Create a custom Jupyter Lab

Owned by RosettaHub
Last updated: Jan 11, 2024
2 min read

Step 1: Select the Container Apps perspective then select the Docker Labs view

Step 2: In the view Object Storages, create a new object storage "my-new-storage"

Step 3: Click on the new storage, the AWS console will open in a new tab. Once you access your S3 bucket, you can upload any relevant files (notebooks, datasets). Note that only your bucket will be accessible via through the AWS console and no other services apart from S3 will be accessible.

Step 4: In the view Docker Labs, right click on JupyterLab 3.7.4 then select Clone and name the cloned Docker Lab "my-new-lab".

Step 5: Right-click on the new formation that appeared under Docker Labs and choose Configure.

Step 5.1: Choose the tab Volumes, and select "my-new-storage" under Working Volume.

Step 6: Click on the formation "my-new-lab" to launch choose "Yes", the session will be ready after about two minutes and you'll see a green tick in the upper left corner of the icon, click on the lab to connect to Jupyter. Find the files that you have uploaded to your Object Storage and start working on your data from your Jupyter lab.

Step 6.1: Click on the session, JupyterLab will open in a new tab, you can access the files you uploaded on your storage and install any missing package.

Step 7: Right-click on the session, choose Create Machine Image, Keep " Update Originator Formation On Success" checked. This will snapshot your current working session with all newly installed packages into a new AWS image, your formation will be updated with the new image. Any files that you would have saved in your working directory will be saved into your object storage which was automatically mounted as the working directory of your JupyterLab.

 

 

 

Now the new lab is ready and linked to your new image, next time when you launch the lab it will create a machine based on your new image and your storage will be mounted automatically to your new JupyterLab.

Step 8: You can share the lab with a user, an organization or a group to allow them to reproduce your lab environment.