation.