Any way to rebuild missing EFI volume from a ZFS mirrored pair?

[victorhooi]

Hi,

OK, I have a Proxmox 6.4 cluster I set up.

The boot volume is a mirrored ZFS setup, with 2 x 1TB disks (/dev/nvme1n1 and /dev/nvme5n1).

I fat-fingered it, and accidentally ran ceph-volume lvm zap on one of the two disks (/dev/nvme5n1) accidentally *sad face*.

The ZFS volume itself on the disk seems to be intact - ceph-volume hit a resource busy issue.

--> Zapping: /dev/nvme5n1
Running command: /usr/bin/dd if=/dev/zero of=/dev/nvme5n1p2 bs=1M count=10 conv=fsync
 stderr: 10+0 records in
10+0 records out
 stderr: 10485760 bytes (10 MB, 10 MiB) copied, 0.00988992 s, 1.1 GB/s
--> Destroying partition since --destroy was used: /dev/nvme5n1p2
Running command: /usr/sbin/parted /dev/nvme5n1 --script -- rm 2
 stderr: wipefs: error: /dev/nvme5n1p3: probing initialization failed: Device or resource busy
--> failed to wipefs device, will try again to workaround probable race condition
 stderr: wipefs: error: /dev/nvme5n1p3: probing initialization failed: Device or resource busy
--> failed to wipefs device, will try again to workaround probable race condition
 stderr: wipefs: error: /dev/nvme5n1p3: probing initialization failed: Device or resource busy
--> failed to wipefs device, will try again to workaround probable race condition
 stderr: wipefs: error: /dev/nvme5n1p3: probing initialization failed: Device or resource busy
--> failed to wipefs device, will try again to workaround probable race condition
 stderr: wipefs: error: /dev/nvme5n1p3: probing initialization failed: Device or resource busy
--> failed to wipefs device, will try again to workaround probable race condition
 stderr: wipefs: error: /dev/nvme5n1p3: probing initialization failed: Device or resource busy
--> failed to wipefs device, will try again to workaround probable race condition
 stderr: wipefs: error: /dev/nvme5n1p3: probing initialization failed: Device or resource busy
--> failed to wipefs device, will try again to workaround probable race condition
 stderr: wipefs: error: /dev/nvme5n1p3: probing initialization failed: Device or resource busy
--> failed to wipefs device, will try again to workaround probable race condition
-->  RuntimeError: could not complete wipefs on device: /dev/nvme5n1p3

However, I've noticed that the EFI volume (/dev/nvme5n1p2) from that disk (/dev/nvme5n1) is gone:



Firstly - what is the impact of this?

And secondly - is there any way to restore/rebuild this EFI volume please?

Thanks,
Victor

 

[Stoiko Ivanov]

Firstly - what is the impact of this?

This depends on the exact setup of the zpool - if you used the PVE installer and created a RAID1 (mirrored vdev) you should be fine.

* I'd first scrub the zpool to check if the data is still allright (it should repair issues found)
* Afterwards check if the partitions are still allright (fdisk/parted) - compare the partitions to the ones on nvme1n1

the reference documentation on ZFS administration should have all necessary steps:
https://pve.proxmox.com/pve-docs/chapter-sysadmin.html#_zfs_administration

I hope this helps!