systemd vs SysV init

PID 1 is the first userspace process and the ancestor of everything else on your system. For decades it was a simple program called init. Today most Linux distributions use systemd, which works very differently. Here's why it changed and how it works.

SysV init — Sequential Shell Scripts

Traditional SysV init worked simply: it read /etc/inittab, determined the runlevel, and executed shell scripts in /etc/rcX.d/ one by one.

# SysV startup sequence (simplified)
/etc/rc0.d/  ← runlevel 0 (halt) scripts
/etc/rc1.d/  ← runlevel 1 (single user) scripts
/etc/rc5.d/  ← runlevel 5 (graphical) scripts

# Scripts named: S##name (start) or K##name (kill)
# Sorted and run in order:
S01networking
S10syslog
S20mysql
S30apache2    ← all run sequentially, one after another

Problems with SysV init:

Sequential = slow. Each service waits for the previous to finish.
Shell scripts are fragile and hard to get right.
No standard way to check service status, restart on failure, or limit resources.
Logs scattered across /var/log/ in different formats.

systemd — Parallel Dependency Graph

What makes systemd fundamentally different? systemd starts services in parallel. It builds a dependency graph of all units and starts as many as possible simultaneously. Instead of "start A then B then C," it's "start A, B, and C in parallel unless B requires A."

Other key systemd features:

Socket activation: systemd creates network sockets on behalf of services, activating them only when a connection arrives. Faster boot, lazy start.
cgroup integration: Every service gets its own cgroup — clean process accounting, no orphaned processes.
Declarative unit files: Replace shell scripts with structured configuration.
Unified logging: journald collects all logs in one searchable binary format.
Automatic restarts: Restart=on-failure in unit file.

Side-by-Side Comparison

	SysV init	systemd
Startup model	Sequential	Parallel dependency graph
Config format	Shell scripts	Declarative .ini-style unit files
Service status	Fragile (exit codes)	Structured (systemctl status)
Logging	Scattered text files	Unified journald binary log
Resource limits	Manual (ulimit in scripts)	Built-in via cgroups
Boot speed	Slow (serial)	Fast (parallel)

Essential systemctl Commands

# Service lifecycle
systemctl start nginx
systemctl stop nginx
systemctl restart nginx
systemctl reload nginx      # reload config without restart

# Enable/disable at boot
systemctl enable nginx      # start on boot
systemctl disable nginx

# Status and inspection
systemctl status nginx      # detailed status + recent logs
systemctl is-active nginx   # active or inactive
systemctl list-units --type=service --state=running

# System targets
systemctl isolate multi-user.target    # switch target
systemctl get-default                  # show default target
systemctl set-default graphical.target