We currently have an automated python script which updates end-users with pertinent business information that is currently disabled because we do not have enough disk space in the ML Workbench it is residing in. Can we get some help clearing space?

Forget df. Use du to see the size of files and directories.

Anyway, will be faster going in each directory and deleting old files. Most likely will be old input files, output files, and logs

For example from the home directory, du -h * will show you something like this......

base) jovyan@workbench-794744c89f-6q89n:~$ du -h *

4.0K  DAPpy

12K  Eduardo/__pycache__

60K  Eduardo/.ipynb_checkpoints

132K  Eduardo

48K  examples_ml-suite

180K  examples_pycelonis/data_deduplication

24K  examples_pycelonis/api_tutorial/.ipynb_checkpoints

428K  examples_pycelonis/api_tutorial

24K  examples_pycelonis/root_cause_analysis

636K  examples_pycelonis

84K  GGB/.ipynb_checkpoints

172K  GGB

......

1.4M  logs

16K  lost+found

4.0K  xxxx/Untitled Folder

1.4M  xxxx/.ipynb_checkpoints

3.2M  xxxx

8.0K  Predictions (Beta)

4.0K  work

... and then you can go the directories with more storage used to find files and remove them

HTH

Much appreciated! It turned out to be a cache of files that needed a conda delete call

To save you further searching: the command for that is: conda clean --all in the linux terminal.

It deletes unused packages and conda cache which takes most space.

I've tested and it removed a few GBs, however I'm not sure if it won't brake anything existing scripts (not tested yet). I will provide some update on that later.

Hi Mateusz, I checked and you can do a dry run with '-d'

conda clean -a -d

I checked and the --all option doesn't delete unused packages. It seems only removes tarballs and indexes (-t and -I )

So I am pretty sure you shouldn't have broken anything. I am running it now (It says will free 346 MB)

Yup! Couldn't remember it.