Metadata-based data copying
A technique for copying data from a source range to a destination range includes identifying a chunk of the source range. The chunk includes multiple logical blocks that map, through a series of source-mapping pages, to a set of virtual pages that point to physical data of the chunk. The technique further includes copying the series of source-mapping pages to form a corresponding series of destination-mapping pages pointed to by the destination range, so that a corresponding chunk of the destination range maps, via the series of destination-mapping pages, to the same set of virtual pages mapped to by the series of source mapping pages.
1. A method of copying data, comprising:
receiving a request to copy a set of data from a source range to a destination range;
identifying a chunk of the source range, the chunk including multiple logical blocks that are mapped, through a series of source-mapping pages, to a set of virtual pages that point to physical data of the chunk;
copying the series of source-mapping pages to provide a corresponding series of destination-mapping pages pointed to by the destination range, the corresponding series of destination-mapping pages mapping multiple logical blocks in a corresponding chunk of the destination range to the set of virtual pages, the destination range thereby pointing to the physical data of the chunk of the source range and providing a copy thereof.
2. The method of claim 1 , wherein the source range and the destination range are both located within a logical address space of a storage system.
3. The method of claim 2 , wherein the series of source-mapping pages includes a source-leaf page that points to the set of virtual pages, the source-leaf page including pointers that map every one of the logical blocks of the chunk.
4. The method of claim 2 , wherein the method further comprises, prior to copying the series of source-mapping pages:
accessing the series of destination-mapping pages with a write-lock; and
allocating a new destination-mapping page for each of the series of destination-mapping pages.
5. The method of claim 4 , wherein at least one of the new destination-mapping pages does not replace any previously allocated destination-mapping page.
6. The method of claim 4 , wherein at least one of the new destination-mapping pages replaces an old destination mapping page that already exists.
7. The method of claim 6 , wherein a parent-child relationship exists between the old destination-mapping page and one of the series of source-mapping pages, and wherein allocating the new destination-mapping page causes the new destination mapping page to have no parent-child relationship with any of the series of source-mapping pages.
8. The method of claim 4 , further comprising accessing the series of source-mapping pages for reading using a try-lock, the try-lock configured to return a failure result if the try-lock cannot be obtained.
9. The method of claim 8 , wherein the method further comprises, responsive to the try-lock returning the failure result:
committing, in a transaction log, a transaction that involves at least one of the series of source-mapping pages; and
after committing the transaction, accessing the series of source-mapping pages using a read-lock.
10. The method of claim 9 , wherein copying the series of source-mapping pages to the corresponding series of destination-mapping pages includes copying mapping pointers from the series of source-mapping pages to the new destination-mapping pages.
11. A computerized apparatus, comprising control circuitry that includes a set of processors coupled to memory, the control circuitry constructed and arranged to:
receive a request to copy a set of data from a source range to a destination range;
identify a chunk of the source range, the chunk including multiple logical blocks that are mapped, through a series of source-mapping pages, to a set of virtual pages that point to physical data of the chunk;
copy the series of source-mapping pages to provide a corresponding series of destination-mapping pages pointed to by the destination range, the corresponding series of destination-mapping pages mapping multiple logical blocks in a corresponding chunk of the destination range to the set of virtual pages, the destination range thereby pointing to the physical data of the chunk of the source range and providing a copy thereof.
12. A computer program product including a set of non-transitory, computer-readable media having instructions which, when executed by control circuitry of a computerized apparatus, cause the computerized apparatus to perform a method of copying data, the method comprising:
receiving a request to copy a set of data from a source range to a destination range;
identifying a chunk of the source range, the chunk including multiple logical blocks that are mapped, through a series of source-mapping pages, to a set of virtual pages that point to physical data of the chunk;
copying the series of source-mapping pages to provide a corresponding series of destination-mapping pages pointed to by the destination range, the corresponding series of destination-mapping pages mapping multiple logical blocks in a corresponding chunk of the destination range to the set of virtual pages, the destination range thereby pointing to the physical data of the chunk of the source range and providing a copy thereof.
13. The computer program product of claim 12 , wherein the source range and the destination range are both located within a logical address space of a storage system.
14. The computer program product of claim 13 , wherein the series of source-mapping pages includes a source-leaf page that points to the set of virtual pages, the source-leaf page including pointers that map every one of the logical blocks of the chunk.
15. The computer program product of claim 13 , wherein the method further comprises, prior to copying the series of source-mapping pages:
accessing the series of destination-mapping pages with a write-lock; and
allocating a new destination-mapping page for each of the series of destination-mapping pages.
16. The computer program product of claim 15 , wherein at least one of the new destination-mapping pages does not replace any previously allocated destination-mapping page.
17. The computer program product of claim 15 , wherein at least one of the new destination-mapping pages replaces an old destination mapping page that already exists.
18. The computer program product of claim 17 , wherein a parent-child relationship exists between the old destination-mapping page and one of the series of source-mapping pages, and wherein allocating the new destination-mapping page causes the new destination mapping page to have no parent-child relationship with any of the series of source-mapping pages.
19. The computer program product of claim 15 , further comprising accessing the series of source-mapping pages for reading using a try-lock, the try-lock configured to return a failure result if the try-lock cannot be obtained.
20. The computer program product of claim 19 , wherein the method further comprises, responsive to the try-lock returning the failure result:
committing, in a transaction log, a transaction that involves at least one of the series of source-mapping pages; and
after committing the transaction, accessing the series of source-mapping pages using a read-lock.
21. The method of claim 1 , wherein copying the series of source-mapping pages to the corresponding series of destination-mapping pages includes adjusting pointers of at least one of the destination-mapping pages for offsets so that relative pointing between the series of destination-mapping pages is analogous to relative pointing between the series of source-mapping pages.