Storage pool scrubbing with concurrent snapshots

• US 8,280,858 B2
• Filed: 06/29/2009
• Issued: 10/02/2012
• Est. Priority Date: 06/29/2009
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Oracle America, Inc. (Oracle Corporation)
• Inventors: Ahrens, Matthew A., Maybee, Mark J.
• Primary Examiner(s): Mahmoudi, Tony
• Assistant Examiner(s): Weinrich, Brian E.
• Application Number:
  US12/494,090
• Publication Number:
  US 20100332446A1
• Time in Patent Office: 1,191 Days
• Field of Search: 707/692, 707/657, 707/790
• US Class Current: 707/692
• CPC Class Codes: G06F 16/10 File systems; File servers

Abstract

A method for scrubbing a storage pool is disclosed. The method includes loading a scrub queue with a number of identifiers corresponding to a number of datasets, selecting a first identifier from the scrub queue, where the first identifier corresponds to a first dataset, initiating a scrubbing of the first dataset. The method further includes, upon receiving an indication of a requirement to pause the scrubbing, pausing the scrubbing of the first dataset, creating a bookmark recording a last location within the first dataset that was scrubbed before pausing the scrubbing of the first dataset, detecting at least one change to a storage pool, and performing a modification in response to detecting the at least one change to the storage pool.

Claims

1. A non-transitory computer readable storage medium having a computer readable program code embodied therein, the computer readable program code adapted to be executed by a processor to perform a method, the method comprising:

loading a scrub queue with a plurality of identifiers corresponding to a plurality of datasets in a storage pool;

selecting a first identifier from the scrub queue, wherein the first identifier corresponds to a first dataset, wherein the first dataset is a dataset of the plurality of datasets;

initiating a scrubbing of the first dataset, wherein the first dataset is a hierarchical block tree (HBT) comprises;

a plurality of data blocks and at least one indirect block, wherein the indirect block comprises a first block pointer that references at least one of the plurality of data blocks, and wherein the first block pointer comprises a birth time associated with the at least one of the plurality of data blocks, and

a root block comprising a second block pointer referencing the at least one indirect block, wherein the second block pointer comprises a birth time associated with the at least one indirect block; and

upon receiving an indication of a requirement to pause the scrubbing;

pausing the scrubbing of the first dataset;

creating a bookmark recording of a last location within the first dataset that was scrubbed before pausing the scrubbing of the first dataset;

detecting at least one change among the plurality of datasets in the storage pool; and

performing a modification to one of the plurality of identifiers in the scrub queue in response to detecting the at least one change.

2. The non-transitory computer readable storage medium of claim 1, wherein scrubbing the first dataset comprises scrubbing only blocks having a birth time after a birth time of a previously scrubbed dataset.

3. The non-transitory computer readable storage medium of claim 1, wherein scrubbing the first dataset comprises validating that a checksum is correct, wherein the checksum is associated with at least one selected from a group comprising a data block and an indirect block.

4. The non-transitory computer readable storage medium of claim 1, wherein scrubbing the first dataset comprises resilvering at least one selected from a group comprising a data block and an indirect block.

5. The non-transitory computer readable storage medium of claim 1, further comprising, upon completing the scrubbing of the first dataset:

removing the first identifier from the scrub queue.

6. The non-transitory computer readable storage medium of claim 5, further comprising:

selecting a second identifier from the scrub queue, wherein the second identifier corresponds to the oldest dataset remaining in the scrub queue.

7. The non-transitory computer readable storage medium of claim 1, wherein the storage pool comprises at least one physical disk.

8. The non-transitory computer readable storage medium of claim 1, wherein each of the plurality of datasets is associated with a plurality of pointers comprising:

a snapshot pointer; and

a clones pointer.

9. The non-transitory computer readable storage medium of claim 8, wherein detecting the at least one change to the storage pool comprises detecting a deletion of the first dataset from the storage pool, and wherein performing the modification comprises:

determining, based on a snapshot pointer associated with the first dataset, a descendant dataset of the first dataset;

resetting the bookmark to a beginning of the descendant dataset; and

initiating a scrubbing of the descendant dataset according to the bookmark.

10. The non-transitory computer readable storage medium of claim 8, wherein detecting the at least one change to the storage pool comprises detecting a deletion of a second dataset from the storage pool, and wherein performing the modification comprises:

determining, based on a snapshot pointer associated with the second dataset, a descendant dataset of the second dataset;

replacing, within the scrub queue, an identifier corresponding to the second dataset with an identifier corresponding to the descendant dataset; and

resuming the scrubbing of the first dataset according to the bookmark.

11. The non-transitory computer readable storage medium of claim 8, wherein detecting the at least one change to the storage pool comprises detecting a clone swap between the first dataset and a clone dataset, and wherein performing the modification comprises:

swapping an identifier of the first dataset with an identifier of the clone dataset;

resetting the bookmark to a beginning of the clone dataset; and

initiating a scrubbing of the clone dataset according to the bookmark.

12. The non-transitory computer readable storage medium of claim 1, wherein detecting the at least one change to the storage pool comprises detecting that a snapshot has been created from the first dataset, and wherein performing the modification comprises initiating a scrubbing of the snapshot according to the bookmark.

13. The non-transitory computer readable storage medium of claim 1, wherein detecting the at least one change to the storage pool comprises detecting that a snapshot has been created from a second dataset, wherein the second dataset is associated with an identifier loaded in the scrub queue, and wherein performing the modification comprises:

replacing, within the scrub queue, the identifier associated with the second dataset with an identifier associated with the snapshot; and

resuming the scrubbing of the first dataset according to the bookmark.

14. The non-transitory computer readable storage medium of claim 1, wherein detecting the at least one change to the storage pool comprises detecting a clone swap between at least one dataset and a clone dataset, and wherein performing the modification comprises:

swapping an identifier of the at least one dataset with an identifier of the clone dataset; and

resuming the scrubbing of the first dataset according to the bookmark.

15. A system, comprising:

a processor;

a computer readable storage medium having computer readable program code embodied therein, the computer readable program code adapted to, when executed by the processor, implement a method, the method comprising;

loading a scrub queue with a plurality of identifiers corresponding to a plurality of datasets to be scrubbed, wherein the plurality of datasets are stored in a storage pool;

selecting a first identifier from the scrub queue, wherein the first identifier corresponds to a first dataset, wherein the first dataset is an oldest dataset of the plurality of datasets;

initiating a scrubbing of the first dataset, wherein the first dataset is a hierarchical block tree (HBT) comprises;

a plurality of data blocks and at least one indirect block, wherein the indirect block comprises a first block pointer that references at least one of the plurality of data blocks, and wherein the first block pointer comprises a birth time associated with the at least one of the plurality of data blocks, and

a root block comprising a second block pointer referencing the at least one indirect block, wherein the second block pointer comprises a birth time associated with the at least one indirect block; and

upon receiving an indication of a requirement to pause the scrubbing;

pausing the scrubbing of the first dataset;

creating a bookmark recording of a last location within the first dataset that was scrubbed before pausing the scrubbing of the first dataset;

detecting at least one change among the plurality of datasets in the storage pool; and

performing a modification to one of the plurality of identifiers in the scrub queue in response to detecting the at least one change.

16. The system of claim 15, wherein scrubbing the first dataset comprises validating that a checksum is correct, wherein the checksum is associated with at least one selected from a group comprising a data block and an indirect block.

17. The system of claim 15, wherein the bookmark comprises the first identifier, an object identifier corresponding to a file in a dataset being scrubbed prior to pausing the scrubbing, a level identifier of the dataset being scrubbed prior to pausing the scrubbing, and a block identifier corresponding to the block within the level being scrubbed prior to pausing the scrubbing.