Scope: Boot chain, systemd targets, and GRUB configuration on Debian-based and RHEL-based systems
Topics: UEFI and GRUB, kernel and initramfs, systemd targets, boot timing, kernel parameters, boot recovery.

A Linux server boots through a fixed chain: the firmware (UEFI or BIOS) starts the GRUB 2 bootloader, GRUB loads the kernel and initramfs, the kernel hands control to systemd, systemd reaches the default target, and a login prompt appears. Each stage fails in a distinct way, so knowing the chain narrows a boot problem quickly. The steps below cover diagnosis first, then GRUB configuration per distribution family, then recovery.

1. Inspect Boot Timing#

Measure how long the last boot took:

systemd-analyze time

Expected output:

Startup finished in 2.103s (firmware) + 1.842s (loader) + 1.201s (kernel) + 8.412s (userspace) = 13.558s
graphical.target reached after 8.390s in userspace.

Rank the slowest services:

systemd-analyze blame

Inspect the critical path to a target:

systemd-analyze critical-chain

Options:

• time: Shows the duration of each boot stage.
• blame: Lists services sorted by startup time.
• critical-chain: Shows the dependency chain that delayed the target.

2. Select a systemd Target#

Inspect the default target:

systemctl get-default

Expected output:

multi-user.target

Change the default target:

systemctl set-default multi-user.target

List active targets:

systemctl list-units --type target --state active

Switch to a target for the current boot only:

systemctl isolate rescue.target

3. Recover a Broken Boot#

Interrupt the boot at the GRUB menu and press e to edit the selected entry. Find the line starting with linux and append a recovery parameter at the end of it.

Options:

• systemd.unit=rescue.target: Boots into rescue mode with the root filesystem read-write.
• systemd.unit=emergency.target: Boots the smallest possible environment.
• init=/bin/bash: Bypasses systemd entirely and starts a root shell.
• rd.break (RHEL-based): Stops inside the initramfs before the real root is mounted.

Boot the edited entry with Ctrl+x (BIOS) or F10 (UEFI). Changes made this way affect one boot only. When rd.break is used, remount the real root filesystem and reset the root password:

mount -o remount,rw /sysroot
chroot /sysroot
passwd root
touch /.autorelabel
exit
reboot

Inspect the previous boot after a failure or an unattended reboot:

journalctl -b -1
journalctl -b -1 -p err

Options:

• -b -1: Shows the log of the boot before the current one.
• -b: Shows the log of the current boot.
• -p err: Filters messages of priority error and higher.

Part 1: Debian-Based Systems#

Examples: Debian, Ubuntu

1. Inspect the GRUB Configuration#

Inspect the defaults file:

cat /etc/default/grub

Example:

GRUB_DEFAULT=0
GRUB_TIMEOUT=5
GRUB_DISTRIBUTOR=`lsb_release -i -s 2> /dev/null || echo Debian`
GRUB_CMDLINE_LINUX_DEFAULT="quiet"
GRUB_CMDLINE_LINUX=""

2. Set Kernel Parameters#

Edit /etc/default/grub. Parameters that apply to every boot belong in GRUB_CMDLINE_LINUX:

GRUB_CMDLINE_LINUX="net.ifnames=0 console=ttyS0,115200n8"

Options:

• GRUB_CMDLINE_LINUX: Applied to all entries, including recovery entries.
• GRUB_CMDLINE_LINUX_DEFAULT: Applied to the default entries only. Removing quiet shows full boot messages.
• GRUB_DEFAULT=0: Boots the first menu entry. Use a menu index or a title string to select another entry.
• GRUB_TIMEOUT: Seconds the menu is displayed before the default entry boots.

3. Regenerate the GRUB Configuration#

Run after every change to /etc/default/grub:

update-grub

Expected output:

Generating grub configuration file ...
Found linux image: /boot/vmlinuz-6.1.0-18-amd64
Found initrd image: /boot/initrd.img-6.1.0-18-amd64
done

update-grub is a wrapper for grub-mkconfig -o /boot/grub/grub.cfg. Never edit /boot/grub/grub.cfg by hand.

Part 2: RHEL-Based Systems#

Examples: RHEL, Rocky Linux, AlmaLinux

1. Inspect the GRUB Configuration#

Inspect the defaults file:

cat /etc/default/grub

Example:

GRUB_TIMEOUT=5
GRUB_DISTRIBUTOR="$(sed 's, release .*$,,g' /etc/system-release)"
GRUB_DEFAULT=saved
GRUB_DISABLE_SUBMENU=true
GRUB_CMDLINE_LINUX="crashkernel=1G-4G:192M,4G-64G:256M,64G-:512M rhgb quiet"
GRUB_DISABLE_RECOVERY="true"
GRUB_ENABLE_BLSCFG=true

2. Set the Default Entry and Kernel Parameters#

With GRUB_DEFAULT=saved, the default entry is stored in the GRUB environment block. Set it explicitly:

grub2-set-default <MENU_INDEX_OR_TITLE>
grub2-editenv list

Alternatively, set GRUB_DEFAULT=0 in /etc/default/grub to always boot the first entry.

Edit GRUB_CMDLINE_LINUX to add parameters for every boot. Remove rhgb quiet on troubleshooting systems to see full boot messages:

GRUB_CMDLINE_LINUX="crashkernel=1G-4G:192M,4G-64G:256M,64G-:512M"

3. Regenerate the GRUB Configuration#

Run after every change to /etc/default/grub:

grub2-mkconfig -o /boot/grub2/grub.cfg

On RHEL 9 and its rebuilds, /boot/grub2/grub.cfg is the authoritative configuration for BIOS and UEFI systems. Older layouts keep the file on the EFI system partition at /boot/efi/EFI/<VENDOR>/grub.cfg; check the path with:

find /boot -name grub.cfg

Kernel entries are managed as BLS entries in /boot/loader/entries/. Never edit grub.cfg or the entry files by hand.

Part 3: Important Notes#

Boot Stage Reference#

Command Reference#

Security Guidelines#

• Keep console or out-of-band access open before changing GRUB; a broken configuration can leave the system unreachable.
• Always regenerate grub.cfg with the distribution’s own tool: update-grub on Debian-based systems, grub2-mkconfig on RHEL-based systems.
• Never edit grub.cfg directly; the changes are lost on the next regeneration and are not validated.
• Protect the GRUB menu with a password on physically accessible systems so that kernel parameters cannot be edited at boot.
• Remember that e at the GRUB menu and rd.break grant root access; lock down console access accordingly.
• Keep at least one older kernel installed and bootable for rollback after a bad kernel or driver update.
• Do not remove recovery entries or the initramfs images; they are the fastest path back into a broken system.
• Restrict write access to /boot, /etc/default/grub, and /boot/loader/entries/ to root.
• After changing the default target or kernel parameters, reboot once and verify with systemctl get-default and systemd-analyze time.
• Use journalctl -b -1 to investigate unexpected reboots before cleaning old journals.