Recovering an Unbootable Arch Linux System with arch‑chroot from a Live ISO

Desired Outcome

Restore a bootable Arch Linux installation that has become unbootable due to a corrupted bootloader, missing initramfs, or mis‑configured /etc/fstab. The goal is to enter the installed system from a live environment, perform targeted repairs, and return to a working boot state.

Prerequisites

• Bootable Arch Linux live ISO (matching the installed system’s architecture, e.g., x86_64) on USB or CD.
• Physical access to the machine and a keyboard/mouse/monitor.
• Knowledge of the disk layout: root, /boot, and EFI System Partition (ESP) if UEFI is used.
• Network connectivity from the live environment if package updates are required.
• Root privileges in the live session.

Focused Procedure

1. Boot the Live Environment

   # Boot from the USB/CD and choose "Boot Arch Linux (x86_64)".
   # Once the live shell appears, you should be root.
   

2. Identify the Partitions

   # List all block devices with filesystem info
   lsblk -f
   # Or use fdisk for a more detailed view
   fdisk -l
   

   Note the device names for the root partition (e.g., /dev/sda2), the ESP (e.g., /dev/sda1), and any separate /boot partition (e.g., /dev/sda3). If LUKS or LVM is in use, unlock or activate them first.

3. Unlock Encrypted Volumes (if applicable)

   # For LUKS
   cryptsetup luksOpen /dev/sda2 cryptroot
   # For LVM
   vgchange -ay
   

4. Mount the Filesystems

   # Mount the root filesystem
   mount /dev/sda2 /mnt
   # If you have a separate /boot
   mount /dev/sda3 /mnt/boot
   # If UEFI, mount the ESP
   mount /dev/sda1 /mnt/boot/efi
   

   Confirm the mounts with ls /mnt and ls /mnt/boot.

5. Bind Mount System Directories

   arch-chroot handles this automatically, but you can do it manually for clarity:

   for dir in /dev /proc /sys /run; do mount --bind $dir /mnt$dir; done
   

6. Enter the Installed System

   arch-chroot /mnt /bin/bash
   # You are now in the installed system’s environment.
   

   Verify that you can run pacman -Qe without errors and that /etc/fstab is readable.

7. Perform Targeted Repairs
   • Update the System

     pacman -Syu
     

   • Reinstall the Kernel and Initramfs

     # Replace linux with the kernel you use (linux, linux-lts, linux-zen, etc.)
     pacman -S linux linux-firmware
     mkinitcpio -P
     

   • Reinstall or Update the Bootloader

     # For systemd-boot (UEFI)
     bootctl update
     # For GRUB (BIOS or UEFI)
     grub-install /dev/sda
     grub-mkconfig -o /boot/grub/grub.cfg
     

   • Check and Fix /etc/fstab

     Ensure UUIDs match the current partitions. Use blkid to confirm and edit if needed.

8. Exit the Chroot and Unmount

   exit
   for dir in /run /sys /proc /dev; do umount /mnt$dir; done
   umount /mnt/boot/efi
   umount /mnt/boot
   umount /mnt
   

9. Reboot

   reboot
   

   Observe the boot sequence. If it reaches a login prompt or display manager, the repair succeeded.

Expected Checks

• After reboot, run journalctl -b | grep -i error to confirm no critical boot failures.
• Confirm that the kernel version matches the installed packages: uname -r.
• Verify that systemctl status systemd-boot (or grub) reports active.
• Check that all filesystems listed in /etc/fstab mount correctly: mount -a inside the system.

Recovery Options If the First Attempt Fails

• Re‑run the live environment, double‑check partition UUIDs, and ensure no hidden LVM or RAID layers were missed.
• Reinstall the bootloader from scratch: remove systemd-boot or grub packages and reinstall them.
• Restore /etc/fstab and /boot from a recent backup or from the installer’s /etc/fstab template.
• If the root filesystem is severely corrupted, consider reinstalling the base and linux packages while preserving /home and other data partitions.

Limitations and Practical Checks

• All commands assume the live ISO architecture matches the installed system. Using a mismatched ISO (e.g., arm instead of x86_64) will fail.
• Mounting the wrong partition can overwrite data. Always verify device names with lsblk before mounting.
• Network access is required only if pacman needs to fetch new packages. If the system is offline, skip the update step.
• After each major change, test the boot by rebooting before proceeding to the next step.

Concrete Example

Assume a UEFI system with the following layout:

Mount Point	Device
root (/)	/dev/sda2
/boot	/dev/sda3
/boot/efi	/dev/sda1

Following the procedure above, the commands would be:

# Mount root
mount /dev/sda2 /mnt
# Mount boot and ESP
mount /dev/sda3 /mnt/boot
mount /dev/sda1 /mnt/boot/efi

# Enter chroot
arch-chroot /mnt /bin/bash

# Reinstall kernel
pacman -S linux linux-firmware
mkinitcpio -P

# Update systemd-boot
bootctl update

# Exit and reboot
exit
umount -R /mnt
reboot

After reboot, the system should boot normally.