Recovering a failed GPT disk: Difference between revisions

From Notes to self
Jump to navigation Jump to search
No edit summary
No edit summary
Line 5: Line 5:
* Insert a new physical disk
* Insert a new physical disk
* Copy the partition table to the new disk from a healthy one
* Copy the partition table to the new disk from a healthy one
:<code># gpart backup ada0 | gpart restore -F ada5</code>
* Label all partitions with the commands:
* Label all partitions with the commands:
:<code># gpart modify -l LABEL -i1 ada5</code>
:<code># gpart modify -l LABEL -i1 ada5</code>

Revision as of 22:14, 13 December 2025

  • before hot replacing a disk (say ada5) run
# zpool offline zroot ada5
and check
# zpool status
  • Insert a new physical disk
  • Copy the partition table to the new disk from a healthy one
# gpart backup ada0 | gpart restore -F ada5
  • Label all partitions with the commands:
# gpart modify -l LABEL -i1 ada5
  • Fill the EFI partition:
# newfs_msdos -F 32 -c 1 /dev/ada5p1
# mount -t msdosfs /dev/ada5p1 /mnt
# mkdir -p /mnt/EFI/BOOT /mnt/EFI/FREEBSD
# cp /boot/loader.efi /mnt/EFI/BOOT/BOOTX64.efi
# cp /boot/loader.efi /mnt/EFI/FREEBSD/
# umount /mnt
  • Fill the freebsd-boot partition:
# gpart bootcode -b /boot/pmbr -p /boot/gptzfsboot -i 2 ada5
  • Check:
# loaders-update show-me
  • Return ada5p3 to the ZFS pool:
# zpool replace zroot ada5p3