Research Computing

The /tigress-hsm file system is a large, shared, hierarchical filesystem, available on the head and compute nodes (if any) of Artemis, Della, Hecate, Sesame, Tigressdata and Woodhen. Space usage is restricted. Presently, an application for space is made by sending a request to cses@princeton.edu.

The implementation of this file system uses GPFS (General Parallel File System), an IBM commercial file system specifically designed to support high reliability and computing clusters. There is just over 270 TB of space allocated to /tigress-hsm. The data and files on /tigress-hsm are backed up to a TSM (Tivoli Storage Manager) server.

Although /tigress-hsm presents itself as one file system, it is composed of two tiers. The first (system) tier consists of fibre channel disk (high speed) - all newly created files exist here. SATA (Serial Advanced Technology Attachment) disk makes up the second tier (tier2). If a file on the system tier has not been accessed in 30 days, it is migrated to tier2. This migration is transparent to the end user. After 200 days of residence in tier2, the file is migrated to near line tape. If the file is accessed at any point, it will either be recalled back to the system tier from tier2 or recalled from tape to the system tier. These recall processes are also transparent to the end user.

/tigress-hsm is backed up to PICSciE's private TSM server. Only the latest copy of any file residing on /tigress-hsm is kept on the server, no previous versions of a file are kept. These active files are kept on the backup server indefinitely. Should a file be deleted from /tigress-hsm, it will remain on the backup server for 30 days after which it will be deleted from the backup server.

/tigress-hsm was set up to act as a repository for output or data that requires more processing on another system or cluster. It is not suitable to use as scratch disk or as a home filesystem.