Storage

• The login nodes are connected directly to the RIS Storage Platforms through NFS.

• The computation nodes are connected directly to the RIS Storage Platforms through NFS.

• There may be a slight delay (up to 5 seconds) of viewing changes made on one connection via another connection. (e.g. Between execution hosts)

• Storage Platforms are NOT designed as a place to share office docs.

• The path for the Storage Platforms on Compute2 is the same as Compute1.

  /storageN/fs1/allocation_name/Active

Scratch2

• Scratch2 space is considered a “global temporary space”. Users should not store their only copy of data in scratch.

• Each group will receive a 10TB allocation in Scratch2.

• Files and directories will be automatically deleted from /scratch2 after 28 days, based on creation time.

• There is no notification for a group that is near the limit of their quota, old data will be deleted to make room.

• The path for scratch2 on Compute2 follows the same scheme as Compute1.

  /scratch2/fs1/allocation_name

• Groups can monitor their quota through use of the df command.

  df -h /scratch2/fs1/ris