Asynchronous movement of in-line metadata for cached volumes at storage gateways
Methods and apparatus for supporting cached volumes at storage gateways are disclosed. A storage gateway appliance is configured to cache at least a portion of a storage object of a remote storage service at local storage devices. In response to a client's write request, directed to at least a portion of a data chunk of the storage object, the appliance stores a data modification indicated in the write request at a storage device, and asynchronously uploads the modification to the storage service. In response to a client's read request, directed to a different portion of the data chunk, the appliance downloads the requested data from the storage service to the storage device, and provides the requested data to the client.
1. A system, comprising:
one or more computing devices comprising one or more hardware processors and one or more memories storing program instructions that, when executed by at least one of the one or more hardware processors, configure the at least one of the one or more computing devices to:
cache, by a storage appliance, at least a portion of one or more data chunks of a storage object;
store, by the storage appliance, inline metadata for the one or more data chunks;
store, by the storage appliance, contiguous metadata for the one more data chunks;
in response to transfer of a new data block from a remote storage service for a particular data chunk of the one or more data chunks, synchronously update the inline metadata without synchronously updating the contiguous metadata;
perform an asynchronous update to the contiguous metadata with respect to the transfer of the new data block and the synchronous update of the inline metadata, wherein the asynchronous update causes one or more changes to the contiguous metadata stored on the storage appliance; and
access the contiguous metadata to determine a state of at least a portion of the particular data chunk without reading all the inline metadata for the particular data chunk.
2. The system recited in claim 1 , wherein the contiguous metadata indicates offset and state information for the one or more data chunks of the storage object.
3. The system recited in claim 1 , wherein the one or more computing devices are further configured to determine, using the inline metadata associated with the one or more data chunks, whether an I/O request directed to at least a portion of the one or more data chunks results in a cache miss.
4. The system recited in claim 3 ,
wherein the one or more data chunks and a corresponding inline metadata for the one or more data chunks are stored together in an alternating sequence of cached data blocks of the particular data chunk and inline metadata sections corresponding to the cached data blocks, and
wherein the one or more computing devices are further configured to update a particular portion of the contiguous metadata based at least in part on a modification of at least one element of inline metadata, wherein said updating the particular portion of the contiguous metadata is performed asynchronously with respect to the modification of the at least one element of inline metadata.
5. The system recited in claim 3 , wherein to modify the at least one element of inline metadata, the one or more computing devices are further configured to:
identify a first inline metadata section adjacent to a particular data block within the particular data chunk, and a different inline metadata section adjacent to the particular data block within the particular data chunk; and
modify the first inline metadata section and the different inline metadata section prior to providing a response to the I/O request.
6. The system recited in claim 5 , wherein to modify the at least one elements of inline metadata, the one or more computing devices are further configured to identify an offset at which the at least one element of inline metadata is stored, based at least in part on an offset of the particular data block within the particular data chunk.
7. A method, comprising:
caching, by a storage appliance, at least a portion of one or more data chunks of a storage object;
storing, by the storage appliance, inline metadata for the one or more data chunks;
storing, by the storage appliance, contiguous metadata for the one more data chunks;
subsequent to transfer of a new data block from a remote storage service for a particular data chunk of the one or more data chunks, synchronously updating the inline metadata without synchronously updating the contiguous metadata;
asynchronously updating the contiguous metadata with respect to the transfer of the new data block and the synchronous updating of the inline metadata, wherein asynchronously updating causes one or more changes to the contiguous metadata stored on the storage appliance; and
accessing the contiguous metadata to determine a state of at least a portion of the particular data chunk without reading all the inline metadata for the particular data chunk.
8. The method recited in claim 7 , wherein the contiguous metadata indicates offset and state information for the one or more data chunks of the storage object.
9. The method recited in claim 7 , further comprising:
receiving a read request directed to a particular data block of the particular data chunk;
downloading at least a portion of the particular data chunk, comprising the particular data block, from the remote storage service;
determining whether the at least a portion of the particular data chunk is to be merged with one or more cached data blocks;
subsequent to determining that a merge is to be performed, merging the at least a portion of the particular data chunk with the one or more cached data blocks prior to storing the at least a portion of the particular data chunk; and
extracting the particular data block from the at least a portion of the particular data chunk and providing the particular data block in response to the read request.
10. The method recited in claim 7 , further comprising:
receiving an I/O request comprising a modification directed to a particular data block of the particular data chunk;
modifying the particular data block in accordance with the I/O request;
adding an entry to an upload buffer configured for asynchronous uploads of modified data to the remote storage service, wherein the entry indicates modifications performed on the particular data block in response to the I/O request; and
verifying that modifications indicated in the entry have been uploaded to the remote storage service prior to replacing the particular data block.
11. The method recited in claim 10 ,
wherein the one or more data chunks and a corresponding inline metadata for the one or more data chunks are stored together in an alternating sequence of cached data blocks of the particular data chunk and inline metadata sections corresponding to the cached data blocks, and
wherein the method further comprises updating a particular portion of the contiguous metadata based at least in part on modifying at least one element of inline metadata, wherein said updating the particular portion of the contiguous metadata is performed asynchronously with respect to the modification of the at least one element of inline metadata.
12. The method recited in claim 11 , wherein modifying the at least one element of inline metadata includes identifying an offset at which the at least a portion of inline metadata is stored, based at least in part on an offset of the particular data block within the particular data chunk.
13. The method recited in claim 12 , wherein modifying the at least one element of inline metadata includes:
identifying a first inline metadata section adjacent to the particular data block, and a different inline metadata section adjacent to the particular data block; and
modifying the first inline metadata section and the different inline metadata section prior to providing a response to the I/O request.
14. A non-transitory computer-readable storage medium storing program instructions that when executed by one or more computer processors:
cache at least a portion of one or more data chunks of a storage object;
store inline metadata for the one or more data chunks;
store contiguous metadata for the one more data chunks;
subsequent to transfer of a new data block from a remote storage service for a particular data chunk of the one or more data chunks, synchronously update the inline metadata without a synchronous update of the contiguous metadata;
asynchronously update the contiguous metadata with respect to the transfer of the new data block and the synchronous update of the inline metadata to cause one or more changes to the contiguous metadata; and
access the contiguous metadata to determine a state of at least a portion of the particular data chunk without reading all the inline metadata for the particular data chunk.
15. The non-transitory computer-readable storage medium of claim 14 , wherein the contiguous metadata indicates offset and state information for the one or more data chunks of the storage object.
16. The non-transitory computer-readable storage medium of claim 14 , wherein the instructions when executed by one or more computer processors:
receive a read request directed to a particular data block of the particular data chunk;
download at least a portion of the particular data chunk, comprising the particular data block, from the remote storage service;
determine whether the at least a portion of the particular data chunk is to be merged with one or more cached data blocks;
subsequent to a determination that a merge is to be performed, merge the at least a portion of the particular data chunk with the one or more cached data blocks prior to storage of the at least a portion of the particular data chunk; and
extract the particular data block from the at least a portion of the particular data chunk and provide the particular data block in response to the read request.
17. The non-transitory computer-readable storage medium of claim 14 , wherein the instructions when executed by one or more computer processors:
receive an I/O request that include a modification directed to a particular data block of the particular data chunk;
modify the particular data block in accordance with the I/O request;
add an entry to an upload buffer configured for asynchronous uploads of modified data to the remote storage service, wherein the entry indicates modifications performed on the particular data block in response to the I/O request; and
verify that modifications indicated in the entry have been uploaded to the remote storage service prior to replacement of the particular data block.
18. The non-transitory computer-readable storage medium of claim 17 ,
wherein to cache the at least the portion of the one or more data chunks of the storage object and to store the inline metadata for the one or more data chunks, the instructions when executed by one or more computer processors store the at least the portion of the one or more data chunks and a corresponding inline metadata for the at least the portion of the one or more data chunks together in an alternating sequence of cached data blocks of the particular data chunk and inline metadata sections corresponding to the cached data blocks, and
wherein instructions when executed by one or more computer processors update a particular portion of the contiguous metadata based at least in part on modification of at least one element of inline metadata, wherein said update the particular portion of the contiguous metadata is performed asynchronously with respect to the modification of the at least one element of inline metadata.
19. The non-transitory computer-readable storage medium of claim 18 , wherein to perform said modification of the at least one element of inline metadata the instructions when executed by one or more computer processors identify an offset at which the at least a portion of inline metadata is stored, based at least in part on an offset of the particular data block within the particular data chunk.
20. The non-transitory computer-readable storage medium of claim 19 , wherein to perform said modification of the at least one element of inline metadata the instructions when executed by one or more computer processors:
identify a first inline metadata section adjacent to the particular data block, and a different inline metadata section adjacent to the particular data block; and
modify the first inline metadata section and the different inline metadata section prior to providing a response to the I/O request.