Customizing Your Shell Environment

Every SciServer Compute container comes with a default configuration (aliases, paths, modules). However, because containers are ephemeral, any changes you make to ~/.bashrc inside a running container are lost as soon as you stop it.

To solve this, SciServer automatically sources a Persistent Bashrc file from your User Storage at the end of every login session.

How it Works

When you open a terminal or launch a JupyterLab session, the container runs the following logic at the very end of its startup sequence:

if [ -f $HOME/workspace/Storage/$SCISERVER_USER_NAME/persistent/.bashrc ]; then
    . $HOME/workspace/Storage/$SCISERVER_USER_NAME/persistent/.bashrc
fi

What this means for you:

If you create a file named .bashrc inside your persistent storage, it will be automatically loaded.
Because it runs last, it overrides any default aliases or variables set by the system.
It persists across all containers and images you use.

Step-by-Step Setup

1. Create the file

Open a terminal inside a Compute container and create the file using your preferred editor (or create it via the jupyterlab interface) at this specific location:

nvim $HOME/workspace/Storage/$SCISERVER_USER_NAME/persistent/.bashrc

2. Add your customizations

Here are common examples of what you might want to add.

A. Permanent Aliases

Save typing by creating shortcuts for long commands.

# .bashrc content

# Shortcut to activate your persistent mamba environment
alias myenv='mamba activate ~/workspace/Temporary/$SCISERVER_USER_NAME/scratch/envs/my-analysis'

# Shortcut to print "hello {name}"
alias sayhello='echo "hello $SCISERVER_USER_NAME"'

B. Environment Variables

Set default paths or API keys.

# .bashrc content

# Tell Python to look in your persistent storage for modules
export PYTHONPATH="$HOME/workspace/Temporary/$SCISERVER_USER_NAME/scratch/custom_libs:$PYTHONPATH"

C. Auto-load Mamba Environments

If you want a specific environment to be active every time you open a terminal:

# .bashrc content
# Activate environment silently
mamba activate ~/workspace/Temporary/$SCISERVER_USER_NAME/scratch/envs/my-analysis 2>/dev/null

D. Automatic SSH Setup

If you follow the Accessing Containers via SSH guide, you can avoid manually creating the symlink every time you launch a new container by adding this logic to your persistent .bashrc:

# .bashrc content

# Automatically restore SSH symlink if persistent storage exists
if [ -d "$HOME/workspace/Storage/$SCISERVER_USER_NAME/persistent/.ssh" ]; then
    ln -s "$HOME/workspace/Storage/$SCISERVER_USER_NAME/persistent/.ssh" "$HOME/.ssh" 2>/dev/null
fi

3. Save and Reload

Save the file and exit the editor.

To test it immediately without restarting the container, run:

source $HOME/workspace/Storage/$SCISERVER_USER_NAME/persistent/.bashrc

Verify your aliases work (e.g., type sayhello).

Important: Precedence and Conflicts

Because this file is sourced at the end of the container’s startup, your settings take precedence over the system defaults.

Example: Overriding System Aliases

If the system defines:

alias ls='ls --color=auto'

And you add this to your persistent .bashrc:

alias ls='ls -G'

Your version (ls -G) will win.

Troubleshooting

If your container behaves strangely after adding this file (e.g., commands not found):

Rename the file temporarily to disable it:

mv $HOME/workspace/Storage/$SCISERVER_USER_NAME/persistent/.bashrc $HOME/workspace/Storage/$SCISERVER_USER_NAME/persistent/.bashrc.bak

Start a new terminal session to confirm the issue is gone.
Edit the backup file to fix the syntax, then rename it back to .bashrc.

Use Case: Linking to Persistent Environments

This feature works perfectly with the Persistent Environments guide.

Instead of remembering the full path to your environment every time, add this to your persistent .bashrc:

# Auto-register your persistent kernel on login
if [ -f ~/workspace/Temporary/$SCISERVER_USER_NAME/scratch/envs/my-analysis/bin/python ]; then
    python -m ipykernel install --user --name=my-analysis --display-name "Python (My Analysis)" 2>/dev/null
    echo "Persistent kernel 'my-analysis' ready."
fi

Now, every time you launch a container, your Jupyter kernel is automatically available.