I have a small test system that was using cheap SSDs (just 2 disks mirror, ZRAID-1), and both disks failed in a short time period. I had access to it after the event I am describing below, someone else proceeded to replace disks, etc. as follows :
I suspect we should've copied partitions first with sgdisk first. Is this correct ? Would grub-install to a new disk after re-silvering have the same effect ?
The documentation indicates this clearly, I am just no sure why resilvering would have worked with a new disk.
I am a bit confused as I thought the resilvering would take care of boot partitions too, the new disks had not been formatted or partitioned before.
I'd also like to understand if there is a difference between using only grub or the proxmox-boot-tool by itself. I'll explore more in the host boooloader docs.
- When the first disk failed it was replaced and re-silvering was fine. At some point the system was rebooted fine.
- When the second disk failed, resilvering went fine too. Eventually the system was restarted and upon reboot it wouldn't start. Only zpool offline / replace commands were used.
- This being a test system it was reinstalled before I got to further check logs, etc.
The documentation indicates this clearly, I am just no sure why resilvering would have worked with a new disk.
I am a bit confused as I thought the resilvering would take care of boot partitions too, the new disks had not been formatted or partitioned before.
I'd also like to understand if there is a difference between using only grub or the proxmox-boot-tool by itself. I'll explore more in the host boooloader docs.
Last edited: