Contents

• Contents
• 1 Historical Overview and New Goals
• 2 Pipeline Overview
  • 2.1 Data flow
  • 2.2 General caracteristics of the archive daemons
    • 2.2.1 Source code
    • 2.2.2 Working directory
    • 2.2.3 Parfile
    • 2.2.4 Logfile
    • 2.2.5 Listfile
  • 2.3 Geometry of the pipeline - Master parfile
  • 2.4 Where's what ? A guided tour of the new directory tree of the archive machine kapu
    • 2.4.1 Filesystems
    • 2.4.2 Archive Devices
  • 2.5 The archive user
• 3 Administration and Maintenance
  • 3.1 Checking the pipeline
    • 3.1.1 Interactive checking
    • 3.1.2 Crontab job checking
  • 3.2 Killing and Restarting daemons
    • 3.2.1 Killing daemons
      • Killing a single daemon
      • Killing (stopping) the whole pipeline
    • 3.2.2 Starting daemons
      • Starting a single daemon
      • Starting the whole pipeline
        • Automatic restart
  • 3.3 Changing media
  • Crontab eight_o_clockjob
• 4 Archive Pipeline: detailed review
  • 4.1 copyd daemon
  • 4.2 distd daemon
  • 4.3 Archiver daemons
    • 4.3.1 archd common code
    • 4.3.2 Media handlers
    • 4.3.3 Media initializers
  • 4.4 Miscellaneous Process daemons
• 5 Remodeling and improving the archive pipeline
  • 5.1 Modifying the geometry of the pipeline
    • 5.1.1 Removing an archiver/miscellaneous daemon from the pipeline
      • WARNING:
    • 5.1.2 Adding an archiver/miscellaneous process daemon to the pipeline
  • 5.2 Pipeline tune-up
• 6 Failure: what to do ?
  • 6.1 Failure of type I: a daemon is dead
    • 6.1.1 Case one: the error was detected
    • 6.1.2 Case two: the error was not detected
  • 6.2 failure of type II: a daemon is hung
  • 6.3 Reboot of the archive machine
  • 6.4 Warning from copyd
  • 6.5 Media are full
• 7 Retrieving files from the archive
  • 7.1 From Sony optical disk
  • 7.2 From an Exabyte tape
    • 7.2.1 TAR format
    • 7.2.2 DUMP format
• References
• About this document ...