Child pages
  • WARNING - Backups
Skip to end of metadata
Go to start of metadata

Please be aware that research data stored on JCU HPC is not backed up in the conventional sense.  A disaster recovery capability exists.  However, due to scale, a large amount of data could be lost in the event of a distaster (e.g., fire in datacentre).  This capability does not provide protection against accidental file deletion or ransomware.

JCU provides (will provide) storage to house the primary copy of research data you require access to.  JCU does not provide resources for housing multiple copies of research data (multiple backups or similar).  Researchers are not permitted to perform regular/routine backups to HPC infrastructure.

If you want to backup research data, or your personal computing device(s), you will need to find and fund another target for such backup requirements.  Researchers wanting additional data protection may want to look into:

  1. https://bitbucket.org/
  2. https://aws.amazon.com/glacier/ - you will need to cover the costs.
  3. OneDrive (JCU Office365 subscription) - suitable for users consuming less than 5TB of disk space.
  4. https://secure.backblaze.com/buy.htm - potentially cheap option for "personal" backups.

IMPORTANT:  Data protection tasks can consume more storage/network bandwidth than most other IT tasks.  Such tasks will reduce performance for research workflows jobs running on HPC infrastructure.

1-Mar-2021:  JCU are likely to invest in cloud storage (AARNet CloudStor?) as a means for long-term housing of JCU research data.  Such an investment is likely to change the current storage configuration on HPC.  E.g., The next generation of HPC cluster storage may be dominated by scratch or short-term storage (home directory quotas could be as low as 20GiB).

  • No labels