Fail-Slow at Scale

Author:

Gunawi Haryadi S.1,Suminto Riza O.1ORCID,Sears Russell2,Golliher Casey2,Sundararaman Swaminathan3,Lin Xing4,Emami Tim4,Sheng Weiguang5,Bidokhti Nematollah5,McCaffrey Caitie6,Srinivasan Deepthi7,Panda Biswaranjan7,Baptist Andrew8,Grider Gary9,Fields Parks M.9,Harms Kevin10,Ross Robert B.10,Jacobson Andree11,Ricci Robert12,Webb Kirk12,Alvaro Peter13,Runesha H. Birali14,Hao Mingzhe1,Li Huaicheng1

Affiliation:

1. University of Chicago

2. Pure Storage

3. Parallel Machines

4. NetApp

5. Huawei

6. Twitter

7. Nutanix

8. IBM

9. Los Alamos National Laboratory

10. Argonne National Laboratory

11. New Mexico Consortium

12. University of Utah

13. University of California, Santa Cruz

14. University of Chicago Research Computing Center

Abstract

Fail-slow hardware is an under-studied failure mode. We present a study of 114 reports of fail-slow hardware incidents, collected from large-scale cluster deployments in 14 institutions. We show that all hardware types such as disk, SSD, CPU, memory, and network components can exhibit performance faults. We made several important observations such as faults convert from one form to another, the cascading root causes and impacts can be long, and fail-slow faults can have varying symptoms. From this study, we make suggestions to vendors, operators, and systems designers.

Funder

DOE Office of Science User Facility

NSF

Publisher

Association for Computing Machinery (ACM)

Subject

Hardware and Architecture

Reference43 articles.

1. 2011. NAND Flash Media Management Through RAIN. Micron. 2011. NAND Flash Media Management Through RAIN. Micron.

Cited by 48 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

1. DRACO: Distributed Resource-aware Admission Control for large-scale, multi-tier systems;Journal of Parallel and Distributed Computing;2024-10

2. Understanding Silent Data Corruption in Processors for Mitigating its Effects;ACM Transactions on Architecture and Code Optimization;2024-09-02

3. Asymmetric RAID: Rethinking RAID for SSD Heterogeneity;Proceedings of the 16th ACM Workshop on Hot Topics in Storage and File Systems;2024-07-08

4. Chronos: Finding Timeout Bugs in Practical Distributed Systems by Deep-Priority Fuzzing with Transient Delay;2024 IEEE Symposium on Security and Privacy (SP);2024-05-19

5. Detection Is Better Than Cure: A Cloud Incidents Perspective;Proceedings of the 31st ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering;2023-11-30

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3