Method and apparatus for avoiding redundant data recovery
A method for use in a storage system, comprising: detecting that at least one storage device in a redundant storage device array has failed, the redundant storage device array including a plurality of storage devices; storing, in a memory, a recovery cache containing recovered data blocks, the recovered data blocks containing data previously stored on the failed storage device that has been calculated as part of executing read commands before the redundant storage device array is rebuilt; receiving an I/O command that is associated with the redundant storage device array, the I/O command including one of a read command or a write command; and executing the I/O command by using the recovery cache, wherein the redundant storage device array includes a plurality of storage regions, wherein any of the storage regions includes a respective portion of each of the plurality of storage devices.
1. A method for use in a storage system, comprising:
detecting that at least one storage device in a redundant storage device array has failed, the redundant storage device array including a plurality of storage devices;
storing, in a memory, a recovery cache containing recovered data blocks, the recovered data blocks containing data previously stored on the failed storage device that has been calculated as part of executing read commands before the redundant storage device array is rebuilt;
receiving an I/O command that is associated with the redundant storage device array, the I/O command including one of a read command or a write command;
executing the I/O command by using the recovery cache; and
rebuilding the redundant storage device array by using the recovery cache, the rebuilding including; identifying one or more data blocks that are stored on the failed storage device performing one or more searches of the recovery cache to detect whether any of the identified data blocks has already been recovered; if not all of the identified data blocks have been recovered yet, recovering, based on parity data, each of the identified data blocks that has not been recovered yet, and copying the identified data blocks to the plurality of storage devices that remain operational,
wherein the redundant storage device array includes a plurality of storage regions, and
wherein any of the storage regions includes a respective portion of each of the plurality of storage devices.
2. The method of claim 1 , wherein the redundant storage device array includes a Redundant Array of Independent Disks 6 (RAID 6) array, and each of the plurality of storage regions includes a RAID 6 stripe.
3. The method of claim 1 , wherein the recovery cache is used to ensure that any data block that is stored on the failed storage device is recovered at most once before the redundant storage device array is rebuilt.
4. The method of claim 3 , wherein;
rebuilding the redundant storage device array further includes selecting one of the plurality of storage regions;
identifying the one or more data blocks includes identifying the one or more data blocks in the selected storage region; and
the parity data is stored in the selected storage region.
5. The method of claim 1 , wherein the I/O command includes a read command, and executing the I/O command includes:
detecting whether a given data block associated with the read command is stored on the failed storage device or an operational storage device;
when the given data block is stored on an operational storage device, retrieving the given data block from the operational storage device;
when the given data block is stored on the failed storage device, detecting whether the given data block has already been recovered by performing a search of the recovery cache;
when the given data block has already been recovered, retrieving the given data block from the recovery cache; and
when the given data block has not been recovered yet, recovering the given data block, and storing the given data block in the recovery cache.
6. A method for use in a storage system, comprising;
detecting that at least one storage device in a redundant storage device array has failed, the redundant storage device array including a plurality of storage devices;
storing in a memory, a recovery cache containing data blocks, the recovered data blocks containing data previously stored on the failed storage device that has been calculated as part of executing read commands before the redundant storage device array is rebuilt;
receiving and I/O command that is associated with the redundant storage device array, the I/O command including one of a read command or a write command; and
executing the I/O command by using the recovery cache,
wherein the redundant storage device array incudes a plurality of storage regions,
wherein any of the storage regions includes a respective portion of each of the plurality of storage devices, and
wherein the I/O command includes a write command, and executing the I/O command includes: when a type-1 storage region is available in the redundant storage device array to receive a data block associated with the write command, storing the data block in the type-1 storage region, the type-1 storage region being a storage region that contains no blocks that need to be recovered; and when a type-1 storage region is not available in the redundant storage device array to receive the data block and a type-2 storage region is available to receive the data block, storing the data block in the type-2 storage region, the type-2 storage region being a storage region that contains no blocks that have been recovered and stored in the recovery cache.
7. The method of claim 6 , wherein, when neither one of a type-1 storage region nor a type-2 storage region is available to receive the data block, executing the I/O command includes at least one of (i) allowing the write command to time out, (ii) generating an error indicating that the write command cannot be completed successfully, or (iii) storing the data block in a temporary storage location and generating an acknowledgement indicating that the write command has been completed successfully.
8. A system, comprising:
at least one processor that is configured to perform the operations of:
detecting that at least one storage device in a redundant storage device array has failed, the redundant storage device array including a plurality of storage devices;
storing, in a memory, a recovery cache containing recovered data blocks, the recovered data blocks containing data previously stored on the failed storage device that has been calculated as part of executing read commands before the redundant storage device array is rebuilt;
receiving an I/O command that is associated with the redundant storage device array, the I/O command including one of a read command or a write command;
executing the I/O command by using the recovery cache; and
rebuilding the redundant storage device array by using the recovery cache, the rebuilding including, identifying one or more data blocks that are stored on the failed storage device, performing one or more searches of the recovery cache to detect whether any of the identified data blocks has already been recovered; if not all of the identified data blocks have been recovered yet, recovering, based on parity data, each of the identified data blocks that has not been recovered yet, and copying the identified data blocks to the plurality of storage devices that remain operational,
wherein the redundant storage device array includes a plurality of storage regions,
wherein any of the storage regions includes a respective portion of each of the plurality of storage devices.
9. The system of claim 8 , wherein the redundant storage device array includes a Redundant Array of Independent Disks 6 (RAID 6) array, and each of the plurality of storage regions includes a RAID 6 stripe.
10. The system of claim 8 , wherein the recovery cache is used to ensure that any data block that is stored on the failed storage device is recovered at most once before the redundant storage device array is rebuilt.
11. The system of claim 10 , wherein;
rebuilding the redundant storage device array further includes selecting one of the plurality of storage regions;
identifying the one or more data blocks includes identifying the one or more data blocks in the selected storage region; and
the parity data is stored in the selected storage region.
12. The system of claim 8 , wherein the I/O command includes a read command, and executing the I/O command includes:
detecting whether a given data block associated with the read command is stored on the failed storage device or an operational storage device;
when the given data block is stored on an operational storage device, retrieving the given data block from the operational storage device;
when the given data block is stored on the failed storage device, detecting whether the given data block has already been recovered by performing a search of the recovery cache;
when the given data block has already been recovered, retrieving the given data block from the recovery cache; and
when the given data block has not been recovered yet, recovering the given data block, and storing the given data block in the recovery cache.
13. A system comprising;
at least one processor that is configured to perform the operation of,
detecting that at least one storage device in a redundant storage device array has failed, the redundant storage device array including a plurality of storage devices;
storing, in a memory, a recovery cache containing recovered data blocks, the recovered data blocks containing data previously stored on the failed storage device that has been calculated as part of executing read commands before the redundant storage device array is rebuilt;
receiving an I/O command that is associated with the redundant storage device array, the I/O command including one of a read command or a write command, and
executing the I/O command by using the recovery cache,
wherein the redundant storage device array includes a plurality of storage regions,
wherein any of the storage regions includes a respective portion of each of the plurality of storage devices
wherein the I/O command includes a write command, and executing the I/O command includes: when a type-1 storage region is available in the redundant storage device array to receive a data block associated with the write command, storing the data block in the type-1 storage region, the type-1 storage region being a storage region that contains no blocks that need to be recovered; and when a type-1 storage region is not available in the redundant storage device array to receive the data block and a type-2 storage region is available to receive the data block, storing the data block in the type-2 storage region, the type-2 storage region being a storage region that contains no blocks that have been recovered and stored in the recovery cache.
14. The system of claim 13 , wherein, when neither one of a type-1 storage region nor a type-2 storage region is available to receive the data block, executing the I/O command includes at least one of (i) allowing the write command to time out, (ii) generating an error indicating that the write command cannot be completed successfully, or (iii) storing the data block in a temporary storage location and generating an acknowledgement indicating that the write command has been completed successfully.