Method and apparatus for efficiently detecting and logging file system changes

• US 8,055,613 B1
• Filed: 04/29/2008
• Issued: 11/08/2011
• Est. Priority Date: 04/29/2008
• Status: Active Grant

• Current Assignee: NetApp, Inc.
• Original Assignee: NetApp, Inc.
• Inventors: Aravindan, Sajeev, Srinivasan, Chandrasekar, Mu, Paul Yuedong
• Primary Examiner(s): DANG, THANH HA T
• Application Number:
  US12/111,756
• Time in Patent Office: 1,288 Days
• Field of Search: 707/610, 707/615, 707/616, 707/624, 707/637, 707/639, 707/640, 707/646, 707/647, 707/648, 707/649, 707/651, 707/682, 707/685
• US Class Current: 707/610
• CPC Class Codes: G06F 11/1451 by selection of backup contents
• CPC Class Codes: G06F 16/1734 Details of monitoring file system events, e.g. by the use of hooks, filter drivers, logs

Abstract

A method and apparatus for efficiently logging changes to a file system are disclosed. A file system filter driver intercepts I/O requests directed to a file system, and analyzes the I/O requests to determine whether the requests are write or modify requests directed to a file included in a list of files to be monitored for changes. If the I/O request is directed to such a file, the file system filter driver logs the I/O request to a b-tree structured change log file. Furthermore, in one embodiment, the file system filter driver is configured to log file system changes to multiple log files to coincide with snapshot operations, such that the file system changes to be included in each snapshot are included in separate log files.

Claims

1. A computer-implemented method, comprising:

identifying, in real time and at a block level, a change made to a user-selected set of directories or files of a volume of a file system as a result of an I/O request having been processed;

writing to a log file, in real time, block change information indicating one or more blocks from a particular directory or file of a particular volume that have changed as a result of the I/O request having been processed, wherein the log file is structured to store the block change information non-sequentially and grouped by the particular directory or file to which the change relates;

reading block change information from the log file to identify blocks of data to be included in a block-level incremental backup;

responsive to a snapshot event, generating a new log file; and

subsequent to the snapshot event, writing to the new log file, in real time, block change information indicating one or more blocks from the particular directory or file of the particular volume that have changed as a result of an I/O request having been processed after the snapshot event.

2. The computer-implemented method of claim 1, further comprising:

receiving, in real time, data selection criteria indicating a user-selected set of directories or files to be monitored for changes, and adapting to monitor the user-selected set of directories or files without interrupting logging operations.

3. The computer-implemented method of claim 1, wherein the steps of identifying the change and writing the block change information to the log file are performed by a file system filter driver configured to register with a file system process to receive a notification when an I/O request results in a change to the user-selected set of directories or files of the volume of the file system.

4. The computer-implemented method of claim 3, wherein the file system filter driver is configured as a mini-filter integrated to operate with a file system filter manager.

5. The computer-implemented method of claim 3, wherein the log file is structured as a b-tree, and each nodal element of the b-tree corresponds with the particular directory or file of a volume of the file system for which a block-level change has occurred.

6. The computer-implemented method of claim 3, wherein one or more portions of the log file are stored in memory, and said writing to the log file involves writing the block change information to a portion of the log file in memory.

7. A computer system, comprising:

an I/O subsystem configured to service client-initiated I/O requests by writing data blocks to, and reading data blocks from, a disk-based file system;

a file system filter driver configured to i) receive, in real time, data selection criteria specifying a user-selected set of directories or files of the disk-based file system to be monitored for changes, ii) identify when an I/O request has caused a change to one or more blocks included in a directory or file of the user-selected set of directories or files to be monitored, and iii) write block change information to a log file when an I/O request has caused a change to one or more blocks included in a directory or file of the user-selected set of directories or files to be monitored, wherein the log file is structured having the block change information grouped within the log file based on the directory or file to which the change relates;

a backup agent to prepare an incremental backup by identifying only blocks that have changed since a previous backup operation by reading block change information from the log file; and

a snapshot manager configured to generate, in response to a snapshot event, a new log file and subsequent to the snapshot event, to write to the new log file, in real time, block change information indicating one or more blocks from a particular directory or file of a particular volume that have changed as a result of an I/O request having been processed after the snapshot event.

8. The computer system of claim 7, wherein the file system filter driver is configured to adapt, without interruption, to monitor for block-level changes made to files or directories specified in additional data selection criteria received after receiving initial data selection criteria.

9. The computer system of claim 7, further comprising:

a file system filter manager, wherein the file system filter driver is configured as a mini-filter integrated with the file system filter manager, and the mini-filter is configured to register with the file system filter manager to receive notification when an I/O request results in a change to a directory or file included in the user-selected set of directories or files of the file system to be monitored for changes.

10. The computer system of claim 7, wherein the log file is structured as a b-tree, and each nodal element of the b-tree corresponds with a particular directory or file of the user-selected set of directories or files for which a change has been identified.

11. The computer system of claim 7, wherein the file system filter driver is configured to store one or more portions of the log file in memory and to write block change information to the one or more portions of the log file in memory.

12. A computer-implemented method, comprising:

analyzing I/O requests directed to a file system to determine whether the I/O requests satisfy logging criteria;

writing block change information indicating blocks of a directory or file that have changed as a result of processing an I/O request to a log file structured as a b-tree when the logging criteria are satisfied, wherein nodal elements of the b-tree correspond with files or directories for which one or more blocks have changed, and when written to the b-tree the block change information is associated with a nodal element of the b-tree;

prior to writing the block change information to the log file structured as the b-tree, storing the block change information in a queue; and

when the queue contains a predetermined number of block change information entries, writing all the block change information entries in the queue to the log file structured as the b-tree.

13. The computer-implemented method of claim 12, wherein writing the block change information to the log file structured as the b-tree includes writing the block change information to a log file that corresponds with a particular snapshot, wherein the block change information for a first snapshot operation is located in a first log file, and block change information for a second snapshot is located in a second log file.

14. The computer-implemented method of claim 12, wherein the block change information includes i) a volume ID and file index of the directory or file with blocks that have changed, ii) a starting block indicating a block position of a first block in the directory or file that has changed as a result of the I/O request, iii) a number indicating a number of blocks that have changed as a result of the I/O request, and iv) a time indicating the time at which the I/O request was processed and the blocks changed.

15. The computer-implemented method of claim 12, wherein the block change information is a first block change information and a second block change information from a second I/O request directed to the same file in the same volume as the first block change information is stored in a single block record, and an address of a block record is stored in a nodal element of the b-tree that corresponds with the directory or file to which the block information relates.

16. The computer-implemented method of claim 12, wherein the block change information includes block change information from multiple I/O requests directed to the same file in the same volume and the block change information is written to the same nodal element of the b-tree.

17. The computer-implemented method of claim 12, wherein non-leaf nodes of the b-tree are stored in memory and written to disk storage only when writing block change information to the log file results in a structural change to the b-tree.