Scholarlink

A substantial body of research identifies Check/Repair (C/R) as a common initial solution for system recovery in data centers following crashes or malfunctions. While effective, traditional offline C/R methods face limitations, such as data inaccessibility and inefficiencies when addressing growing data volumes or modern hardware capabilities. This paper introduces WFSCK, a parallel C/R approach designed to leverage CPU parallelism and high-bandwidth storage to accelerate file system checks and repairs both offline and online. WFSCK also optimizes Garbage Collection (GC) to reduce system downtime, significantly improving data and system reliability. Experimental results demonstrate a 20%–50% performance enhancement, making WFSCK a valuable advancement in data center management and resilience.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

WFSCK: An Intelligent File System Check Recovery and Task Management Framework

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Juehao Chen,
  • Hang Liu,
  • Shiyi Li,
  • Hang Hui,
  • Qiang Liu,
  • Wen Xia

摘要

A substantial body of research identifies Check/Repair (C/R) as a common initial solution for system recovery in data centers following crashes or malfunctions. While effective, traditional offline C/R methods face limitations, such as data inaccessibility and inefficiencies when addressing growing data volumes or modern hardware capabilities. This paper introduces WFSCK, a parallel C/R approach designed to leverage CPU parallelism and high-bandwidth storage to accelerate file system checks and repairs both offline and online. WFSCK also optimizes Garbage Collection (GC) to reduce system downtime, significantly improving data and system reliability. Experimental results demonstrate a 20%–50% performance enhancement, making WFSCK a valuable advancement in data center management and resilience.