Is it right for you?
A good fit
- Lab and project data that must stay online and shared within the group
- Instrument output, imaging and sequencing libraries
- Data accessed from scripts, pipelines or the cluster over S3
Not a fit
- Data rated P3 or P4
- Real-time document collaboration (use Google Drive)
- Scratch space for running jobs (use HPCC storage)
- Use cases that need a fixed long-term capacity commitment during the pilot
- Mounting as a network drive over NFS or SMB (S3 access only)
- Automatic deletion (lifecycle) rules
What it is
CephRDS is Research Computing's on-campus storage platform for active research data. It is built on Ceph, with the storage spread across campus data centers, and is accessed through an S3-compatible interface. The system was funded in part by an NSF Campus Cyberinfrastructure grant.
What pilot means
CephRDS is in pilot. During the pilot, capacity is limited, terms may change, and allocations are reviewed individually. Pilot participation is not a commitment to a specific capacity, price or term beyond what is agreed in writing for your allocation.
Costs
Rates are under review, not yet approved. Terms are set out when your storage is allocated. See Costs.
How to get access
The PI or lab lead requests storage through Research Computing. KB013: CephRDS onboarding and access walks through the request, the access keys and the first connection.
Connect and use
- KB018: CephRDS with Python (boto3)
- KB019: CephRDS with Cyberduck and other desktop clients
- KB020: Mounting CephRDS buckets as local folders with rclone
Keep in mind
Keep your access keys private. Whether a given allocation is backed up, and how, is part of its terms. Ask before relying on CephRDS as the only copy of irreplaceable data. See backups are not archives.