RESILIENT PROGRAMMING FRAMEWORKS FOR HANDLING FAILURES IN PARALLEL PROGRAMS

• US 20160266983A1
• Filed: 06/25/2015
• Published: 09/15/2016
• Est. Priority Date: 03/13/2015
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: MILTHORPE, Joshua J., IYENGAR, Arun
• Granted Patent: US 9,652,337 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/1438 Restarting or rejuvenating
• CPC Class Codes: G06F 11/1451 by selection of backup contents
• CPC Class Codes: G06F 11/1469 Backup restoration techniques
• CPC Class Codes: G06F 2201/84 Using snapshots, i.e. a logical point-in-time copy of the data

Abstract

An information processing system, computer readable storage medium, and method for supporting resilient execution of computer programs. A method provides a resilient store wherein information in the resilient store can be accessed in the event of a failure. The method periodically checkpoints application state in the resilient store. A resilient executor comprises software which executes applications by catching failures. The method uses the resilient executor to execute at least one application. In response to the resilient executor detecting a failure, restoring application state information to the at least one application from a checkpoint stored in the resilient store, the resilient executor resuming execution of the at least one application with the restored application state information.

Claims

1. A method for supporting resilient execution of computer programs comprising the steps of:

providing a resilient store wherein information in the resilient store can be accessed in the event of a failure;

periodically checkpointing application state in the resilient store;

providing a resilient executor which comprises software which executes applications by catching failures;

using the resilient executor to execute at least one application; and

in response to the resilient executor detecting a failure, restoring application state information from a checkpoint in the resilient store, the resilient executor resuming execution of the at least one application.

2. The method of claim 1, in which the resilient executor further comprises:

an interface allowing applications to use the resilient executor;

a resilient run method which an application invokes via the interface which executes the application while detecting and catching place failures as exceptions, the place failures comprising failures at a place of computation in the application; and

a recovery method which is invoked when the resilient run method catches an exception resulting from a failed place, wherein the recovery method recovers from the place failure, restores the application to application state information from a checkpoint stored in the resilient store, and resumes execution of the application with the application state information restored from the checkpoint stored in the resilient store.

3. The method of claim 1, further comprising:

providing an interface allowing programs to explicitly reference a place to communicate with or execute at least one computation on the place, wherein each place is an entity executing a computation;

providing a virtual place abstraction layer which defines a mapping between virtual places and physical places;

providing an interface allowing an application to communicate with or execute at least one computation on a place p1 by referencing a virtual place p2 which is mapped to physical place p1; and

in response to a physical place p3 failing, wherein virtual place p4 maps to physical place p3, updating the mapping so that virtual place p4 maps to physical place p5 wherein p5 is live.

4. The method of claim 3, in which a place is at least one of a process and at least one thread.

5. The method of claim 1, in which the resilient executor includes a run method which runs an application while catching exceptions.

6. The method of claim 1, in which in response to catching at least one exception, the resilient executor handles the at least one exception, restores computation of the application from a previous checkpoint, and resumes execution of the computation.

7. The method of claim 1, in which the resilient executor invokes application-specific code to restore computation of the application from a previous checkpoint.

8. The method of claim 1, in which the resilient executor includes a method for resiliently executing an iterative computation.

9. The method of claim 1, in which the resilient executor handles an iterative computation by invoking at least one of an application-specific method to execute an iteration of computation of the application and an application-specific method to determine if the computation has finished.

10. The method of claim 1, in which the resilient executor handles at least one exception by recursively catching and handling additional exceptions which occur.

11. The method of claim 1, in which the resilient executor calls an application-specific method to checkpoint data across multiple places.

12. The method of claim 1, in which the resilient executor is provided to an application as an object.

13. The method of claim 12, in which the application invokes a method on the object to resiliently run the application.

14. The method of claim 12, in which the application program invokes a method on the object to resiliently checkpoint the application.