Skip to content

Instantly share code, notes, and snippets.

@kinow
Last active June 26, 2022 05:58
Show Gist options
  • Save kinow/11c5081d9d2e902913d22cfe2a3bb117 to your computer and use it in GitHub Desktop.
Save kinow/11c5081d9d2e902913d22cfe2a3bb117 to your computer and use it in GitHub Desktop.
Running my first experiment with Autosubmit

Here's the step by step that I used to have a running workflow.

  1. Synced master branch
  2. Re-installed dependencies in venv, pip install -U -e .
  3. Removed old files, rm -rf ~/.autosubmit* and rm -rf ~/autosubmit
  4. Configured and installed Autosubmit, autosubmit configure, followed by autosubmit install
  5. Created a new dummy experiment with autosubmit expid -H local --dummy -d DummyExp
  6. Updated my experiment's platform file
# File: /home/kinow/autosubmit/a000/conf/platforms_a000.conf

[local]
TYPE = PS
HOST = localhost
PROJECT = local
USER = root
SCRATCH_DIR = /tmp
  1. Started a Docker container with SSH (didn't want to install OpenSSH), sudo docker run -d -P --name test_sshd -p22:22 rastasheep/ubuntu-sshd:14.04 and copied my SSH pub ID there with ssh-copy-id root@localhost (pass is root)
  2. Ran the experiment autosubmit run a000
  3. Then monitored it, autosubmit monitor a000

Screenshot from 2022-06-26 17-36-58

  1. Changed the chunk size in the experiment definition file
# File: /home/kinow/autosubmit/a000/conf/expdef_a000.conf

#...
NUMCHUNKS = 2
#...

  1. Ran again, and monitored it again

Screenshot from 2022-06-26 17-26-51

@kinow
Copy link
Author

kinow commented Jun 26, 2022

The plot appears to have some issue with how it is displaying the members and chunks 😕

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment