~1 min read


HPSS (High Performance Storage System) is a mass file storage system for NERSC. Lots of data outputs of E3SM are stored in HPSS.

  • To see what’s in HPSS, use hsi from any NERSC node

  • To extract file from HPSS to NERSC, need to use zstash

    • load conda and activate the zstash environment
      module load python
      conda create -n zstash_env -c e3sm -c conda-forge zstash`
      source activate zstash_env 
    • then the zstash environment is activated, should see (zstash_env) leading command line
    • to extract files, use
        zstash extract --hpss=<path to HPSS> [files]
  • Normally the files will be tared in a directory named as 000000.tar, 000001.tar … etc. and there is a index.db file describing what files are tared. The .db file is a database file can’t be read directly. Need to extract to nersc directory first.

  • To read the .db file, use following command to see the content of the index.db, or other commands
    sqlite3 zstash/index.db "select * from files;>filelist.txt"
  • Once find the .tar files that contain the E3SM outputs need to be extracted, extract them and then untar them using tar -xvf