0%done
Overall Completion
0 / 53 steps
INITPHASE 1

Bootstrapping, System Management & Process Control

// invariant: you do not own a machine you cannot boot into by hand

Target Chapters

Ch 2Booting & System Management Daemons — UEFI, GRUB2, systemd, targets, emergency mode
Ch 4Process Control — lifecycle, SIGTERM vs SIGKILL, nice/renice, /proc/$PID

Technical Deep Dive — Kernel vs Userspace

The kernel hands control to PID 1 (systemd) which pivots userspace into a target. Unit definitions live in userspace at /etc/systemd/system/ but the *live* runtime graph is materialised under /run/systemd/system/. Never trust a unit file — inspect the running state. A process is a kernel object; its live descriptors, signals and scheduling class are all exposed as text under /proc/[PID].

/etc/systemd/system/
/proc/[PID]/status
/run/systemd/system/
/proc/[PID]/fd/

Interactive Labs & Break-Fix Runbooks

### SENIOR MENTOR DIRECTIVE

$ Stop reading man pages without inspecting live state. If you cannot drop into an emergency root shell from GRUB and remount a read-only filesystem as read-write, you do not own the machine. Never use SIGKILL (9) until you have proven SIGTERM (15) failed to clean up descriptors.