Method and system to implement directory reads for a database file system

Patent metadata

Field	Value
Publication number	US-12001406-B2
Application number	US-202217877689-A
Country	US
Kind code	B2
Filing date	Jul 29, 2022
Priority date	Jul 29, 2022
Publication date	Jun 4, 2024
Grant date	Jun 4, 2024

Abstract

Disclosed is an improved approach to access the contents of database file system. The contents of the database file system can be acquired and stored in a directory buffer. A parallel tree structure is used to retrieve directory contents quickly and in a consistent fashion.

First claim

What is claimed is:

1. A method, comprising:
implementing a database file system (DBFS) where a plurality of files in a file system are stored in a database backend;
receiving a first file system command to access a directory in the DBFS, the directory having multiple files of the plurality of files in the file system;
constructing a node tree for the multiple files in the directory in response to the first file system command, wherein the node tree is constructed based on a table for the directory in the database backend having columns that hold metadata associated with the files;
loading file contents for the files in the directory into a cache in response to the first file system command to access the directory, wherein the node tree references the file contents for the files in the cache;
receiving a second file system command to access a file of the multiple files in the directory; and
accessing, in response to the second file system command, the previously constructed node tree to identify file contents of the file in the cache based on a reference to the file contents for the file in the node tree.

2. The method of claim 1 , wherein the node tree is constructed as a red black tree.

3. The method of claim 1 , wherein an individual row in the table in the database backend corresponds to a specific file, and the individual row corresponds to a LOB that holds file content for the file.

4. The method of claim 3 , wherein the individual row comprises a pointer to a location for the LOB located outside of the table.

5. The method of claim 1 , wherein the DBFS provides an abstraction layer that is embodied as a mounted file system at a client location, and where the first file system command is translated into a database operation request that is processed at the database backend to handle the first file system command.

6. The method of claim 1 , wherein the node tree comprises nodes that correspond to the files in the directory, and where an individual node comprises a key and an offset associated with a specific file.

7. The method of claim 6 , wherein the file contents are loaded into a buffer in the cache and the offset is used to reference a location in the buffer for file system content for the specific file.

8. The method of claim 1 , wherein the node tree is walked to process the first file system command, wherein an iterative approach is taken to walk different portions of the node tree to process different subsets of the directory to implement the first file system command.

9. The method of claim 1 , wherein the node tree is not deleted until after a change to the directory or after a timeout threshold has been reached.

10. A system, comprising:
a processor;
a memory for storing a set of instructions; and
wherein the a set of instructions, when executed by the processor cause a set of acts, the set of acts comprise: implementing a database file system (DBFS) where a plurality of files in a file system are stored in a database backend; receiving a first file system command to access a directory in the DBFS, the directory having multiple files of the plurality of files in the file system; constructing a node tree for the multiple files in the directory in response to the first file system command, wherein the node tree is constructed based on a table for the directory in the database backend having columns that hold metadata associated with the files; loading file contents for the files in the directory into a cache in response to the first file system command to access the directory, wherein the node tree references the file contents for the files in the cache; receiving a second file system command to access a file of the multiple files in the directory; and accessing, in response to the second file system command, the previously constructed node tree to identify file contents of the file in the cache based on a reference to the file contents for the file in the node tree.

11. The system of claim 10 , wherein the node tree is constructed as a red black tree.

12. The system of claim 10 , wherein an individual row in the table in the database backend corresponds to a specific file, and the individual row corresponds to a LOB that holds file content for the file.

13. The system of claim 12 , wherein the individual row comprises a pointer to a location for the LOB located outside of the table.

14. The system of claim 10 , wherein the DBFS provides an abstraction layer that is embodied as a mounted file system at a client location, and where the first file system command is translated into a database operation request that is processed at the database backend to handle the first file system command.

15. The system of claim 10 , wherein the node tree comprises nodes that correspond to the files in the directory, and where an individual node comprises a key and an offset associated with a specific file.

16. The system of claim 15 , wherein the file contents are loaded into a buffer in the cache and the offset is used to reference a location in the buffer for file system content for the specific file.

17. The system of claim 10 , wherein the node tree is walked to process the first file system command, wherein an iterative approach is taken to walk different portions of the node tree to process different subsets of the directory to implement the first file system command.

18. A computer program product embodied on a non-transitory computer readable medium having stored thereon a sequence of instructions which, when executed by a processor, performs:
implementing a database file system (DBFS) where a plurality of files in a file system are stored in a database backend;
receiving a first file system command to access a directory in the DBFS, the directory having multiple files of the plurality of files in the file system;
constructing a node tree for the multiple files in the directory in response to the first file system command, wherein the node tree is constructed based on a table for the directory in the database backend having columns that hold metadata associated with the files;
loading file contents for the files in the directory into a cache in response to the first file system command to access the directory, wherein the node tree references the file contents for the files in the cache;
receiving a second file system command to access a file of the multiple files in the directory; and
accessing, in response to the second file system command, the previously constructed node tree to identify file contents of the file in the cache based on a reference to the file contents for the file in the node tree.

19. The computer program product of claim 18 , wherein the node tree is constructed as a red black tree.

20. The computer program product of claim 18 , wherein an individual row in the table in the database backend corresponds to a specific file, and the individual row corresponds to a LOB that holds file content for the file.

21. The computer program product of claim 20 , wherein the individual row comprises a pointer to a location for the LOB located outside of the table.

22. The computer program product of claim 18 , wherein the DBFS provides an abstraction layer that is embodied as a mounted file system at a client location, and where the first file system command is translated into a database operation request that is processed at the database backend to handle the first file system command.

23. The computer program product of claim 18 , wher

Assignees

• Oracle Int Corp

Inventors

• Susairaj Margaret
• Lasa Gutierrez Jose Antonio
• Chatterjee Sumanta

Classifications

• G06F16/185 Primary

  Hierarchical storage management [HSM] systems, e.g. file migration or policies thereof (details of archiving G06F16/11) · CPC title

• G06F16/172 Primary

  Caching, prefetching or hoarding of files · CPC title

• G06F16/2219

  Large Object storage; Management thereof · CPC title