0

I am getting a bunch of TBs of data files from a remote read-only tape archive into my work directory on my working Linux cluster. Unfortunately I am not allowed to install expect on the computer I am working with. I would like to write a command by which I can submit a slurm job that gets my files through sftp. I need the following sequence

  1. sftp user@remote.server

  2. enter the password: "my_password"

  3. cd /source/folder/where/files/are

  4. set destination folder (if necessary)

  5. get data_files_*.nc

Important is that, it takes days to get this amount of data and I would like to use slurm job option to save time and continuously get the files and avoid keeping connected to the remote server through my laptop the whole time. Any ideas and suggestions are appreciated :)

Behnam
  • 501
  • 1
  • 5
  • 21
  • You can replace expect with [pexpect](https://pexpect.readthedocs.io/en/stable/). I don't know anything about slurm, sorry. – JGK Jan 23 '19 at 21:04
  • Thanks @JGK. As I said, unfortunately I am not allowed to install something new on the supercomputer. – Behnam Jan 23 '19 at 21:21
  • 1
    If what you want is avoid the need for being connected during the transfer, a Slurm job is more complicated than using a terminal multiplexer such as screen or tmux. Is any of those installed? – damienfrancois Jan 24 '19 at 14:59

0 Answers0