State based configuration failure detection using checkpoint comparison

• US 20040260678A1
• Filed: 06/18/2003
• Published: 12/23/2004
• Est. Priority Date: 06/18/2003
• Status: Abandoned Application

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Wang, Yi-Min, Verbowski, Chad
• Application Number:
  US10/464,118
• Publication Number:
  US 20040260678A1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/008 Reliability or availability analysis
• CPC Class Codes: G06F 11/0706 the processing taking place on a specific hardware platform or in a specific software environment
• CPC Class Codes: G06F 11/079 Root cause analysis, i.e. error or fault diagnosis in a hardware test environment G06F11/22; in a software test environment G06F11/36
• CPC Class Codes: G06F 11/3476 Data logging G06F11/14, G06F11/2205 take precedence

Abstract

A system and method for determining configuration failure root cause of an application uses persistent-state checkpoints. Checkpoints are periodic snapshots of configuration data saved at different points in a machine'"'"'s history. One component in the system compares checkpoints, records configuration file accesses, and tracks change frequencies of configuration file values. Another component is configured to record actions of a user interface and configuration file modifications and to search one or more databases for information related to the configuration failure. The components determine a deviation from known operation of the application without the application specifying configuration settings that control the deviation. The method includes identifying a set of configuration data modified since a last known working state of the application, intersecting the set of configuration data with data associated with access by the application, removing frequently changing configuration data from the intersection, and ranking each entry by likelihood of each entry being the cause of the failure.

Claims

1. A system for determining a root cause for configuration failure of an application, the system including:

a narrow-down component configured to compare a plurality of checkpoints, each checkpoint being any configuration data saved at different points in a machine history, record configuration file accesses, and track change frequencies of configuration file values;

a solution query component coupled to the narrow-down component, the solution query component configured to record actions of the user interface and configuration file modifications and to search one or more databases for information related to the configuration failure, the solution query phase, the narrow-down component and the solution query component operative to determine a deviation from known operation of the application without requiring the application to specify configuration settings that control the deviation.

2. The system of claim 1 wherein the narrow-down component includes:

a tracing tool configured to record configuration file accesses;

a state ranking tool configured to track change frequencies, the state ranking tool determining a relative importance of tracked configuration file values; and

a checkpoint comparison tool for the comparing the checkpoints and determining a reduced set of configuration data items.

3. The system of claim 2 wherein the checkpoint comparison tool includes:

a checkpoint reader configured to load a plurality of checkpoints;

a differencing unit configured to compare and identify differences between configuration data items; and

a storage component to store a resulting set of configuration data items from the differencing unit.

4. The system of claim 2 wherein the tracing tool is configured to track all configuration file values accessed by the application and processes run on behalf of the application.

5. The system of claim 2 wherein the state ranking tool performs an order ranking, the order ranking listing a plurality of configuration file values in order of appearance.

6. The system of claim 2 wherein the state ranking tool performs an inverse change frequency ranking, the inverse change frequency ranking determining configuration file values that are frequent and unlikely to cause configuration failure.

7. The system of claim 6 wherein the frequent configuration file values are found by determining how frequently each configuration file value changes.

8. The system of claim 6 wherein the state ranking tool scores each configuration file value using log(1+N/F_V), wherein N represents a total number of documents representing state differences between pairs of checkpoints, and F_V represents a number of documents in which V appears.

9. The system of claim 8 wherein the state ranking tool sorts an intersection of a set of data from the tracing tool with the reduced set of configuration file values according to the score.

10. The system of claim 2 wherein the state ranking tool periodically checks for newly created checkpoints, invokes the checkpoint comparison tool to compare the newly created checkpoints with existing checkpoints, and updates a database with a new appearance count.

11. The system of claim 1 wherein the checkpoint is a persistent state checkpoint.

12. A method for determining a cause for failure of an application on a machine, the method comprising:

identifying a set of configuration data modified since a last known working state of an application;

intersecting the set of configuration data with a second set of configuration data including data associated with access by the application;

removing frequently changing configuration data from the intersection to provide a reduced set of configuration data; and

ranking each entry in the reduced set of configuration data by likelihood of each entry being the cause of the failure.

13. The method of claim 12 wherein the intersecting the set of configuration data with the second set of configuration data includes data associated with access by each program called by the application.

14. The method of claim 12 wherein the ranking includes tracking change frequencies and determining a relative importance of tracked configuration file values.

15. The method of claim 12 further comprising comparing the reduced set of configuration data with one or more persistent state checkpoints outside the machine.

16. The method of claim 12 wherein the intersecting is performed by a checkpoint comparison tool, the checkpoint comparison tool loading a plurality of configuration file values, comparing and removing duplicate configuration file values and storing a resulting set of configuration file values.

17. The method of claim 12 further comprising tracking all configuration file values accessed by the application and processes run on behalf of the application.

18. The method of claim 12 wherein the ranking includes an order ranking, the order ranking listing a plurality of configuration file values in order of appearance.

19. The method of claim 12 wherein the ranking includes determining configuration file values that are frequent and unlikely to cause configuration failure.

20. The method of claim 19 wherein the frequent configuration file values are found by determining a fraction of a number of configuration file value changes between each pair of consecutive checkpoints.

21. The method of claim 12 wherein the ranking includes scoring each configuration file value using log(1+N/F_V), wherein N represents a total number of documents representing state differences between pairs of checkpoints, and F_V represents a number of documents in which V appears.

22. The method of claim 12 further comprising:

periodically checking for newly created checkpoints;

comparing the newly created checkpoints with existing checkpoints; and

updating a database with a new appearance count based on the comparing the newly created checkpoints.

23. A computer readable medium having computer-executable instructions to perform acts for determining a cause for failure of an application on a machine, the acts comprising:

identifying a set of configuration data modified since a last known working state of an application;

intersecting the set of configuration data with a second set of configuration data including data associated with access by the application;

removing frequently changing configuration data from the intersection to provide a reduced set of configuration data; and

ranking each entry in the reduced set of configuration data by likelihood of each entry being the cause of the failure.

24. The computer readable medium of claim 23 wherein the intersecting the set of configuration data with the second set of configuration data includes data associated with access by each program called by the application.

25. The computer readable medium of claim 23 wherein the ranking includes tracking change frequencies and determining a relative importance of tracked configuration file values.

26. The computer readable medium of claim 23 further comprising comparing the reduced set of configuration data with one or more persistent state checkpoints outside the machine.

27. The computer readable medium of claim 23 wherein the intersecting is performed by a checkpoint comparison tool, the checkpoint comparison tool loading a plurality of configuration file values, comparing a removing duplicate configuration file values and storing a resulting set of configuration file values.

28. The computer readable medium of claim 23 further comprising tracking all configuration file values accessed by the application and processes run on behalf of the application.

29. The computer readable medium of claim 23 wherein the ranking includes an order ranking, the order ranking listing a plurality of configuration file values in order of appearance.

30. The computer readable medium of claim 23 wherein the ranking includes determining configuration file values that are frequent and unlikely to cause configuration failure.