

systemd
systemd is a suite of basic building blocks for a Linux system. It provides a system and service manager that runs as PID 1 and starts the rest of the system.
Cost / License
- Free
- Open Source (GPL-2.0)
Since
- Since 2012
Releases
- 42 releases last year42 last year
GitHub
- GitHub stars: 17.1k
Features
systemd News & Activities
Recent News
- Paul published news article about openSUSE
Recent activities
Pandadad added systemd as alternative to Watchdog PandaWe schedule one-off and recurring jobs with systemd-run --on-calendar instead of cron. Each run gets its own unit, its own logs in journalctl and survives a reboot, and systemctl list-timers shows everything that is due in one place. We also keep SSH tunnels alive as services with Restart=always. Minus one star for the learning curve: unit file syntax, transient units that vanish after they fire, and reset-failed before reusing a unit name all bit us at least once.
- Hamze Zare liked systemd
- humble-advisor liked systemd
We run a small job board as two units on one droplet, a web role and a worker role that re-crawls sources every six hours, and splitting them was the best call we made: a crawler crash no longer takes the site down, and journalctl -u gives each role its own log with zero setup. Restart=on-failure handles most incidents by itself. The catch is that a restart only helps when the process actually exits. A worker stuck on a hung network call stays "active (running)" indefinitely, so we ended up...
- Anton Kylikov liked systemd
Alaik added systemd as alternative to Hummingbird Init







Comments and Reviews
We schedule one-off and recurring jobs with systemd-run --on-calendar instead of cron. Each run gets its own unit, its own logs in journalctl and survives a reboot, and systemctl list-timers shows everything that is due in one place. We also keep SSH tunnels alive as services with Restart=always. Minus one star for the learning curve: unit file syntax, transient units that vanish after they fire, and reset-failed before reusing a unit name all bit us at least once.
We run a small job board as two units on one droplet, a web role and a worker role that re-crawls sources every six hours, and splitting them was the best call we made: a crawler crash no longer takes the site down, and journalctl -u gives each role its own log with zero setup. Restart=on-failure handles most incidents by itself. The catch is that a restart only helps when the process actually exits. A worker stuck on a hung network call stays "active (running)" indefinitely, so we ended up writing our own watchdog that checks a heartbeat instead of trusting the unit state. Drop-in override files are the quiet win: limits and environment sit next to the unit instead of being edited into it.