Rescue mode
How to use Rescue mode to recover broken environments
JupyterHub Rescue mode Image Documentation
Overview
The rescue mode
image is a special-purpose container image designed to help users recover from broken environments that prevent their standard JupyterLab sessions from starting correctly.
This image launches a minimal JupyterLab environment with a temporary, clean $HOME
directory (/rescue-home
) and provides read/write
access to the original home directory via a mounted path (/rescue-home/original-home
).
When Should You Use the Rescue Image?
Use the rescue mode image when:
- JupyterLab fails to start due to:
- You accidentally installed a broken package (e.g., in
.local
) - You edited or corrupted shell configuration files (e.g.,
.bashrc
) - Your user environment is broken due to incompatible extensions or software
- You accidentally installed a broken package (e.g., in
- You want to safely delete or fix files in your original home without interfering with a full JupyterLab session.
What Does the Rescue Mode Image Do?
- Creates a temporary, clean
$HOME
at/rescue-home
- Mounts your original home directory to
/rescue-home/original-home
- Starts a minimal JupyterLab session (no extensions or user settings loaded)
- Lets you safely view, modify, or delete files in your real home
What You Can Do in Rescue Image
- View and delete broken files
- Remove or rename corrupt packages in
.local/
- Reset or move
.bashrc
,.profile
, or.jupyter
folders - Confirm what’s in your original home directory
- Copy files from your original home to rescue-safe space
How to Fix a Broken Environment
- Start the
Rescue mode
image from the image dropdown. - Navigate to
original-home/
. - Identify and fix/remove problematic content:
- e.g.,
rm -rf ~/original-home/.local/lib/python3.12/site-packages/broken_package
. Abrocken package
could be, for example,jupyterlab
and you might have to remove every file and folder starting withjupyterlab
. Please also pay attention to the python version as well and replace it with the version already installed in the jupyterlab einvironment. Sometimes you have to remove everything undersite-packages
so that your environment become clean. - e.g.,
mv original-home/.bashrc original-home/.bashrc.bak
- Log out and re-launch your regular JupyterLab image.
Note: If you want to see the hidden files in the browser, you can activate it via the menu: view/Show Hidden Files