Skip to content

09. Bash scripts

Why this matters

Imagine typing 15 commands by hand every time you deploy: git pull, source .venv/bin/activate, pip install -r requirements.txt, python manage.py migrate, python manage.py collectstatic, sudo systemctl restart myapp... It is slow, and sooner or later you will skip something or make a typo.

A Bash script is a file that contains all of these commands in order. You run a single file, and it does everything for you.


Simple explanation

A Bash script is a list of commands written to a file. Instead of typing them by hand one after another, you run a single file and it does everything for you.


Shebang: the first line of a script

#!/usr/bin/env bash

The first line of a script always starts with #! (the shebang). It specifies which program should execute the file.

/usr/bin/env bash means "find bash via env" (a more reliable option than #!/bin/bash, because bash may live in different locations).


Your first script

#!/usr/bin/env bash

echo "Starting Django project setup..."

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
python manage.py migrate
python manage.py collectstatic --noinput

echo "Done!"

Save it to a file named setup.sh, make it executable, and run it:

chmod +x setup.sh
./setup.sh

Variables

#!/usr/bin/env bash

PROJECT_DIR="/var/www/myapp"
VENV_DIR="$PROJECT_DIR/.venv"
APP_USER="www-data"

echo "Project: $PROJECT_DIR"
echo "Virtualenv: $VENV_DIR"

Rules for variables in Bash: - No spaces: NAME="value" ✓, NAME = "value" ✗ - Read them with $: $NAME or ${NAME} - Quotes matter: "$NAME" preserves spaces


Conditions (if/else)

#!/usr/bin/env bash

if [ -f ".env" ]; then
    echo ".env file found"
else
    echo "ERROR: .env file is missing"
    exit 1
fi

Useful checks:

Condition What it checks
[ -f file ] the file exists
[ -d dir ] the directory exists
[ -z "$VAR" ] the variable is empty
[ "$A" = "$B" ] the strings are equal
[ $N -gt 0 ] the number is greater than 0

Exit codes in scripts

#!/usr/bin/env bash

# Stop at the first error
set -e

git pull origin main
source .venv/bin/activate
pip install -r requirements.txt
python manage.py migrate

echo "Deployment successful!"

set -e is a very useful option: if any command returns a non-zero exit code, the script stops. Without it, the script keeps running even after an error.


Loops

#!/usr/bin/env bash

# Run a command for each argument
for app in users notes tags; do
    echo "Running migrations for $app..."
    python manage.py migrate $app
done

# Iterate over files
for file in *.py; do
    echo "File: $file"
done

Command-line arguments

#!/usr/bin/env bash

ENVIRONMENT=${1:-development}   # $1 is the first argument, development by default

echo "Deploying to environment: $ENVIRONMENT"

if [ "$ENVIRONMENT" = "production" ]; then
    echo "Warning: deploying to production!"
fi

Running it:

./deploy.sh production
./deploy.sh            # will use development


A real deployment script

#!/usr/bin/env bash
set -e

echo "=== Deploying the Django project ==="

# Variables
APP_DIR="/var/www/myapp"
VENV="$APP_DIR/.venv"
SERVICE_NAME="myapp"

# Update the code
cd "$APP_DIR"
git pull origin main

# Update dependencies
source "$VENV/bin/activate"
pip install -r requirements.txt --quiet

# Apply migrations
python manage.py migrate --noinput

# Collect static files
python manage.py collectstatic --noinput

# Restart the service
sudo systemctl restart "$SERVICE_NAME"

echo "=== Deployment finished! ==="
echo "Service status:"
systemctl status "$SERVICE_NAME" --no-pager

Debugging a script

bash -x script.sh         # run with tracing (shows every command)
set -x                    # turn on tracing inside the script
set +x                    # turn off tracing

Common beginner mistakes

Mistake 1: Permission denied when running the script

You forgot chmod +x script.sh

Mistake 2: ./script.sh: line 1: $'\r': command not found

The file was saved in Windows format (CRLF). Fix it with: sed -i 's/\r//' script.sh

Mistake 3: The script keeps going after an error and breaks the server

Add set -e at the start of the script

Mistake 4: Spaces around = in variables

NAME="value" ✓, NAME = "value" ✗ — Bash will treat NAME as a command


Practical exercise

Task 1

Write setup.sh for a Django project that: 1. Checks that .env exists 2. Creates a virtualenv (if it does not exist) 3. Installs dependencies 4. Runs migrations

Task 2

Write deploy.sh that: 1. Runs git pull 2. Runs pip install -r requirements.txt 3. Runs python manage.py migrate 4. Runs python manage.py collectstatic --noinput 5. Checks the exit code of each step

Task 3

bash -x setup.sh 2>&1 | head -30
Run the script with tracing and analyze the output.


Self-check

  • [ ] I understand what a shebang is and why it is needed
  • [ ] I can write a script with variables, conditions, and loops
  • [ ] I know that set -e stops the script on an error
  • [ ] I know how to make a script executable with chmod +x
  • [ ] I can write a simple deployment script for Django

Short summary

A Bash script is a set of commands in a file. The shebang specifies the interpreter. set -e stops the script on an error. Real deployment starts with writing a script that automates all the manual steps. The next step is the Makefile, another way to organize commands.