Fast cache reheat
Embodiments of the present invention allow for fast cache reheat by periodically storing a snapshot of information identifying the contents of the cache at the time of the snapshot, and then using the information from the last snapshot to restore the contents of the cache following an event that causes loss or corruption of cache contents such as a loss of power or system reset. Since there can be a time gap between the taking of a snapshot and such an event, the actual contents of the cache, and hence the corresponding data stored in a data store, may have changed since the last snapshot was taken. Thus, the information stored at the last snapshot is used to retrieve current data from the data store for use in restoring the contents of the cache.
1. A method for fast cache reheat in a data storage system, the method comprising:
providing a volatile memory cache that includes content corresponding to a set of data blocks in a non-volatile second data store, wherein the non-volatile second data store is disk storage or solid state storage;
periodically storing, in a non-volatile first data store, a snapshot of an index identifying storage locations of the set of data blocks, wherein such periodically storing of the snapshot is performed prior to a restart of the data storage system;
upon a restart of the data storage system causing loss or corruption of the contents of the volatile memory cache, restoring the cache from data in the non-volatile second data store that is persistently stored and available for retrieval upon the restart of the data storage system by:
retrieving from the first data store the index from the last snapshot stored prior to the restart, the last snapshot identifying content of the cache at a time of storing the last snapshot;
retrieving, from the non-volatile second data store, data from the storage locations identified in the index from the last snapshot; and
storing the retrieved data in the cache, wherein the cache is separate from the first and second data stores.
2. A method according to claim 1 , wherein the first data store and the second data store are the same data store.
3. A method according to claim 1 , wherein the first data store and the second data store are different data stores.
4. A method according to claim 1 , wherein at least one of the first data store or the second data store includes a set of block storage devices.
5. A method according to claim 1 , wherein the index identifies physical storage locations.
6. A method according to claim 1 , wherein the index identifies virtual storage addresses.
7. A method according to claim 6 , wherein retrieving data from storage locations identified in the index comprises:
translating a virtual storage address to a physical storage address for the second data store.
8. A method according to claim 1 , wherein the cache includes an application program interface, and wherein the cache provides the index via the application program interface.
9. A method according to claim 8 , wherein retrieving data from storage locations identified in the index comprises:
providing the retrieved index to the cache via the application program interface; and
retrieving the data by the cache.
10. A method according to claim 9 , wherein the data storage system includes a cache miss handler, and wherein retrieving the data by the cache includes making calls to the cache miss handler by the cache.
11. A data storage system comprising:
a storage processor configured to manage storage of data in at least one non-volatile data store; and
a volatile memory cache that includes content corresponding to a set of data blocks in a non-volatile second data store, wherein the non-volatile second data store is disk storage or solid state storage, wherein:
the storage processor is configured to periodically store, in a non-volatile first data store, a snapshot of an index identifying storage locations of the set of data blocks, wherein such periodically storing of the snapshot is performed prior to a restart of the data storage system, and, upon a restart of the data storage system causing loss or corruption of the contents of the volatile memory cache, retrieve the index from the last snapshot stored prior to the restart, the last snapshot identifying content of the cache at a time of storing the last snapshot; and
at least one of the storage processor and the cache is configured to retrieve, from the non-volatile second data store, data from the storage locations identified in the index from the last snapshot and store the retrieved data in the cache, data in the non-volatile second data store is persistently stored and available for retrieval upon the restart of the data storage system, wherein the cache is separate from the first and second data stores.
12. A system according to claim 11 , wherein the first data store and the second data store are the same data store.
13. A system according to claim 11 , wherein the first data store and the second data store are different data stores.
14. A system according to claim 11 , wherein at least one of the first data store or the second data store includes a set of block storage devices.
15. A system according to claim 11 , wherein the index identifies physical storage locations.
16. A system according to claim 11 , wherein the index identifies virtual storage addresses.
17. A system according to claim 16 , wherein retrieving data from storage locations identified in the index comprises:
translating a virtual storage address to a physical storage address for the second data store.
18. A system according to claim 11 , wherein:
the cache includes an application program interface;
the cache provides the index to the storage processor via the application program interface; and
the storage processor stores the snapshot of the index in the second data store.
19. A system according to claim 18 , wherein retrieving data from storage locations identified in the index comprises:
retrieving the index from the second data store by the storage processor;
providing the retrieved index by the storage processor to the cache via the application program interface; and
retrieving the data by the cache.
20. A system according to claim 19 , wherein the storage processor includes a cache miss handler, and wherein retrieving the data by the cache includes making calls to the cache miss handler by the cache.