Skip to content

Latest commit

 

History

History
39 lines (30 loc) · 1.83 KB

SAGEMAKER.md

File metadata and controls

39 lines (30 loc) · 1.83 KB
Raw

Running on AWS SageMaker

If you're using AWS SageMaker, make sure what version of JupyterLab it is running. If you are not sure, you can open a terminal and running the following command:

jupyter --version

Look for the line with the name "jupyterlab".

If the version is JupyterLab 1.x or 2.x:

  1. Stop and start your sagemaker instance (to make sure you're starting fresh)
  2. Open a terminal, and run source activate JupyterSystemEnv to switch to JupyterLab's conda environment
  3. Run jupyter labextension install jupyterlab-s3-browser to install the lab extension
  4. Run pip install jupyterlab-s3-browser to install the server extension
  5. Run jupyter serverextension enable --py jupyterlab_s3_browser to make sure the server extension is enabled
  6. Run sudo initctl restart jupyter-server --no-wait to restart your jupyterlab server
  7. Refresh the page

If the version is JupyterLab 3.x:

  1. Stop and start your sagemaker instance (to make sure you're starting fresh)
  2. Open a terminal, and run conda activate studio to switch to JupyterLab's conda environment
  3. Run pip install jupyterlab-s3-browser to install the server extension
  4. Run jupyter serverextension enable --py jupyterlab_s3_browser to make sure the server extension is enabled
  5. Run restart-jupyter-server to restart your jupyterlab server
  6. Refresh the page

You should now have the bucket icon on your sidebar. Use https://s3.amazonaws.com as your endpoint. Enter your access key and secret key generated on this page: https://console.aws.amazon.com/iam/home#security_credential.

You'll need to perform these instructions every time you log in, because SageMaker doesn't save the state of your installed extensions. However, you can create a lifecycle configuration that executes those commands (except restarting the server) every time an instance is created