Replace Failed Disk in ZFS Pool - Boot / EFI Partition

ots-24

Member
Jun 13, 2021
4
0
6
I replaced a failed disk (nvme1n1) that was a member of a ZFS mirror on a system running 6.4-8 with the command zpool replace rpool /dev/disk/by-id/xxxxxx /dev/disk/by-id/xxxxxx. The resilver completed successfully, but the "BIOS boot" and "EFI" partitions did not copy to the new disk. Attached are screenshots showing the original and new partition layout. How can I regain the original fully mirrored config, including boot & EFI?

Any help would be greatly appreciated.
 

Attachments

  • Failed Disk.PNG
    Failed Disk.PNG
    64.8 KB · Views: 38
  • New Disk.PNG
    New Disk.PNG
    18.2 KB · Views: 35
Funny, I wanted to ask nearly the same question. I am planning on replacing an SSD from a healthy zfs-mirror, since it is getting close to be worn out. I am planning on adding the new SSD ... resilvering ... at that point it would be a 3-way mirror ... and then removing the old SSD.
 
Funny, I wanted to ask nearly the same question. I am planning on replacing an SSD from a healthy zfs-mirror, since it is getting close to be worn out. I am planning on adding the new SSD ... resilvering ... at that point it would be a 3-way mirror ... and then removing the old SSD.

Let me know how that goes if you don't mind. I've reviewed the documentation at https://pve.proxmox.com/wiki/ZFS_on_Linux#_zfs_administration, but I suspect that since I basically skipped to the zpool replace part, that I need to break the pool and start over. It's a production system that I haven't been able to get another maintenance window on yet and I'm still learning the Linux / Proxmox file system.
 
Did you partition the replacement drive first?
I did not. The only command I issued was zpool replace rpool /dev/disk/by-id/xxxxxx /dev/disk/by-id/xxxxxx.

The output of proxmox-boot-tool status is
Re-executing '/usr/sbin/proxmox-boot-tool' in new private mount namespace.. 7666-6B67 is configured with: uefi WARN: /dev/disk/by-uuid/7666-988E does not exist - clean '/etc/kernel/proxmox-boot-uuids'! - skipping
 
Have a read of the manual
https://pve.proxmox.com/pve-docs/chapter-sysadmin.html#_zfs_administration
I'm not a 100% certain but you should be able to remove a drive then follow the instructions on the manual regarding partitioning (section on failed boot drive replacement). A google should bring up a few clues.
Something along the lines of partition, format partitions, add drive to rpool the sort out the boot using the boot tool.