We Need to Talk About Systemd: Boot Time Optimization for the New init daemon
Talkat Open Source Summit Europe + Embedded Linux Conference Europe 2019 · on Tuesday, 2019-10-29· at 11:30· Forum 2·
summary
Systemd defaults often lengthen boots because unused units and unnecessary ordering hold the critical path; diagnosing with systemd-analyze critical-chain and systemctl enable/disable exposes the offenders. On a PocketBeagle example user‑space boot drops from ~47s to ~12s (≈35s saved) by switching the default target from graphical to multi-user, removing a hanging getty@ttyGS0 and disabling unneeded services. Also shows using watchdog settings and CPUQuota to limit and auto‑restart misbehaving services.
AI-written summary · about
chapters (22)
- 0:00Intro
- 0:36About Chris Simmonds
- 2:32What is an init daemon?
- 3:53Init daemons for embedded use cases
- 5:03Systemd is not just an init daemon
- 6:30What are the advantages of systemd?
- 8:24Units, services and targets
- 11:27Unit dependencies
- 12:23Order: Before and After
- 15:44The default target
- 16:14Reverse dependencies: WantedBy
- 17:02The Install section
- 17:48systemcti
- 19:14Reducing boot time
- 21:16Measuring systemd boot time
- 22:31First attempt
- 23:03systemd-analyze 3/3
- 25:002nd attempt
- 26:45Other useful systemd features
- 27:10Watchdog
- 28:14Resource limits
- 35:03systemd-analyze 2/3
related talks
37mAccelerating Linux Boot Time: Techniques and Strategies for Optimal Performance
40mA Pragmatic Guide to Boot-Time Optimization
53mReducing the boot time of Linux devices
36mSystemd, from /dev/null to root
23mReboot Without Rebooting: Updating Linux Userspace at the Speed of a Service Restart
16mlightning talk, CentOS Dojo, DevConf.us
01h 45mDebootstrapping Embedded Linux Root Filesystems - Building and Optimization for the...
54mSystemd-Nspawn is Chroot on Steroids