Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Protocol-Aware Recovery for Consensus-Based Distributed Storage

Protocol-Aware Recovery for Consensus-Based Distributed Storage We introduce protocol-aware recovery (Par), a new approach that exploits protocol-specific knowledge to correctly recover from storage faults in distributed systems. We demonstrate the efficacy of Par through the design and implementation of <underline>c</underline>orruption-<underline>t</underline>olerant <underline>r</underline>ep<underline>l</underline>ication (Ctrl), a Par mechanism specific to replicated state machine (RSM) systems. We experimentally show that the Ctrl versions of two systems, LogCabin and ZooKeeper, safely recover from storage faults and provide high availability, while the unmodified versions can lose data or become unavailable. We also show that the Ctrl versions achieve this reliability with little performance overheads. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png ACM Transactions on Storage (TOS) Association for Computing Machinery

Protocol-Aware Recovery for Consensus-Based Distributed Storage

Download PDF
30 pages

Loading next page...
 
/lp/association-for-computing-machinery/protocol-aware-recovery-for-consensus-based-distributed-storage-2s0Y42Fp2A

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Association for Computing Machinery
Copyright
Copyright © 2018 ACM
ISSN
1553-3077
eISSN
1553-3093
DOI
10.1145/3241062
Publisher site
See Article on Publisher Site

Abstract

We introduce protocol-aware recovery (Par), a new approach that exploits protocol-specific knowledge to correctly recover from storage faults in distributed systems. We demonstrate the efficacy of Par through the design and implementation of <underline>c</underline>orruption-<underline>t</underline>olerant <underline>r</underline>ep<underline>l</underline>ication (Ctrl), a Par mechanism specific to replicated state machine (RSM) systems. We experimentally show that the Ctrl versions of two systems, LogCabin and ZooKeeper, safely recover from storage faults and provide high availability, while the unmodified versions can lose data or become unavailable. We also show that the Ctrl versions achieve this reliability with little performance overheads.

Journal

ACM Transactions on Storage (TOS) – Association for Computing Machinery

Published: Oct 3, 2018

Keywords: Storage faults

References

  • Byzantine disk paxos: Optimal resilience with byzantine shared memory
    Abraham, Ittai; Chockler, Gregory; Keidar, Idit; Malkhi, Dahlia
  • Protocol-aware recovery for consensus-based storage
    Alagappan, Ramnatthan; Ganesan, Aishwarya; Lee, Eric; Albarghouthi, Aws; Chidambaram, Vijay; Arpaci-Dusseau, Andrea; Arpaci-Dusseau, Remzi
  • Correlated crash vulnerabilities
    Alagappan, Ramnatthan; Ganesan, Aishwarya; Patel, Yuvraj; Pillai, Thanumalayan Sankaranarayana; Arpaci-Dusseau, Andrea C.; Arpaci-Dusseau, Remzi H.
  • Kakfa
    Apache,
  • ZooKeeper
    Apache,
  • ZooKeeper Guarantees, Properties, and Definitions
    Apache,
  • Cassandra Replication
    Cassandra, Apache
  • Applications and Organizations using ZooKeeper
    ZooKeeper, Apache
  • Operating Systems: Three Easy Pieces (0
    Arpaci-Dusseau, Remzi H.; Arpaci-Dusseau, Andrea C.
  • An analysis of data corruption in the storage stack
    Bairavasundaram, Lakshmi N.; Arpaci-Dusseau, Andrea C.; Arpaci-Dusseau, Remzi H.; Goodson, Garth R.; Schroeder, Bianca
  • An analysis of latent sector errors in disk drives
    Bairavasundaram, Lakshmi N.; Goodson, Garth R.; Pasupathy, Shankar; Schindler, Jiri
  • Analyzing the effects of disk-pointer corruption
    Bairavasundaram, Lakshmi N.; Rungta, Meenali; Agrawal, Nitin; Arpaci-Dusseau, Andrea C.; Arpaci-Dusseau, Remzi H.; Swift, Michael M.
  • Characteristics, Impact, and Tolerance of Partial Disk Failures
    Bairavasundaram, Lakshmi Narayanan
  • CORFU: A shared log design for flash clusters
    Balakrishnan, Mahesh; Malkhi, Dahlia; Prabhakaran, Vijayan; Wobber, Ted; Wei, Michael; Davis, John D.
  • Grapevine: An exercise in distributed computing
    Birrell, Andrew D.; Levin, Roy; Schroeder, Michael D.; Needham, Roger M.
  • Paxos replicated state machines as the basis of a high-performance data store
    Bolosky, William J.; Bradshaw, Dexter; Haagens, Randolph B.; Kusters, Norbert P.; Li, Peng
  • The chubby lock service for loosely-coupled distributed systems
    Burrows, Mike
  • Paxos made live: An engineering perspective
    Chandra, Tushar D.; Griesemer, Robert; Redstone, Joshua
  • Optimistic crash consistency
    Chidambaram, Vijay; Pillai, Thanumalayan Sankaranarayana; Arpaci-Dusseau, Andrea C.; Arpaci-Dusseau, Remzi H.
  • Consistency without ordering
    Chidambaram, Vijay; Sharma, Tushar; Arpaci-Dusseau, Andrea C.; Arpaci-Dusseau, Remzi H.
  • Upright cluster services
    Clement, Allen; Kapritsos, Manos; Lee, Sangmin; Wang, Yang; Alvisi, Lorenzo; Dahlin, Mike; Riche, Taylor
  • Practical hardening of crash-tolerant systems
    Correia, Miguel; Ferro, Daniel Gómez; Junqueira, Flavio P.; Serafini, Marco
  • Building Large-Scale Internet Services
    Dean, Jeff
  • Dynamo: Amazon’s highly available key-value store
    DeCandia, Giuseppe; Hastorun, Deniz; Jampani, Madan; Kakulapati, Gunavardhan; Lakshman, Avinash; Pilchin, Alex; Sivasubramanian, Swaminathan; Vosshall, Peter; Vogels, Werner
  • Raft TLA+ Specification
    Ongaro, Diego
  • epaxos Source Code
    epaxos,
  • etcd
    etcd,
  • etcd: Production Users
    etcd,
  • Checking the integrity of transactional mechanisms
    Fryer, Daniel; Qin, Dai; Sun, Jack; Lee, Kah Wai; Brown, Angela Demke; Goel, Ashvin
  • Recon: Verifying file system consistency at runtime
    Fryer, Daniel; Sun, Kuei; Mahmood, Rahat; Cheng, TingHao; Benjamin, Shaun; Goel, Ashvin; Brown, Angela Demke
  • Redundancy does not imply fault tolerance: Analysis of distributed storage reactions to file-system faults
    Ganesan, Aishwarya; Alagappan, Ramnatthan; Arpaci-Dusseau, Andrea C.; Arpaci-Dusseau, Remzi H.
  • Redundancy does not imply fault tolerance: Analysis of distributed storage reactions to single errors and corruptions
    Ganesan, Aishwarya; Alagappan, Ramnatthan; Arpaci-Dusseau, Andrea C.; Arpaci-Dusseau, Remzi H.
  • The Google file system
    Ghemawat, Sanjay; Gobioff, Howard; Leung, Shun-Tak
  • Evaluation of applied intra-disk redundancy schemes to improve single disk reliability
    Grawinkel, Matthias; Schafer, Thorsten; Brinkmann, Andre; Hagemeyer, Jens; Porrmann, Mario
  • Building flexible, fault-tolerant flash-based storage systems
    Greenan, Kevin M.; Long, Darrell D. E.; Miller, Ethan L.; Schwarz, Thomas; Wildani, Avani
  • Characterizing flash memory: Anomalies, observations, and applications
    Grupp, Laura M.; Caulfield, Adrian M.; Coburn, Joel; Swanson, Steven; Yaakobi, Eitan; Siegel, Paul H.; Wolf, Jack K.
  • On designing and deploying internet-scale services
    Hamilton, James
  • Data Integrity in Solid State Drives
    Myers, James
  • Silent Data Corruption Is Real
    Goerzen, John
  • Responding to ext4 Journal Corruption
    Corbet, Jonathan
  • Zab: High-performance broadcast for primary-backup systems
    Junqueira, Flavio P.; Reed, Benjamin C.; Serafini, Marco
  • HAFT: Hardware-assisted fault tolerance
    Kuvaiskii, Dmitrii; Faqeh, Rasha; Bhatotia, Pramod; Felber, Pascal; Fetzer, Christof
  • Paxos made simple
    Lamport, Leslie
  • XFT: Practical fault tolerance beyond crashes
    Liu, Shengyun; Viotti, Paolo; Cachin, Christian; Quéma, Vivien; Vukolic, Marko
  • LogCabin
    LogCabin,
  • The SMART way to migrate replicated stateful services
    Lorch, Jacob R.; Adya, Atul; Bolosky, William J.; Chaiken, Ronnie; Douceur, John R.; Howell, Jon
  • Filo: Consolidated consensus as a cloud service
    Marandi, Parisa Jalili; Gkantsidis, Christos; Junqueira, Flavio; Narayanan, Dushyanth
  • A large-scale study of flash memory failures in the field
    Meza, Justin; Wu, Qiang; Kumar, Sanjev; Mutlu, Onur
  • MongoDB Replication
    MongoDB.,
  • There is more consensus in egalitarian parliaments
    Moraru, Iulian; Andersen, David G.; Kaminsky, Michael
  • SSD failures in datacenters: What? When? and Why?
    Narayanan, Iyswarya; Wang, Di; Jeon, Myeongjae; Sharma, Bikash; Caulfield, Laura; Sivasubramaniam, Anand; Cutler, Ben; Liu, Jie; Khessib, Badriddine; Vaid, Kushagra
  • Consensus: Bridging Theory and Practice
    Ongaro, Diego
  • In search of an understandable consensus algorithm
    Ongaro, Diego; Ousterhout, John
  • Data integrity
    Panzer-Steindel, Bernd
  • Application crash consistency and performance with CCFS
    Pillai, Thanumalayan Sankaranarayana; Alagappan, Ramnatthan; Lu, Lanyue; Chidambaram, Vijay; Arpaci-Dusseau, Andrea C.; Arpaci-Dusseau, Remzi H.
  • All file systems are not created equal: On the complexity of crafting crash-consistent applications
    Pillai, Thanumalayan Sankaranarayana; Chidambaram, Vijay; Alagappan, Ramnatthan; Al-Kiswany, Samer; Arpaci-Dusseau, Andrea C.; Arpaci-Dusseau, Remzi H.
  • IRON file systems
    Prabhakaran, Vijayan; Bairavasundaram, Lakshmi N.; Agrawal, Nitin; Gunawi, Haryadi S.; Arpaci-Dusseau, Andrea C.; Arpaci-Dusseau, Remzi H.
  • Redis
    Redis,
  • Redis Replication
    Redis,
  • Introducing CloudLab: Scientific infrastructure for advancing cloud architectures and applications
    Ricci, Robert; Eide, Eric; Team, CloudLab
  • Data Corruption Is Worse than You Know
    Harris, Robert
  • Implementing fault-tolerant services using the state machine approach: A tutorial
    Schneider, Fred B.
  • Understanding latent sector errors and how to protect against them
    Schroeder, Bianca; Damouras, Sotirios; Gill, Phillipa
  • Flash reliability in production: The expected and the unexpected
    Schroeder, Bianca; Lagisetty, Raghav; Merchant, Arif
  • Experience with grapevine: The growth of a distributed system
    Schroeder, Michael D.; Birrell, Andrew D.; Needham, Roger M.
  • RESAR: Reliable storage at exabyte scale
    Schwarz, Thomas; Amer, Ahmed; Kroeger, Thomas; Miller, Ethan L.; Long, Darrell D. E.; Pâris, Jehan-François
  • Arakoon: A distributed consistent key-value store
    Slootmaekers, Romain; Trangez, Nicolas
  • Can ext4 Detect Corrupted File Contents? Retrieved April 21, 2017 from http://stackoverflow
    Stackoverflow,
  • ZooKeeper Clear State
    Stackoverflow,
  • Improving the reliability of commodity operating systems
    Swift, Michael M.; Bershad, Brian N.; Levy, Henry M.
  • Managing update conflicts in Bayou, a weakly connected replicated storage system
    Terry, D. B.; Theimer, M. M.; Petersen, Karin; Demers, A. J.; Spreitzer, M. J.; Hauser, C. H.
  • HARDFS: Hardening HDFS with selective and lightweight versioning
    Do, Thanh; Harter, Tyler; Liu, Yingchao; Gunawi, Haryadi S.; Arpaci-Dusseau, Andrea C.; Arpaci-Dusseau, Remzi H.
  • What to Do when the Journal Checksum is Incorrect
    Ts’o, Theodore
  • Vive la différence: Paxos vs
    Van Renesse, Robbert; Schiper, Nicolas; Schneider, Fred B.
  • Robustness in the Salus scalable block store
    Wang, Yang; Kapritsos, Manos; Ren, Zuocheng; Mahajan, Prince; Kirubanandam, Jeevitha; Alvisi, Lorenzo; Dahlin, Mike
  • End-to-end data integrity for file systems: A ZFS case study
    Zhang, Yupu; Rajimwale, Abhishek; Arpaci-Dusseau, Andrea C.; Arpaci-Dusseau, Remzi H.
  • Unable to Load Database on Disk when Restarting after Node Freeze
    Issues, ZooKeeper Jira