Method and apparatus for defragmentation and for detection of relocated blocks

A method of operating a storage system, the method comprising:

• maintaining a first data structure and a second data structure, each defining an organization of stored data, the first and second data structure each including a first pointer to a particular stored block, the first data structure further including a second pointer associated with the particular stored block;
• moving the block to a new location;
• and
• in response to a need to access the block,
• using the first pointer in the first data structure to attempt to locate the block,
• determining that the block may have been moved, and in response to such determination,
• using the second pointer to reference into the second data structure, to locate the block at the new location.

A method of operating a storage system, the method comprising:

• maintaining a first data structure and a second data structure, each defining an organization of a plurality of blocks stored in an aggregate of storage, the first and second data structures each including a first pointer to a particular block, the first data structure further including a second pointer associated with the particular block, the first pointer representing a physical address of the block in an aggregate of storage, the second pointer representing an address of the block in a virtual data container defined within the aggregate, the first data structure associating the first pointer with the second pointer;
• moving the block to a new location in the aggregate as part of a defragmentation process, including updating the first pointer in the second data structure without requiring any update of the first pointer in the first data structure;
• and
• in response to a request to access the block after the block has been moved to the new location,
• determining that the first pointer is incorrect in the first data structure,
• using the second pointer to reference into the second data structure to locate a correct value of the first pointer in the second data structure, the correct value reflecting the new location of the block, and
• accessing the block using the correct value of the first pointer.

A storage server comprising:

• a communications interface through which to receive a data access request from a client;
• a network interface through which to access a mass storage subsystem on behalf of the client in response to the data access request;
• a file system manager to maintain a first data structure and a second data structure, each defining an organization of a plurality of blocks stored in an aggregate of storage, the first and second data structures each including a first pointer to a particular block which corresponds to the data access request, the first data structure further including a second pointer associated with the block and an association between the first and second pointers;
• a write allocator to move the block to a new location, including updating the first pointer in the second data structure, without requiring any update of the first pointer in the first data structure;
• and
• a read handler to respond to the data access request by accessing the block, after the block has been moved to the new location, by
• determining that the first pointer is incorrect in the first data structure,
• using the second pointer and the second data structure to locate a correct value of the first pointer, the correct value reflecting the new location of the block, and
• accessing the block using the correct value of the first pointer.