Method for implementing persistent file pre-allocation

• US 8,639,656 B2
• Filed: 02/02/2007
• Issued: 01/28/2014
• Est. Priority Date: 02/02/2007
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Bhattacharya, Suparna, Cao, Mingming, Kleikamp, David J., Tso, Theodore Y. T.
• Primary Examiner(s): BETIT, JACOB F
• Assistant Examiner(s): Hocker, John P
• Application Number:
  US11/670,539
• Publication Number:
  US 20080189342A1
• Time in Patent Office: 2,552 Days
• Field of Search: 707 1- 10, 707/205, 707/609, 707/694
• US Class Current: 707/609
• CPC Class Codes: G06F 16/10 File systems; File servers

Abstract

A method and apparatus are provided for pre-allocating non-initialized blocks of data in a file system. Each file contains a primary i-node to maintain file metadata. A secondary i-node is created and associated with the primary i-node. The secondary i-node contains non-initialized data blocks, and the primary i-node contains valid and initialized data blocks. A write operation to a non-initialized data block in a pre-allocated range stored in the secondary i-node directs the data block from the secondary i-node to the primary i-node.

Claims

1. A computer implemented method for persistent pre-allocation of data blocks comprising:

maintaining a first i-node for a file to manage file operations on a file server, said first i-node having initialized blocks;

associating said first i-node with a second i-node, including creating said second i-node for said file, establishing a relationship between primary and secondary nodes for the file, and allocating pre-allocated blocks into the second i-node; and

using a processor for transferring corresponding blocks from said second i-node to said first i-node responsive to a write operation requiring non-initialized blocks in a pre-allocated range of blocks, wherein said block transfer includes a contiguous block allocation for said file.

2. The method of claim 1, further comprising creating said second i-node and associating said second i-node with said first i-node if said second i-node is absent for a specified file.

3. The method of claim 2, further comprising allocating a range of specified blocks for said second i-node.

4. The method of claim 1, wherein the step of transferring corresponding blocks from said second i-node to said first i-node includes transferring a mapping for said blocks from said second i-node to said first i-node.

5. The method of claim 4, wherein the step of transferring blocks from said second i-node to said first i-node includes validating said second i-node prior to said transfer.

6. The method of claim 1, wherein said mapping technique includes system extended attributes.

7. A computer system comprising:

a processor in communication with memory;

a file system to log metadata;

a file having a first i-node in the file system to manage file operations, said first i-node having initialized blocks;

an allocation manager in the file system to associate the first i-node with a second i-node, including creation of the second i-node for the file, to establish a relationship between first and second i-nodes for the file, and allocation of pre-allocated blocks in the second i-node, and

said allocation manager to direct transfer of blocks from said second i-node to said first i-node in response to a requirement for non-initialized blocks in a pre-allocated range of blocks present on said second i-node in a write operation, wherein said block transfer includes a contiguous block allocation for said file.

8. The system of claim 7, further comprising a file system manager to create said second i-node for a specified file and to associate said second i-node with said first i-node if it is determined that said second i-node is absent for a specified file.

9. The system of claim 8, wherein said file system manager allocates a range of specified blocks for said created second i-node.

10. The system of claim 7, wherein said allocation manager transfers corresponding 2 blocks from said second i-node to said first i-node, and transfers a mapping for said blocks from said second i-node to said first i-node.

11. The system of claim 10, wherein said allocation manager validates presence of said second i-node in said file system prior to said transfer.

12. The system of claim 7, wherein said mapping technique includes system extended attributes.

13. An article comprising:

a tangible computer readable data storage media including computer program instructions configured to pre-allocate data blocks to a file in conjunction with a first i-node to manage file operations and a second i-node having pre-allocated non-initialized blocks, comprising;

instructions to associate said first i-node with said second i-node through a mapping technique; and

instructions to transfer corresponding blocks from said second i-node to said first i-node in response to a write operation requiring non-initialized blocks in a pre-allocated range of blocks absent from said first i-node, wherein said block transfer includes a series of blocks having a continuous connection.

14. The article of claim 13, further comprising instructions to create said second i-node and to associate said second i-node with said first i-node if said second i-node is absent for a specified file.

15. The article of claim 14, further comprising instructions to allocate a range of specified blocks for said second i-node.

16. The article of claim 13, wherein the instructions to transfer corresponding blocks from said second i-node to said first i-node includes transferring a mapping for said blocks from said second i-node to said first i-node.

17. The article of claim 16, wherein the instructions to transfer blocks from said second i-node to said first i-node includes instructions to validate said second i-node prior to said transfer.

18. The article of claim 13, wherein said mapping technique includes system extended attributes.

19. A computer implemented method for persistent pre-allocation of data blocks comprising:

maintaining a primary i-node for a file to manage file operations on a file server, said primary i-node having a contiguous set of mapped data blocks;

for said file backing said primary i-node with a first level secondary i-node for mapping initialized data blocks in a pre-allocated range;

backing said first level secondary i-node with a second level secondary i-node for mapping pre-allocated non-initialized data blocks;

associating said primary i-node with said first level secondary i-node, and associated said first level secondary i-node with said second level secondary i-node through a mapping technique; and

using a processor transferring corresponding data blocks from said second level secondary i-node to said first level secondary i-node, and transferring a contiguous set of initialized pre-allocated data blocks from said first level secondary i-node to said primary i-node responsive to a write operation.