New Hpc Temp Storage
Edit this Page via GitHub Comment by Filing an Issue Have Questions? Ask them here.Who
Research IT and Scientific Computing (RITSC)
What
RITSC now supports a new temporary data storage platform for use by HPC cluster jobs. The system is made by Vast Data, and is a 100% flash storage system. It is clustered and provides increased performance and reliability over our previous temporary data storage platform (called scratch).
How
The new storage platform is mounted on all cluster nodes, including rhinos, at /hpc/temp. In order to provide a high performance shared storage platform for all cluster users, all data written to temp will be deleted 30 days after creation (this is a change from scratch). In order to use this new system, all labs will need to send an email to scicomp@fredhutch.org from lab leadership (data owner/PI/data steward) to have a folder created for read/write access.
When
The storage platform is available now.
Why
Our previous temporary storage platform (scratch) is end of life. It is also an older generation of hardware with deteriorating reliability. It will be retired very soon.