Thin provisioning row snapshot with reference count map
View Patent ↗The storage of single or multiple references of the same data block in a storage pool is disclosed. Indexing of the data includes storing reference information in the storage pool as a mapping table. The mapping table indexes each data block in the storage pool. On any read or write request mapping information is used to retrieve the corresponding data block in storage pool.
1. A method for scalable indexing of a storage pool formed from multiple storage units in which backup information is recorded, comprising:
forming a virtual volume in which the combined storage capacity of the storage pool is represented by a set of chunks have a pre-selected data size with the chunks being assigned to be either storage chunks or reference chunks;
storing in a set of reference chunks metadata information to index the storage chunks and track usage of storage chunks, wherein storage chunks are allocated on demand and are referable more than once to reduce redundant storage for identical content;
adjusting the total number of chunks and a chunk allocation in response to a change in the size of the available physical storage to include a sufficient number of reference chunks to index all of the storage chunks;
wherein the reference chunks are updatable to store indexing information for redirect on write snapshots of the virtual volume; and
wherein the set of reference chunks stores:
a reference count map maintained as a set of contiguous storage pool chunks, the reference count map referencing all the storage chunks in a virtual volume including usage information on storage chunks used more than once including a usage count of each storage chunk, the usage count tracking free and allocated storage chunks and the number of times a storage chunk is referenced; and
a reference count of chunks referred to by any redirect on write snapshots of the virtual volume.
2. The method of claim 1 , wherein the reference chunks store a representation of a radix tree to index each storage chunk and the usage of each chunk.
3. The method of claim 1 , wherein the set of reference chunks further stores:
a reference count of chunks referred to by the virtual volume; and
a reference count of chunks referred to by clones of the virtual volume.
4. The method of claim 1 , wherein the set of contiguous storage pool chunks are disposed at one end of the storage pool or on an external storage device.
5. The method of claim 1 , wherein the reference count map is cached for faster access.
6. The method of claim 1 , wherein the reference count map provides index information for allocating a group of storage pool chunks which can be used by any number of virtual volumes.
7. The method of claim 1 , further comprising maintaining a circular buffer of pre-allocated chunks and storing the circular buffer onto the persistent storage.
8. The method of claim 7 , further comprising reading the circular buffer, the circular buffer saving the chunk allocations for each entry in the circular buffer when it was stored on persistent storage.
9. The method of claim 1 , wherein the maximum amount of potential storage space allocated to users is greater than the physical storage capacity of the storage pool and storage space is allocated on demand based on actual usage.
10. A method for scalable indexing of a storage pool formed from multiple storage units in which backup information is recorded, comprising:
forming a virtual volume in which the combined storage capacity of the storage pool is represented by a set of chunks have a pre-selected data size with the chunks being assigned to be either storage chunks or reference chunks;
storing in a set of reference chunks metadata information to index the storage chunks and track usage of storage chunks, wherein storage chunks are allocated on demand and are referable more than once to reduce redundant storage for identical content; and
adjusting the total number of chunks and a chunk allocation in response to a change in the size of the available physical storage to include a sufficient number of reference chunks to index all of the storage chunks, wherein the set of reference chunks stores a reference count map referencing all of the storage pool chunks in the storage pool, the reference chunks including a usage count of each storage chunk, the usage count tracking free and allocated storage chunks and wherein the reference count map assigns a set of bits in each reference chunk as a bitset to store a usage count of each storage chunk associated with a particular reference chunk to indicate when an individual storage chunk is allocated and referred to by more than one volume or snapshot.
11. The method of claim 10 , wherein a usage count is maintained for each storage chunk, the method further comprising:
incrementing the usage count of a chunk on snapshot creation;
decrementing the usage count of a chunk on snapshot or virtual volume deletion; and
decrementing the usage count of a shared chunk on update.
12. A method for scalable indexing of a storage pool formed from multiple storage units in which backup information is recorded, comprising:
forming a virtual volume in which the combined storage capacity of the storage pool is represented by a set of chunks have a pre-selected data size with the chunks being assigned to be either storage chunks or reference chunks;
storing in a set of reference chunks metadata information to index the storage chunks and track usage of storage chunks, wherein storage chunks are allocated on demand and are referable more than once to reduce redundant storage for identical content; and
adjusting the total number of chunks and a chunk allocation in response to a change in the size of the available physical storage to include a sufficient number of reference chunks to index all of the storage chunks, wherein the set of reference chunks stores a reference count map referencing all of the storage pool chunks in the storage pool, the reference chunks including a usage count of each storage chunk, the usage count tracking free and allocated storage chunks and wherein each reference count chunk in the reference map indexes a set of storage chunks, the method further comprising:
in response to allocation of a chunk, scanning a reference count map bitset in a first reference count chunk to identify the address of a freely available chunk in the storage pool;
responsive to determining that the scanned reference count chunk does not have any free bitset available, sequentially scanning subsequent reference count chunks to identify the first freely available bitset and marking the bitset to identify a new chunk allocation.
13. A method for scalable indexing of a storage pool formed from multiple storage units in which backup information is recorded, comprising:
forming a virtual volume in which the combined storage capacity of the storage pool is represented by a set of chunks have a pre-selected data size with the chunks being assigned to be either storage chunks or reference chunks;
storing in a set of reference chunks metadata information to index the storage chunks and track usage of storage chunks, wherein storage chunks are allocated on demand and are referable more than once to reduce redundant storage for identical content;
adjusting the total number of chunks and a chunk allocation in response to a change in the size of the available physical storage to include a sufficient number of reference chunks to index all of the storage chunks; and
logging a reference count change in a logging string, comprising:
an address of a reference count chunk;
an offset of a storage pool chunk within a reference count chunk;
a previous reference count; and
a new reference count.
14. A method for scalable indexing of a storage pool formed from multiple storage units in which backup information is recorded, comprising:
forming a virtual volume in which the combined storage capacity of the storage pool is represented by a set of chunks have a pre-selected data size with the chunks being assigned to be either storage chunks or reference chunks;
storing in a set of reference chunks metadata information to index the storage chunks and track usage of storage chunks, wherein storage chunks are allocated on demand and are referable more than once to reduce redundant storage for identical content; and
wherein the reference chunks having a sequential order and the reference chunks comprise a reference count map including:
a reference count of chunks referred to by the virtual volume;
a reference count of chunks referred to by clones of the virtual volume; and
a reference count of chunks referred to by any redirect on write snapshots of the virtual volume.
15. The method of claim 14 , wherein the reference count map further comprises a usage count of chunks referred to more than once.
16. The method of claim 14 , further comprising:
adjusting the total number of chunks and a chunk allocation in response to a change in the size of the available physical storage to include a sufficient number of reference chunks to index all of the storage chunks.
17. A storage server to provide thin provisioning of storage, comprising:
at least one interface for the storage server to access storage devices;
at least one processor and a memory having computer program code which when executed on the processor is operative to perform scalable indexing of a storage pool formed from multiple storage units in which backup information is recorded, including:
generating a virtual volume in which the combined storage capacity of the storage pool is represented by a set of chunks have a pre-selected data size with the chunks being assigned to be either storage chunks or reference chunks;
storing in a set of reference chunks metadata information to index the storage chunks and track usage of storage chunks, wherein storage chunks are allocated on demand and are referable more than once to reduce redundant storage for identical content; and
adjusting the total number of chunks and a chunk allocation in response to a change in the size of the available physical storage to include a sufficient number of reference chunks to index all of the storage chunks;
wherein the set of reference chunks stores a reference count map referencing all of the storage pool chunks in the storage pool, the reference chunks including a usage count of each storage chunk, the usage count tracking free and allocated storage chunks; and
wherein the reference count map assigns a set of bits in each reference chunk as a bitset to store a usage count of each storage chunk associated with a particular reference chunk to indicate when an individual storage chunk is allocated and referred to by more than one volume or snapshot.
18. The storage server of claim 17 , wherein the reference chunks store a representation of a radix tree to index each storage chunk and the usage of each chunk.
19. The storage server of claim 17 , wherein the scalable indexing further comprises: updating reference chunks to store indexing information for redirect on write snapshots of the virtual volume.
20. The storage server of claim 17 , wherein the set of reference chunks stores: a reference count map referencing all the storage chunks in a virtual volume including usage information on storage chunks used more than once.
21. The storage server of claim 17 , wherein the set of reference chunks further stores:
a reference count of chunks referred to by the virtual volume;
a reference count of chunks referred to by clones of the virtual volume; and
a reference count of chunks referred to by any redirect on write snapshots of the virtual volume.
22. The storage server of claim 17 , wherein the storage server includes a set of interfaces for file based access and block based access of storage devices.
23. The storage server of claim 17 , wherein the storage server is a Unified Storage Server (USS).
24. The storage server of claim 22 , wherein the scalable indexing further comprises: updating reference chunks to store indexing information for redirect on write snapshots of the virtual volume.
25. The storage server of claim 24 , wherein the set of reference chunks stores:
a reference count map referencing all the storage chunks in a virtual volume including usage information on storage chunks used more than once.
26. The storage server of claim 24 , wherein the set of reference chunks further stores:
a reference count of chunks referred to by the virtual volume;
a reference count of chunks referred to by clones of the virtual volume; and
a reference count of chunks referred to by any redirect on write snapshots of the virtual volume.
27. A storage server to provide thin provisioning of storage, comprising:
at least one interface for the storage server to access storage devices;
at least one processor and a memory having computer program code which when executed on the processor is operative to perform scalable indexing of a storage pool formed from multiple storage units in which backup information is recorded, including:
generating a virtual volume in which the combined storage capacity of the storage pool is represented by a set of chunks have a pre-selected data size with the chunks being assigned to be either storage chunks or reference chunks;
storing in a set of reference chunks metadata information to index the storage chunks and track usage of storage chunks, wherein storage chunks are allocated on demand and are referable more than once to reduce redundant storage for identical content; and
adjusting the total number of chunks and a chunk allocation in response to a change in the size of the available physical storage to include a sufficient number of reference chunks to index all of the storage chunks;
wherein the set of reference chunks stores a reference count map referencing all of the storage pool chunks in the storage pool, the reference chunks including a usage count of each storage chunk, the usage count tracking free and allocated storage chunks and wherein each reference count chunk in the reference map indexes a set of storage chunks, the method further comprising:
in response to allocation of a chunk, scanning a reference count map bitset in a first reference count chunk to identify the address of a freely available chunk in the storage pool;
responsive to determining that the scanned reference count chunk does not have any free bitset available, sequentially scanning subsequent reference count chunks to identify the first freely available bitset and marking the bitset to identify a new chunk allocation.
28. A storage server to provide thin provisioning of storage, comprising:
at least one interface for the storage server to access storage devices;
at least one processor and a memory having computer program code which when executed on the processor is operative to perform scalable indexing of a storage pool formed from multiple storage units in which backup information is recorded, including:
generating a virtual volume in which the combined storage capacity of the storage pool is represented by a set of chunks have a pre-selected data size with the chunks being assigned to be either storage chunks or reference chunks;
storing in a set of reference chunks metadata information to index the storage chunks and track usage of storage chunks, wherein storage chunks are allocated on demand and are referable more than once to reduce redundant storage for identical content;
adjusting the total number of chunks and a chunk allocation in response to a change in the size of the available physical storage to include a sufficient number of reference chunks to index all of the storage chunks; and
logging a reference count change in a logging string, comprising:
an address of a reference count chunk;
an offset of a storage pool chunk within a reference count chunk;
a previous reference count; and
a new reference count.
29. A method for scalable indexing of a storage pool formed from multiple storage units in which backup information is recorded, comprising:
forming a virtual volume in which the combined storage capacity of the storage pool is represented by a set of chunks have a pre-selected data size with the chunks being assigned to be either storage chunks or reference chunks;
storing in a set of reference chunks metadata information to index the storage chunks and track usage of storage chunks, wherein storage chunks are allocated on demand and are referable more than once to reduce redundant storage for identical content;
adjusting the total number of chunks and a chunk allocation in response to a change in the size of the available physical storage to include a sufficient number of reference chunks to index all of the storage chunks;
wherein the set of reference chunks stores:
a reference count map maintained as a set of contiguous storage pool chunks disposed at one end of the storage pool or on an external storage device, the reference count map referencing all the storage chunks in a virtual volume including usage information on storage chunks used more than once including a usage count of each storage chunk.
30. The method of claim 29 , wherein the reference chunks store a representation of a radix tree to index each storage chunk and the usage of each chunk.
31. The method of claim 29 , wherein the reference chunks are updatable to store indexing information for redirect on write snapshots of the virtual volume.
32. The method of claim 29 , wherein the reference count map includes a usage count of each storage chunk, the usage count tracking free and allocated storage chunks and the number of times a storage chunk is referenced.
33. The method of claim 29 , wherein the reference count map includes a reference count of chunks referred to by any redirect on write snapshots of the virtual volume.
34. A method for scalable indexing of a storage pool formed from multiple storage units in which backup information is recorded, comprising:
forming a virtual volume in which the combined storage capacity of the storage pool is represented by a set of chunks have a pre-selected data size with the chunks being assigned to be either storage chunks or reference chunks;
storing in a set of reference chunks metadata information to index the storage chunks and track usage of storage chunks, wherein storage chunks are allocated on demand and are referable more than once to reduce redundant storage for identical content;
adjusting the total number of chunks and a chunk allocation in response to a change in the size of the available physical storage to include a sufficient number of reference chunks to index all of the storage chunks;
wherein the reference chunks having a sequential order and the reference chunks comprise a reference count map referencing all of the storage pool chunks in the storage pool, the reference chunks including a usage count of each storage chunk, the usage count tracking free and allocated storage chunks.