Scalable and Accurate Application-Level Crash-Consistency Testing(…)
summary
Implements representative testing as an update-behavior heuristic and builds Pathfinder to trace POSIX and MMIO executions, construct a persistent graph of operations, group semantically related updates, and generate representative crash states. Claims a 90–99% crash-state reduction, finds 18 bugs (7 new) across 8 production systems in evaluation, and reports finding 4× more POSIX bugs and 8× more MMIO bugs than prior tools while scaling to larger applications.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 0:23Paper overview
- 1:07Crash consistency problem
- 3:39Prior testing approaches
- 5:05State equivalence in crashes
- 8:17Update behaviors testing
- 11:06Pathfinder overview
- 12:30Pix update behaviors
- 13:13MMIO update behaviors
- 14:18Grouping mechanisms
- 16:03Crash testing and bugs
- 16:48Pathfinder results
- 17:33Heuristic state reduction
- 18:18Q&A
related talks
12mUnderstanding and Finding Crash-Consistency Bugs in Parallel File Systems
15mTowards Verifying Crash Consistency
38mTesting with two failure seeking missiles: fuzzing and property based testing
29mScaleCheck: A Single-Machine Approach for Discovering Scalability Bugs...
18mPreview: Verification (Correctness & Debugging)
42mSurprise yourself with Property Based Testing
01h 00mContinuous Regression Testing for Safer and Faster Refactoring in C++
41m0Flake - Reaching reliable non-flaky tests