Directly mapped buffer cache on non-volatile memory

• US 10,719,446 B2
• Filed: 08/31/2017
• Issued: 07/21/2020
• Est. Priority Date: 08/31/2017
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Loaiza, Juan R., Lee, J. William, Hu, Wei-Ming, Umamageswaran, Kothanda, MacNaughton, Neil J. S., Lee, Adam Y.
• Primary Examiner(s): Rones, Charles
• Assistant Examiner(s): Vo, Tong B.
• Application Number:
  US15/693,273
• Publication Number:
  US 20190065383A1
• Time in Patent Office: 1,055 Days
• CPC Class Codes: G06F 12/0866 for peripheral storage systems, e.g. disk cache
• CPC Class Codes: G06F 12/0868 Data transfer between cache memory and other subsystems, e.g. storage devices or host systems
• CPC Class Codes: G06F 12/0873 Mapping of cache memory to specific storage devices or parts thereof
• CPC Class Codes: G06F 16/13 File access structures, e.g. distributed indices arrangements of input from, or output to, record carriers G06F3/06
• CPC Class Codes: G06F 16/172 Caching, prefetching or hoarding of files
• CPC Class Codes: G06F 2212/1016 Performance improvement
• CPC Class Codes: G06F 2212/222 Non-volatile memory
• CPC Class Codes: G06F 2212/463 File
• CPC Class Codes: G06F 2212/465 Structured object, e.g. database record
• CPC Class Codes: G06F 2212/608 Details relating to cache mapping

Abstract

A method and apparatus for implementing a buffer cache for a persistent file system in non-volatile memory is provided. A set of data is maintained in one or more extents in non-volatile random-access memory (NVRAM) of a computing device. At least one buffer header is allocated in dynamic random-access memory (DRAM) of the computing device. In response to a read request by a first process executing on the computing device to access one or more first data blocks in a first extent of the one or more extents, the first process is granted direct read access of the first extent in NVRAM. A reference to the first extent in NVRAM is stored in a first buffer header. The first buffer header is associated with the first process. The first process uses the first buffer header to directly access the one or more first data blocks in NVRAM.

Claims

1. A method comprising:

maintaining a set of data blocks in one or more extents in a non-volatile random-access memory (NVRAM) of a computing device;

allocating a plurality of data buffers in a dynamic random-access memory (DRAM);

allocating a plurality of buffer headers in said DRAM of the computing device, each buffer header of said buffer headers being;

configurable to store a reference to a data buffer of said plurality of data buffers that stores a DRAM-copy of one or more data blocks of said set of data blocks, and

configurable to store a reference to NVRAM of one or more data blocks of said set of data blocks;

determining whether an access request to access one or more first data blocks in a first extent of the one or more extents is a read request, said access request being made by a first process executing on the computing device; and

when said access request is a read request, granting the first process direct read access of the one or more first data blocks in the NVRAM by at least;

storing, in a first buffer header of said plurality of buffer headers, a reference to the one or more first data blocks in the first extent in the NVRAM,

associating the first buffer header with the first process, and

the first process using the first buffer header to directly access the one or more first data blocks in the NVRAM; and

when said access request is not a read request, granting the first process write access by at least;

generating a first DRAM-copy of the one or more first data blocks in a first data buffer of said plurality of data buffers,

storing, in the first buffer header, a reference to the first DRAM-copy of the one or more first data blocks,

associating the first buffer header with the first process, and

the first process using the first buffer header to access the one or more first data blocks in said first data buffer.

2. The method of claim 1, wherein the set of data blocks corresponds to a file in a persistent file storage system implemented in the NVRAM of the computing device.

3. The method of claim 1, further comprising:

in response to a write request by a second process executing on the computing device to access one or more second data blocks in a second extent of the one or more extents, granting the second process write access by at least;

generating a DRAM-copy of the one or more second data blocks in the DRAM,

storing, in a second buffer header of said plurality of buffer headers, a reference to the DRAM-copy of the one or more second data blocks,

associating the second buffer header with the second process, and

the second process using the second buffer header to access the one or more second data blocks in the DRAM.

4. The method of claim 3, wherein the set of data blocks corresponds to a database stored persistently in the NVRAM, wherein the first process and the second process are database server processes, the method further comprising:

executing, by the second process, one or more update operations of a database transaction by modifying the DRAM-copy of the one or more second data blocks; and

after the database transaction is complete, committing the database transaction.

5. The method of claim 1, further comprising, in response to granting the first process direct read access to the one or more first data blocks, incrementing a first reference count for the one or more first data blocks that indicates a number of processes with direct read access to the one or more first data blocks.

6. The method of claim 5, further comprising modifying a particular data block of the set of data blocks by:

checking a reference count for the particular data block that indicates a number of processes with direct read access to the particular data block; and

updating the particular data block in the NVRAM only after determining that the reference count for the particular data block is zero.

7. The method of claim 1, further comprising, in response to a read request by a third process executing on a remote computing device to access one or more third data blocks belonging to a third extent of the one or more extents in the NVRAM of the computing device, granting the third process read access of the one or more third data blocks by at least:

transmitting the one or more third data blocks to the remote computing device;

generating, by the remote computing device, a DRAM-copy of the one or more third data blocks in the DRAM of the remote computing device;

storing, in a particular buffer header of a plurality of buffer headers in the DRAM at the remote computing device, a reference to the DRAM-copy of the one or more third data blocks at the remote computing device; and

associating the particular buffer header with the third process.

8. The method of claim 1, wherein the set of data blocks corresponds to a file in a distributed file system implemented in NVRAMs of a plurality of computing devices of a distributed computing system that includes the computing device.

9. The method of claim 8, further comprising executing a distributed database system on the plurality of computing devices, wherein the file is a database system file.

10. The method of claim 8, further comprising redistributing at least one extent in the distributed file system, wherein redistributing includes:

selecting a particular extent of the one or more extents to move from the NVRAM of the computing device to NVRAM of a different computing device of the distributed computing system;

checking one or more reference counts for one or more data blocks of a particular extent, said one or more reference counts indicating a number of processes with direct read access to said one or more data blocks of said particular extent; and

moving the particular extent to the NVRAM of the different computing device only after determining that the reference count is zero.

11. One or more non-transitory computer-readable media storing sequence of instructions, wherein the sequences of instructions, when executed by one or more hardware processors, cause:

maintaining a set of data blocks in one or more extents in a non-volatile random-access memory (NVRAM) of a computing device;

allocating a plurality of data buffers in a dynamic random-access memory (DRAM);

allocating a plurality of buffer headers in said DRAM of the computing device, each buffer header of said buffer headers being;

configurable to store a reference to a data buffer of said plurality of data buffers that stores a DRAM-copy of one or more data blocks of said set of data blocks, and

configurable to store a reference to NVRAM of one or more data blocks of said set of data blocks;

determining whether an access request to access one or more first data blocks in a first extent of the one or more extents is a read request, said access request being made by a first process executing on the computing device; and

when said access request is a read request, granting the first process direct read access of the one or more first data blocks in the NVRAM by at least;

storing, in a first buffer header of said plurality of buffer headers, a reference to the one or more first data blocks in the first extent in the NVRAM,

associating the first buffer header with the first process, and

the first process using the first buffer header to directly access the one or more first data blocks in the NVRAM; and

when said access request is not a read request, granting the first process write access by at least;

generating a first DRAM-copy of the one or more first data blocks in a first data buffer of said plurality of data buffers,

storing, in the first buffer header, a reference to the first DRAM-copy of the one or more first data blocks,

associating the first buffer header with the first process, and

the first process using the first buffer header to access the one or more first data blocks in said first data buffer.

12. The one or more non-transitory computer-readable media of claim 11, wherein the set of data blocks corresponds to a file in a persistent file storage system implemented in the NVRAM of the computing device.

13. The one or more non-transitory computer-readable media of claim 11, wherein the sequences of instructions include instructions that, when executed by one or more hardware processors, cause:

in response to a write request by a second process executing on the computing device to access one or more second data blocks in a second extent of the one or more extents, granting the second process write access by at least;

generating a DRAM-copy of the one or more second data blocks in the DRAM,

storing, in a second buffer header of said plurality of buffer headers, a reference to the DRAM-copy of the one or more second data blocks,

associating the second buffer header with the second process, and

the second process using the second buffer header to access the one or more second data blocks in the DRAM.

14. The one or more non-transitory computer-readable media of claim 13,

wherein the set of data blocks corresponds to a database stored persistently in the NVRAM;

wherein the first process and the second process are database server processes;

wherein the sequences of instructions include instructions that, when executed by one or more hardware processors, cause;

executing, by the second process, one or more update operations of a database transaction by modifying the DRAM-copy of the one or more second data blocks; and

after the database transaction is complete, committing the database transaction.

15. The one or more non-transitory computer-readable media of claim 11, wherein the sequences of instructions include instructions that, when executed by one or more hardware processors, cause, in response to granting the first process direct read access to the one or more first data blocks, incrementing a first reference count for the one or more first data blocks that indicates a number of processes with direct read access to the one or more first data blocks.

16. The one or more non-transitory computer-readable media of claim 15, wherein the sequences of instructions include instructions that, when executed by one or more hardware processors, cause modifying a particular extent of the one or more extents by:

checking a reference count for the particular data block that indicates a number of processes with direct read access to the particular data block; and

updating the particular data block in the NVRAM only after determining that the reference count for the particular data block is zero.

17. The one or more non-transitory computer-readable media of claim 11, wherein the instructions include instructions that, when executed by one or more hardware processors, cause, in response to a read request by a third process executing on a remote computing device to access one or more third data blocks belonging to a third extent of the one or more extents in the NVRAM of the computing device, granting the third process read access of the one or more third data blocks by at least:

transmitting the one or more third data blocks to the remote computing device;

generating, by the remote computing device, a DRAM-copy of the one or more third data blocks in the DRAM of the remote computing device;

storing, in a particular buffer header of a plurality of buffer headers in the DRAM at the remote computing device, a reference to the DRAM-copy of the one or more third data blocks at the remote computing device; and

associating the particular buffer header with the third process.

18. The one or more non-transitory computer-readable media of claim 11, wherein the set of data corresponds to a file in a distributed file system implemented in NVRAMs of a plurality of computing devices of a distributed computing system that includes the computing device.

19. The one or more non-transitory computer-readable media of claim 18,

wherein the sequences of instructions include instructions that, when executed by one or more hardware processors, cause executing a distributed database system on the plurality of computing devices;

wherein the file is a database system file.

20. The one or more non-transitory computer-readable media of claim 18, wherein the sequences of instructions include instructions that, when executed by one or more hardware processors, cause redistributing at least one extent in the distributed file system, wherein redistributing includes:

selecting a particular extent of the one or more extents to move from the NVRAM of the computing device to NVRAM of a different computing device of the distributed computing system;

checking one or more reference counts for one or more data blocks of a particular extent, said one or more reference counts indicating a number of processes with direct read access to said one or more data blocks of said particular extent; and

moving the particular extent to the NVRAM of the different computing device only after determining that the reference count is zero.