How to Run an App as a systemd Service on Linux

The short answer

To keep an app running in the background, starting with the server and recovering after crashes, turn it into a systemd service:

1. Create /etc/systemd/system/myapp.service.
2. Specify the start command, user, working directory and restart policy.
3. Run systemctl daemon-reload and systemctl enable --now myapp.
4. Read logs with journalctl -u myapp.

Starting with nohup, screen or & does not survive a reboot and does not restart a crashed process — systemd handles both.

The unit file

[Unit]
Description=My web app
After=network-online.target
Wants=network-online.target

[Service]
Type=simple
User=myapp
Group=myapp
WorkingDirectory=/opt/myapp
EnvironmentFile=/etc/myapp/env
ExecStart=/usr/bin/node /opt/myapp/server.js
Restart=on-failure
RestartSec=5

[Install]
WantedBy=multi-user.target

Let us go through the key settings.

What each setting means

[Unit]

• Description — a readable name shown in systemctl status.
• After / Wants — start once the network is up. If the app needs a database on the same server, add its service to After=.

[Service]

• ExecStart — the full start command with absolute paths. systemd does not use your shell PATH.
• User / Group — a dedicated system user without root rights. Create it with sudo useradd --system --no-create-home myapp.
• WorkingDirectory — the folder the app resolves relative files from.
• EnvironmentFile — a file with KEY=value lines. Keep secrets there with 600 permissions, not in the unit file itself. For a couple of non-secret values, Environment= is fine.
• Restart — the restart policy.
• RestartSec — a pause before restarting, to avoid a tight crash loop.

[Install]

• WantedBy=multi-user.target — start during normal system boot.

Restart policies

Web services usually use on-failure or always. If a process crashes too often, systemd stops restarting it — tune this with StartLimitIntervalSec and StartLimitBurst in the [Unit] section.

Managing the service

sudo systemctl daemon-reload        # after any unit file change
sudo systemctl enable --now myapp   # start on boot + start now
sudo systemctl status myapp
sudo systemctl restart myapp
sudo systemctl stop myapp

Logs with journalctl

Everything the app writes to stdout and stderr goes to the journal:

journalctl -u myapp -f                  # follow live
journalctl -u myapp -n 100              # last 100 lines
journalctl -u myapp --since "1 hour ago"
journalctl -u myapp -p err              # errors only

Log to stdout — then you do not need separate log files and their rotation.

Common mistakes

• Forgetting daemon-reload — systemd keeps using the old version of the file.
• Relative paths in ExecStart — the service fails with "not found".
• Running as root without a reason.
• Secrets inside the unit file — any user on the system can read it.
• The app daemonizes itself while Type=simple is set — systemd thinks the process exited. Run the app in the foreground.

The full list of settings is in the systemd.service documentation.