Granted Patent B1

US 12,292,853 · App. 18/503,082 · Granted May 6, 2025

Inventors: Graham Edwin Ellis (Seattle, WA); Ying Fairweather (Sammamish, WA); Thorne Davis Garvin (Spokane, WA); Steven Henry Haber (Seattle, WA); Yuxi Bai (Seattle, WA); Michael Anthony Chmiel (Seattle, WA); Pathirat Kosakanchit (Seattle, WA); Jonathan Michael MacLaren (Seattle, WA); Matthew Christopher McMullan (Bellevue, WA); Tyler Morrison Moody (Seattle, WA); Aaron James Passey (San Rafael, CA); Rowan Arthur Phipps (Seattle, WA); Thomas Gregory Rothschilds (Seattle, WA)

Assignee: Qumulo, Inc.

G06F16/122G06F12/0253G06F2212/702

Abstract

Claims (181)

1. A method for managing data in a file system over a network using one or more processors to execute instructions that are configured to cause performance of actions, comprising:

determining an object store for one or more write requests provided by a file system client based on the file system, wherein each write request includes one or more data blocks associated with a write transaction;

generating one or more write ahead log (WAL) entries that include the one or more data blocks, wherein one or more of a rule, instruction or a library is employed to determine one or more types of index data structures selected from a plurality of types of data structures for efficiently including the one or more data blocks with a WAL entry;

generating a WAL object based on the one or more WAL entries;

storing the WAL object in the object store;

updating an in-memory overlay with the one or more types of determined index data structures to associate the one or more data blocks with the WAL object;

deleting each WAL object in the object store that is unassociated with one or more data blocks and deleting each non-referenced data object in the object store; and

employing a threshold for a number of WAL objects stored in the object store to initiate a checkpoint operation that performs further actions, including:

generating an index object that includes one or more index entries that associate one or more other data blocks with one or more data objects stored in the object store;

updating the index object to include one or more index entries that associate the one or more data blocks with the WAL object;

storing the updated index object in the object store, wherein the WAL object is converted to a data object based on the WAL object being referenced in the updated index object;

updating the in-memory overlay to remove the association of the one or more data blocks and the WAL object; and

initiating garbage collection for one or more data objects stored in the object store based on one or more additional financial costs charged by a cloud computing provider of an application programming interface (API) for the file system due to one or more of an amount of stored data objects being above a predetermined amount for the object store or a total amount of data objects.

2. The method of claim 1 , further comprising:

invalidating a portion of a page cache that includes a copy of the one or more data blocks.

3. The method of claim 1 , further comprising:

determining the object store for one or more read requests provided by the file system client based on the file system, wherein each read request includes address information for one or more read data blocks associated with a read transaction;

employing the in-memory index to determine a data object in the object store that includes the one or more read data blocks;

loading the determined data object from the object store;

determining the one or more read data blocks in the loaded data object based on the address information; and

sending a response to the file system client that includes a copy of the one or more read data blocks.

4. The method of claim 1 , further comprising:

determining the object store for one or more read requests provided by the file system client based on the file system, wherein each read request includes address information for one or more read data blocks associated with a read transaction;

employing the in-memory overlay to determine a WAL object in the object store that includes the one or more read data blocks;

loading the determined WAL object from the object store;

determining the one or more read data blocks in the loaded WAL object based on the address information; and

sending a response to the file system client that includes a copy of the one or more read data blocks.

5. The method of claim 1 , further comprising:

determining an amount of waste data stored in the object store based on one or more metrics;

initiating the garbage collection operation based on one or more of the amount of waste data and a cost to store the waste space in the object store exceeding another cost of executing the garbage collection operation;

determining one or more fragmented data objects in the object store that include the waste data based on one or more reference counters;

generating zero or more consolidated data objects that exclude the waste data based on the one or more fragmented data objects;

storing the zero or more consolidated data objects in the object store; and

updating the overlay to include zero or more references to the zero or more consolidated data objects, or performing additional actions, including:

loading a current index object from the object store;

generating a modified index object based on the current index object and the zero or more consolidated data objects, wherein the modified index object is stored in the object store; and

removing the one or more fragmented data objects and the current index object from the object store.

6. The method of claim 1 , wherein the one or more data objects further comprise:

an index block in each data object that references a location of each data block included in the one or more data objects, wherein the index block is located at the beginning of the one or more data objects.

7. The method of claim 1 , wherein executing a checkpoint operation, further comprises:

determining one or more globally committed WAL objects that correspond to one or more committed write transactions;

determining one or more locally committed WAL objects that are associated with one or more uncommitted write transactions; and

excluding the one or more locally committed WAL objects from the checkpoint operation.

8. A system for managing data in a file system comprising:

a network computer, comprising:

a memory that stores at least instructions; and

one or more processors that execute instructions that are configured to cause performance of actions, including:

determining an object store for one or more write requests provided by a file system client based on the file system, wherein each write request includes one or more data blocks associated with a write transaction;

generating one or more write ahead log (WAL) entries that include the one or more data blocks, wherein one or more of a rule, instruction or a library is employed to determine one or more types of index data structures selected from a plurality of types of data structures for efficiently including the one or more data blocks with a WAL entry;

generating a WAL object based on the one or more WAL entries;

storing the WAL object in the object store;

updating an in-memory overlay with the one or more types of determined index data structures to associate the one or more data blocks with the WAL object;

deleting each WAL object in the object store that is unassociated with one or more data blocks and deleting each non-referenced data object in the object store; and

employing a threshold for a number of WAL objects stored in the object store to initiate a checkpoint operation that performs further actions, including:

generating an index object that includes one or more index entries that associate one or more other data blocks with one or more data objects stored in the object store;

updating the index object to include one or more index entries that associate the one or more data blocks with the WAL object;

storing the updated index object in the object store, wherein the WAL object is converted to a data object based on the WAL object being referenced in the updated index object; and

updating the in-memory overlay to remove the association of the one or more data blocks and the WAL object; and

initiating garbage collection for one or more data objects stored in the object store based on one or more additional financial costs charged by a cloud computing provider of an application programming interface (API) for the file system due to one or more of an amount of stored data objects being above a predetermined amount for the object store or a total amount of data objects; and

one or more client computers, comprising:

a memory that stores at least instructions; and

one or more processors that execute instructions that are configured to cause performance of other actions, including:

providing the one or more write requests to the file system.

9. The system of claim 8 , wherein the one or more network computer processors execute instructions that are configured to cause performance of actions, further comprising:

invalidating a portion of a page cache that includes a copy of the one or more data blocks.

10. The system of claim 8 , wherein the one or more network computer processors execute instructions that are configured to cause performance of actions, further comprising:

determining the object store for one or more read requests provided by the file system client based on the file system, wherein each read request includes address information for one or more read data blocks associated with a read transaction;

employing the in-memory index to determine a data object in the object store that includes the one or more read data blocks;

loading the determined data object from the object store;

determining the one or more read data blocks in the loaded data object based on the address information; and

sending a response to the file system client that includes a copy of the one or more read data blocks.

11. The system of claim 8 , wherein the one or more network computer processors execute instructions that are configured to cause performance of actions, further comprising:

determining the object store for one or more read requests provided by the file system client based on the file system, wherein each read request includes address information for one or more read data blocks associated with a read transaction;

employing the in-memory overlay to determine a WAL object in the object store that includes the one or more read data blocks;

loading the determined WAL object from the object store;

determining the one or more read data blocks in the loaded WAL object based on the address information; and

sending a response to the file system client that includes a copy of the one or more read data blocks.

12. The system of claim 8 , wherein the one or more network computer processors execute instructions that are configured to cause performance of actions, further comprising:

determining an amount of waste data stored in the object store based on one or more metrics;

initiating a garbage collection operation based on one or more of the amount of waste data and a cost to store the waste space in the object store exceeding another cost of executing the garbage collection operation;

determining one or more fragmented data objects in the object store that include the waste data based on one or more reference counters;

generating zero or more consolidated data objects that exclude the waste data based on the one or more fragmented data objects;

storing the zero or more consolidated data objects in the object store; and

updating the overlay to include zero or more references to the zero or more consolidated data objects, or performing additional actions, including:

loading a current index object from the object store;

generating a modified index object based on the current index object and the zero or more consolidated data objects, wherein the modified index object is stored in the object store; and

removing the one or more fragmented data objects and the current index object from the object store.

13. The system of claim 8 , wherein the one or more data objects further comprise:

an index block in each data object that references a location of each data block included in the one or more data objects, wherein the index block is located at the beginning of the one or more data objects.

14. The system of claim 8 , wherein executing a checkpoint operation, further comprises:

determining one or more globally committed WAL objects that correspond to one or more committed write transactions;

determining one or more locally committed WAL objects that are associated with one or more uncommitted write transactions; and

excluding the one or more locally committed WAL objects from the checkpoint operation.

15. A processor readable non-transitory storage media that includes instructions for managing data in a file system over a network, wherein execution of the instructions by one or more processors on one or more network computers performs actions, comprising:

determining an object store for one or more write requests provided by a file system client based on the file system, wherein each write request includes one or more data blocks associated with a write transaction;

generating one or more write ahead log (WAL) entries that include the one or more data blocks, wherein one or more of a rule, instruction or a library is employed to determine one or more types of index data structures selected from a plurality of types of data structures for efficiently including the one or more data blocks with a WAL entry;

generating a WAL object based on the one or more WAL entries;

storing the WAL object in the object store;

updating an in-memory overlay with the one or more types of determined index data structures to associate the one or more data blocks with the WAL object;

deleting each WAL object in the object store that is unassociated with one or more data blocks and deleting each non-referenced data object in the object store; and

employing a threshold for a number of WAL objects stored in the object store to initiate a checkpoint operation that performs further actions, including:

generating an index object that includes one or more index entries that associate one or more other data blocks with one or more data objects stored in the object store;

updating the index object to include one or more index entries that associate the one or more data blocks with the WAL object;

storing the updated index object in the object store, wherein the WAL object is converted to a data object based on the WAL object being referenced in the updated index object; and

updating the in-memory overlay to remove the association of the one or more data blocks and the WAL object; and

initiating garbage collection for one or more data objects stored in the object store based on one or more additional financial costs charged by a cloud computing provider of an application programming interface (API) for the file system due to one or more of an amount of stored data objects being above a predetermined amount for the object store or a total amount of data objects.

16. The media of claim 15 , further comprising:

invalidating a portion of a page cache that includes a copy of the one or more data blocks.

17. The media of claim 15 , further comprising:

determining the object store for one or more read requests provided by the file system client based on the file system, wherein each read request includes address information for one or more read data blocks associated with a read transaction;

employing the in-memory index to determine a data object in the object store that includes the one or more read data blocks;

loading the determined data object from the object store;

determining the one or more read data blocks in the loaded data object based on the address information; and

sending a response to the file system client that includes a copy of the one or more read data blocks.

18. The media of claim 15 , further comprising:

determining the object store for one or more read requests provided by the file system client based on the file system, wherein each read request includes address information for one or more read data blocks associated with a read transaction;

employing the in-memory overlay to determine a WAL object in the object store that includes the one or more read data blocks;

loading the determined WAL object from the object store;

determining the one or more read data blocks in the loaded WAL object based on the address information; and

sending a response to the file system client that includes a copy of the one or more read data blocks.

19. The media of claim 15 , further comprising:

determining one or more of an amount of waste data stored in the object store based on one or more metrics;

initiating a garbage collection operation based on one or more of the amount of waste data and a cost to store the waste space in the object store exceeding another cost of executing the garbage collection operation;

determining one or more fragmented data objects in the object store that include the waste data based on one or more reference counters;

generating zero or more consolidated data objects that exclude the waste data based on the one or more fragmented data objects;

storing the zero or more consolidated data objects in the object store; and

updating the overlay to include zero or more references to the zero or more consolidated data objects, or performing additional actions, including:

loading a current index object from the object store;

generating a modified index object based on the current index object and the zero or more consolidated data objects, wherein the modified index object is stored in the object store; and

removing the one or more fragmented data objects and the current index object from the object store.

20. The media of claim 15 , wherein the one or more data objects further comprise:

an index block in each data object that references a location of each data block included in the one or more data objects, wherein the index block is located at the beginning of the one or more data objects.

21. The media of claim 15 , wherein executing a checkpoint operation, further comprises:

determining one or more globally committed WAL objects that correspond to one or more committed write transactions;

determining one or more locally committed WAL objects that are associated with one or more uncommitted write transactions; and

excluding the one or more locally committed WAL objects from the checkpoint operation.

22. A network computer for managing data in a file system, comprising:

a memory that stores at least instructions; and

one or more processors that execute instructions that are configured to cause performance of actions, including:

determining an object store for one or more write requests provided by a file system client based on the file system, wherein each write request includes one or more data blocks associated with a write transaction;

generating one or more write ahead log (WAL) entries that include the one or more data blocks, wherein one or more of a rule, instruction or a library is employed to determine one or more types of index data structures selected from a plurality of types of data structures for efficiently including the one or more data blocks with a WAL entry;

generating a WAL object based on the one or more WAL entries;

storing the WAL object in the object store;

updating an in-memory overlay with the one or more types of determined index data structures to associate the one or more data blocks with the WAL object;

deleting each WAL object in the object store that is unassociated with one or more data blocks and deleting each non-referenced data object in the object store; and

employing a threshold for a number of WAL objects stored in the object store to initiate a checkpoint operation that performs further actions, including:

generating an index object that includes one or more index entries that associate one or more other data blocks with one or more data objects stored in the object store;

updating the index object to include one or more index entries that associate the one or more data blocks with the WAL object;

storing the updated index object in the object store, wherein the WAL object is converted to a data object based on the WAL object being referenced in the updated index object; and

updating the in-memory overlay to remove the association of the one or more data blocks and the WAL object; and

initiating garbage collection for one or more data objects stored in the object store based on one or more additional financial costs charged by a cloud computing provider of an application programming interface (API) for the file system due to one or more of an amount of stored data objects being above a predetermined amount for the object store or a total amount of data objects.

23. The network computer of claim 22 , wherein the one or more processors execute instructions that are configured to cause performance of actions, further comprising:

invalidating a portion of a page cache that includes a copy of the one or more data blocks.

24. The network computer of claim 22 , wherein the one or more processors execute instructions that are configured to cause performance of actions, further comprising:

determining the object store for one or more read requests provided by the file system client based on the file system, wherein each read request includes address information for one or more read data blocks associated with a read transaction;

employing the in-memory index to determine a data object in the object store that includes the one or more read data blocks;

loading the determined data object from the object store;

determining the one or more read data blocks in the loaded data object based on the address information; and

sending a response to the file system client that includes a copy of the one or more read data blocks.

25. The network computer of claim 22 , wherein the one or more processors execute instructions that are configured to cause performance of actions, further comprising:

determining the object store for one or more read requests provided by the file system client based on the file system, wherein each read request includes address information for one or more read data blocks associated with a read transaction;

employing the in-memory overlay to determine a WAL object in the object store that includes the one or more read data blocks;

loading the determined WAL object from the object store;

determining the one or more read data blocks in the loaded WAL object based on the address information; and

sending a response to the file system client that includes a copy of the one or more read data blocks.

26. The network computer of claim 22 , wherein the one or more processors execute instructions that are configured to cause performance of actions, further comprising:

determining an amount of waste data stored in the object store based on one or more metrics;

initiating a garbage collection operation based on one or more of the amount of waste data and a cost to store the waste space in the object store exceeding another cost of executing the garbage collection operation;

determining one or more fragmented data objects in the object store that include the waste data based on one or more reference counters;

generating zero or more consolidated data objects that exclude the waste data based on the one or more fragmented data objects;

storing the zero or more consolidated data objects in the object store; and

updating the overlay to include zero or more references to the zero or more consolidated data objects, or performing additional actions, including:

loading a current index object from the object store;

generating a modified index object based on the current index object and the zero or more consolidated data objects, wherein the modified index object is stored in the object store; and

removing the one or more fragmented data objects and the current index object from the object store.

27. The network computer of claim 22 , wherein the one or more data objects further comprise:

an index block in each data object that references a location of each data block included in the one or more data objects, wherein the index block is located at the beginning of the one or more data objects.

28. The network computer of claim 22 , wherein executing a checkpoint operation, further comprises:

determining one or more globally committed WAL objects that correspond to one or more committed write transactions;

determining one or more locally committed WAL objects that are associated with one or more uncommitted write transactions; and

excluding the one or more locally committed WAL objects from the checkpoint operation.

Assignments (2)

References Cited (400)

Cited By (9)

US 12,443,559 US 12,443,568 US 12,481,625 US 12,517,684 US 12,585,563 US 12,619,582 US 12,670,081 US 12,743,357 US 12,748,730