Metadata management for virtual volumes

1. A method comprising

loading into memory at least a portion of metadata for a virtual volume (VV) that spans data extents of different persistent storage devices, wherein the metadata comprises virtual metadata block (VMB) descriptors and virtual metadata blocks (VMBs);

mapping an address of the VV to a VMB number and an index of an extent pointer within a VMB identified by the VMB number, wherein the extent pointer indicates an extent within one of the different persistent storage devices;

locating a VMB descriptor in the memory based on the VMB number; and

locating the identified VMB in the memory or in one of the different persistent storage devices based on the located VMB descriptor.

2. The method of claim 1 , wherein locating the identified VMB not in the memory comprises locating the identified VMB in one of the different persistent storage devices, and the method comprises:

selecting for eviction another VMB that is in the memory;

swapping out the selected VMB to free memory space for the identified VMB; and

updating VMB descriptors for the selected VMB and the identified VMB.

3. The method of claim 2 , wherein the selecting comprises using an Adaptive Replacement Cache (ARC), and the updating comprises changing IN_CORE and SWAPPED_OUT flags in the VMB descriptors for the selected VMB and the identified VMB.

4. The method of claim 2 , wherein the loading comprises loading metadata for multiple sparse virtual volumes, and locating the VMB descriptor comprises locating the VMB descriptor in the memory, based on a VV identifier and the VMB number, using a hash table.

5. The method of claim 1 , wherein each of the VMBs includes VV identifier and VMB number meta-metadata, and each of the VMBs for a specified virtual volume includes a same number of extent pointers, each extent referenced by one of the extent pointers in the VMBs for the specified virtual volume has a same number of blocks, and the mapping comprises:

determining the VMB number by dividing a block offset by the number of extent pointers per VMB and by the number of blocks per extent; and

determining the index of the extent pointer by dividing the block offset by the number of blocks per extent, modulo the number of extent pointers per VMB.

6. The method of claim 1 , wherein the different persistent storage devices comprise flash memory devices.

7. The method of claim 1 , wherein the extent pointer is a logical address.

8. An apparatus comprising

a processor,

a memory coupled with the processor; and

a computer-readable medium embodying a program operable to cause the processor to perform operations comprising (i) loading into the memory at least a portion of metadata for a virtual volume (VV) that spans data extents of different persistent storage devices, wherein the metadata comprises virtual metadata block (VMB) descriptors and virtual metadata blocks (VMBs), (ii) mapping an address of the VV to a VMB number and an index of an extent pointer within a VMB identified by the VMB number, wherein the extent pointer indicates an extent within one of the different persistent storage devices, (iii) locating a VMB descriptor in the memory based on the VMB number, and (iv) locating the identified VMB in the memory or in one of the different persistent storage devices based on the located VMB descriptor.

9. The apparatus of claim 8 , wherein locating the identified VMB not in the memory comprises locating the identified VMB in one of the different persistent storage devices, and the operations comprise:

selecting for eviction another VMB that is in the memory;

swapping out the selected VMB to free memory space for the identified VMB; and

updating VMB descriptors for the selected VMB and the identified VMB.

10. The apparatus of claim 9 , wherein the selecting comprises using an Adaptive Replacement Cache (ARC), and the updating comprises changing IN_CORE and SWAPPED_OUT flags in the VMB descriptors for the selected VMB and the identified VMB.

11. The apparatus of claim 9 , wherein the loading comprises loading metadata for multiple sparse virtual volumes, and locating the VMB descriptor comprises locating the VMB descriptor in the memory, based on a VV identifier and the VMB number, using a hash table.

12. The apparatus of claim 8 , wherein each of the VMBs includes VV identifier and VMB number meta-metadata, and each of the VMBs for a specified virtual volume includes a same number of extent pointers, each extent referenced by one of the extent pointers in the VMBs for the specified virtual volume has a same number of blocks, and the mapping comprises:

determining the VMB number by dividing a block offset by the number of extent pointers per VMB and by the number of blocks per extent; and

determining the index of the extent pointer by dividing the block offset by the number of blocks per extent, modulo the number of extent pointers per VMB.

13. The apparatus of claim 8 , wherein the different persistent storage devices comprise flash memory devices.

14. The apparatus of claim 8 , wherein the extent pointer is a logical address.

15. A system comprising

a storage pool comprising different persistent storage devices; and

a data processing apparatus configured to (i) load into memory at least a portion of metadata for a virtual volume (VV) that spans data extents of the different persistent storage devices, wherein the metadata comprises virtual metadata block (VMB) descriptors and virtual metadata blocks (VMBs), (ii) map an address of the VV to a VMB number and an index of an extent pointer within a VMB identified by the VMB number, wherein the extent pointer indicates an extent within one of the different persistent storage devices, (iii) locate a VMB descriptor in the memory based on the VMB number, and (iv) locate the identified VMB in the memory or in one of the different persistent storage devices based on the located VMB descriptor.

16. The system of claim 15 , the data processing apparatus configured to select for eviction a VMB that is in the memory, swap out the selected VMB to free memory space for the identified VMB, and update VMB descriptors for the selected VMB and the identified VMB.

17. The system of claim 16 , the data processing apparatus configured to use an Adaptive Replacement Cache (ARC), and change IN_CORE and SWAPPED_OUT flags in the VMB descriptors for the selected VMB and the identified VMB.

18. The system of claim 16 , the data processing apparatus configured to load metadata for multiple sparse virtual volumes, and locate the VMB descriptor in the memory, based on a VV identifier and the VMB number, using a hash table.

19. The system of claim 15 , wherein each of the VMBs includes VV identifier and VMB number meta-metadata, and each of the VMBs for a specified virtual volume includes a same number of extent pointers, each extent referenced by one of the extent pointers in the VMBs for the specified virtual volume has a same number of blocks, and the data processing apparatus is configured to determine the VMB number by dividing a block offset by the number of extent pointers per VMB and by the number of blocks per extent, and determine the index of the extent pointer by dividing the block offset by the number of blocks per extent, modulo the number of extent pointers per VMB.

20. The system of claim 15 , wherein the storage pool comprises RAID (Redundant Array of Independent Disks) groups.

21. The system of claim 15 , wherein the extent pointer is a logical address.

22. The system of claim 15 , comprising a NAS (Network-Attached Storage) system.