ZFS: Switch Legacy-Boot to Proxmox Boot Tool without having a a second VFAT partition (grub boot)

[git192prox]

ZFS: Switch Legacy-Boot to Proxmox Boot Tool​

https://pve.proxmox.com/wiki/ZFS:_Switch_Legacy-Boot_to_Proxmox_Boot_Tool#3._Finding_potential_ESPs

currently i'm running: pve-manager/6.4-13/9f411e79 (running kernel: 5.4.166-1-pve)

To be best prepared to upgrade to v7.1 i want/need to switch from grub boot to the "proxmox boot tool".

As a prerequisite, i'd need a second VFAT partition (for example /dev/sdd2) with size 512 M.
But currently - i don't:

sdd 8:48 0 894.3G 0 disk
├─sdd1 8:49 0 1007K 0 part
├─sdd2 8:50 0 894.3G 0 part
└─sdd3 8:51 0 8.3M 0 part
sde 8:64 0 894.3G 0 disk
├─sde1 8:65 0 1007K 0 part
├─sde2 8:66 0 894.3G 0 part
└─sde3 8:67 0 8.3M 0 part

Is there any written documentation, how to migrate a current zfs rpool without such a vfat partition, making sure, rpool not get corrupted ?

Thx all to help.

 

[fabian]

that's not really possible. the ESP can be somewhere other than on the disks that contain your rpool though, so if you have other (bootable) disks where you have or can make free space, you can create partition there and format/init it according to the instructions.

 

[git192prox]

Thx Fabian, ok understood.

What is the advantage having more than one ESP ?

To have the chance/possibility to boot the systems from any of the ESP Partitions and so to get some kind of redundancy
(if one of the ESP would fail) ?

 

[Dunuin]

Thx Fabian, ok understood.

What is the advantage having more than one ESP ?

To have the chance/possibility to boot the systems from any of the ESP Partitions and so to get some kind of redundancy
(if one of the ESP would fail) ?

Jup, if your rpool is mirrored and you would only have the ESP on one of those teo disks and the disk with the ESP would fail, your PVE wouldn't be able to boot anymore. So in general you want an ESP on both disks of that mirror and keep them in sync with the proxmox-boot-tool so your server keeps working whatever disk of that mirror might fail.

 

[fabian]

yes, it's basically to reduce downtime / improve the convenience factor in case the disk your primary ESP is on fails (redundant: just select a different one, boot, then fix while everything is up and running vs. single ESP: boot a live system, fix from there, then reboot to get up and running).

the way ESPs are implemented in PVE, they are never original place where things like kernel images/initrds/.. are stored, so if you lose one or all of them, it/they can be re-created from the files on the / partition of your system.

 

[git192prox]

Thx Fabian for your excellent explanation.

 

[Kyle]

Is there any written documentation, how to migrate a current zfs rpool without such a vfat partition, making sure, rpool not get corrupted ?

FWIW, I have undertaken and written up: Switching from legacy boot when there is no space for ESP partition.