Behavior-based problem localization for parallel file systems

Authors:
Michael P. Kasick;Rajeev Gandhi;Priya Narasimhan
Affiliations:
Electrical & Computer Engineering Department, Carnegie Mellon University, Pittsburgh, PA;Electrical & Computer Engineering Department, Carnegie Mellon University, Pittsburgh, PA;Electrical & Computer Engineering Department, Carnegie Mellon University, Pittsburgh, PA
Venue:
HotDep'10 Proceedings of the Sixth international conference on Hot topics in system dependability
Year:
2010

Citing 7
Cited 1

The Paradyn Parallel Performance Measurement Tool

Computer
Performance debugging for distributed systems of black boxes

SOSP '03 Proceedings of the nineteenth ACM symposium on Operating systems principles
The Tau Parallel Performance System

International Journal of High Performance Computing Applications
Problem diagnosis in large-scale computing environments

Proceedings of the 2006 ACM/IEEE conference on Supercomputing
PVFS: a parallel file system for linux clusters

ALS'00 Proceedings of the 4th annual Linux Showcase & Conference - Volume 4
Black-box problem diagnosis in parallel file systems

FAST'10 Proceedings of the 8th USENIX conference on File and storage technologies
Detecting application-level failures in component-based Internet services

IEEE Transactions on Neural Networks

Automated diagnosis without predictability is a recipe for failure

HotCloud'12 Proceedings of the 4th USENIX conference on Hot Topics in Cloud Ccomputing

Quantified Score

Hi-index	0.00

Visualization

Abstract

We present a behavior-based problem-diagnosis approach for PVFS that analyzes a novel source of instrumentation--CPU instruction-pointer samples and function-call traces--to localize the faulty server and to enable root-cause analysis of the resource at fault. We validate our approach by injecting realistic storage and network problems into three different workloads (dd, IO-zone, and PostMark) on a PVFS cluster.