TL;DR

First step after editing any unit file: systemctl daemon-reload, or systemd keeps using the old config. systemctl restart = stop+start; systemctl reload sends SIGHUP without interrupting the process (service must support it). In scripts use systemctl is-active svc (exit code 0=active, 3=inactive, 4=unknown) — never parse status output text. Top pitfalls: ① ExecStart does not support pipes/redirection/$VAR — wrap shell syntax as /bin/sh -c '...'; ② % is an escape character in units — write literal %%; ③ journald's default size limit is 10% of disk or 4G (whichever is smaller), it rotates automatically, but journalctl --vacuum-size=200M frees space immediately.

Service Lifecycle Command Table

Common Query Commands

Core Unit Directives (copy-paste for .service files)

systemd vs SysVinit Comparison

Copy-Paste Full Unit (Python service)

# /etc/systemd/system/myapp.service
[Unit]
Description=My Python Web API
After=network-online.target mysql.service
Wants=mysql.service

[Service]
Type=simple
User=www-data
Group=www-data
WorkingDirectory=/opt/myapp
EnvironmentFile=/etc/myapp.env
ExecStart=/usr/bin/python3 -m uvicorn main:app --host 127.0.0.1 --port 8000
Restart=on-failure
RestartSec=3

[Install]
WantedBy=multi-user.target

# Deploy a service (with permissions + reload)
sudo cp myapp.service /etc/systemd/system/myapp.service
sudo chmod 644 /etc/systemd/system/myapp.service
sudo systemctl daemon-reload
sudo systemctl enable --now myapp
# Verify
systemctl is-active myapp        # prints active, exit code 0
systemctl status myapp --no-pager | head -n 15
journalctl -u myapp -n 50 --no-pager   # last 50 log lines

# Check service state in a script (use exit codes, not status text)
import subprocess
r = subprocess.run(["systemctl", "is-active", "myapp"], capture_output=True, text=True)
print(r.returncode, r.stdout.strip())   # 0 active / 3 inactive / non-zero failed

# Type=notify with sdnotify (systemd considers the service up only after READY=1)
try:
    import sdnotify
    n = sdnotify.SystemdNotifier()
    n.notify("READY=1")
except ImportError:
    pass  # still works when running outside systemd (local dev)

High-Frequency Pitfall Troubleshooting

FAQ

Why don't my systemd service config changes take effect?

Almost always because you skipped systemctl daemon-reload. systemd loads units into memory at startup; after editing a file you must run daemon-reload, then systemctl restart svc for the new config to apply.

Is "active (exited)" in systemctl status normal?

Yes. exited means the main process ran and finished as expected (one-shot tasks, Type=oneshot init scripts) — it is as healthy as active (running). Only failed is abnormal.

How do I make a service start on boot?

sudo systemctl enable svc (or enable --now to also start it immediately). The unit needs an [Install] section with WantedBy=multi-user.target or enable will error. Use disable to remove autostart, mask to fully block it.

journalctl logs are eating the disk — what do I do?

Run journalctl --vacuum-size=200M to free space immediately. To cap it long-term, set SystemMaxUse=200M in /etc/systemd/journald.conf and systemctl restart systemd-journald. The default limit is 10% of the disk or 4G, whichever is smaller.

My service keeps failing to start — fastest way to debug?

systemctl status svc for the error and recent logs → journalctl -u svc -n 50 --no-pager for the full picture → run the ExecStart command manually to prove the command itself works (mind environment variables and working-directory differences).

Sources

• systemd official docs (freedesktop.org/software/systemd), verified 2026-08-11
• man pages: systemctl / systemd.service / journalctl
• Arch Linux Wiki: systemd (wiki.archlinux.org/title/Systemd), verified 2026-08-11

最后更新：2026-08-11