Cache based recovery of corrupted or missing data

1. In a system that includes a storage device storing data and a flash memory configured as a cache for the storage device, a method for recovering data on the storage device that has become corrupted without accessing a backup of the data stored on the storage device, the method comprising:

discovering, with a background process tasked with scanning the storage device, that data that was previously stored on the storage device has become corrupted or is missing, wherein the background process performs a background scan on the storage device;

performing a background scan of a cache to determine whether data in the cache is corrupted or missing, wherein corrupted data or missing data in the cache is invalidated or evicted from the cache;

determining an extent of the corrupted or missing data on the storage device in an area near the discovered data that is corrupted or missing;

searching the cache for a valid copy of the corrupted data or the missing data, wherein the cache is accessible to clients such that the clients can perform read and write operations in the cache and wherein the cache stores data read from the storage device;

searching one or more additional caches associated with one or more additional devices for the valid copy when the valid copy is not found in the cache; and

recovering the valid copy of the corrupted data or the missing data to the storage device when the valid copy is present in the cache or in the one or more additional caches.

2. The method of claim 1 , further comprising preventing the cache from evicting data when the corrupted data or the missing data is discovered.

3. The method of claim 1 , further comprising discovering the corrupted data or the missing data when accessing the storage device.

4. The method of claim 3 , wherein accessing the storage device includes performing a read operation on the corrupted data or the missing data.

5. The method of claim 1 , further comprising recovering the valid copy from a backup of the data or by reconstructing the data when the valid copy is not present in the cache.

6. The method of claim 1 , wherein searching a cache for a valid copy of the corrupted data includes determining a key of the corrupted data or of the missing data and accessing the cache using the key.

7. In a system that includes a storage system and a flash memory configured as a cache for the storage system, a method for recovering data in the storage system that has become corrupted, the method comprising:

identifying data that was previously stored in the storage system that has become corrupted data or is missing data in the storage system with a background process tasked with scanning the storage system, wherein the background process performs a background scan on the storage system, wherein the corrupted data or the missing data is associated with a location in the storage system;

performing a background scan of a cache to determine whether data in the cache is corrupted or missing, wherein corrupted data or missing data in the cache is invalidated or evicted from the cache;

determining an extent of the corrupted or missing data on the storage system in an area near the discovered data that is corrupted or missing;

examining an auxiliary data structure to identify a key associated with the corrupted data or the missing data;

searching the cache using the key to determine if a valid copy of the corrupted data or of the missing data is present in the cache, wherein the cache is accessible to clients such that the clients can perform read and write operations to the cache and wherein the cache stores data read from the storage system;

searching one or more additional caches associated with one or more additional devices using the key for the valid copy when the valid copy is not found in the cache; and

recovering the valid copy to the storage system when the valid copy is present in the cache or in the one or more additional caches.

8. The method of claim 7 , wherein identifying corrupted data or missing data in the storage system includes one or more of:

performing a read operation, wherein the corrupted data or the missing data is identified during the read operation; and

performing a replication operation, wherein the corrupted data or the missing data is identified when an attempt is made to replicate the corrupted data.

9. The method of claim 7 , wherein the auxiliary data structure establishes a relationship between the location and the key, further comprising walking the auxiliary data structure using the location to determine the key of the corrupted data or of the missing data.

10. The method of claim 7 , wherein the auxiliary data structure is accessed to determine all keys associated with additional corrupted data or additional missing data near or adjacent the corrupted data or the missing data.

11. The method of claim 7 , wherein the key comprises at least one of an inode, a fingerprint, a hash, a file and offset, a logical unit number, or an object identifier.

12. The method of claim 7 , further comprising returning the valid copy of the corrupted data or the missing data to a requesting client.

13. The method of claim 1 , further comprising preventing the cache from evicting data until the cache has been searched.

14. A non-transitory computer readable medium having stored thereon computer executable instructions that, when executed, perform a method for recovering data in a storage system that has become corrupted without accessing a backup of the data, the method comprising:

storing data in the storage system;

identifying data previously stored in the storage system that has become corrupted when an attempt to read the corrupted data is performed or when the corrupted data s discovered by a background process tasked with scanning the storage system for data that has become corrupted or is missing, wherein the background process performs a background scan on the storage device;

performing a background scan of a cache to determine whether data in the cache is corrupted or missing, wherein corrupted data or missing data in the cache is invalidated or evicted from the cache;

determining an extent of the corrupted data on the storage device in an area near the discovered data that is corrupted;

searching a cache associated with the storage system for a valid copy of the corrupted data, wherein the cache is accessible to clients such that the clients can perform read and write operations to the cache and wherein the cache stores data read from the storage system;

searching one or more additional caches associated with one or more additional devices for the valid copy when the valid copy is not found in the cache; and

recovering the valid copy to the storage system.

15. The computer readable medium of claim 14 , wherein searching a cache further comprises determining at least one key for the corrupted data and for other corrupted data that is near or adjacent the corrupted data, if other corrupted data is present in the storage system, based on a location of the corrupted data and a location of the other corrupted data, wherein the cache is accessed using the at least one key.

16. The computer readable medium of claim 14 , wherein the at least one key comprises one or more of inode, a fingerprint, a hash, a file and offset, a logical unit number, or an object identifier.