RDA strategy to recover from a major disk failure

General Framework for glade disaster recovery strategy:

  1. DECS: Identify and prioritize what datasets need to be restaged to /glade.
    1. Usage metrics
    2. Datasets with DOIs
    3. Other considerations. - e.g. can legacy datasets be retired from disk?
  2. HPCD: Provide RDA file lists ordered by tape.
    1. DECS provides HPCD with a list of files that need to be recovered from Quasar Tape Storage. HPCD will order this list by tape, and provide the ordered list to DECS for read back processing. By following this strategy, all required data are read off of one tape during one mount.
  3. HPCD: Disaster Recovery Performance Simulations - Yearly simulation every year:
    1. HPCD and DECS will run a yearly tape read back performance test of selected RDA datasets using the strategy listed in “2”. From here, an estimate can be made on how long it would take to restore the full RDA disk copy from tape if needed.
      1. Performance tests will be done yearly.
      2. Estimate how long it would take to repopulate the full RDA on disk.
Disaster Recovery Strategy