[SOLVED] Proxmox not booting at all after disk replacement

[robert1]

Hello community,

I've got a mayor problem after replacing one disk and successfully resilvering the pool (incl. GRUB install, etc.). The system goes over the BIOS and then immediately stops with a blinking prompt (no Bootloader, no error message, nothing).

Some basic information: The installation is on a HP G-Series proliant, HD 4 slots:

• a) SSD 120 GB
• b) SSD 120 GB
• c) HD 5TB
• d) HD 5TB

a) and b) RAID 1 ZFS - "rpool"
c) and d) RAID 1 ZFS - "zpool"

Installed is/was Proxmox 6.1, there have been local and local-zfs on rpool and zdata and backup on the zpool. Within the local/local-zfs all 7 VE's lived.

What have I done:
I identified the degraded pool of the SSDs (a & b) and saw that b's [HD-ID]-part-3 was faulted. I replaced the disk by a new one (same size) and copied the partition table from the other one to it, randomized the IDs, resilvered it and copied the GRUB to the new one (as described here in some threads). Resilvering sucessfully finished and all things seemed ok (remark: I'd this several times on various installations, also on this one once a while ~ 7 months). Finally, after checking I rebooted the machine and ... nothing (it looks like, GRUB is gone and Rescue from the ISO also gives Error 101.

Main issue of this all is that I've got 4 slots for HD's (so I can't put some others inside). I also think that both pools are still ok (zfs and grub are two different things). Only GRUB (or whatever) is gone - but it also can't be installed/repaired (tried it with the ISO). What I won't try is to follow these procedures container-aus-zerstörtem-proxmox-wiederherstellen.61454/ because they don't look comfortable (because I don't know the structure and procedures what I must take and put somewhere) to me AND I don't wont to destroy the existing (and in my opinion still functional pools where everything is in).

Any help would be appreciated (some sort of howto). If you ask me about some more logs etc. - I'm not able to get these right now but I'll try to give you all needed answers if I can ;-)

(ps. Antworten gerne auch auf Deutsch)

 

[robert1]

UPDATE 1:
To figure out whether it's the bootloader or something else I plugged two brand-new SSDs into the machine (ps. it has no UEFI). I installed 6.3 from the USB stick and it finishes successfully. After reboot the same thing happens, blinking prompt. This happened, because I made a reinstall when both HD's (c and d) have been still attached (but left un-configured during installation - I wont destroy my backups ;-)). When I detached them, Proxmox (new installation) came up. It Seems that changing HDs physically on the server it changes boot priority as well so the sda and sdb wont be found AND also not reattached) but the BIOS didn't tell anything (strange).

Nevertheless, the 120 GB disks a/b don't come up again - so I'll have to take the backups from the c/d zdata pool. :-|

UPDATE 2:
Due to the fact that I didn't check everything (and forgott to check the f. boot order again), the boot order within the BIOS sets always and every time the HD's which are all the time attached to those where it should boot from (stupid stuff, but HP's Amibios works like this).

So lesson learned:

Every time you replace a drive physically, check within BIOS that boot order is set again to ... It's not the fault of GRUB or Proxmox (it's probably yours - in that case it was mine ;-) )


(I leave this thread for others to see, how you can run into something ...)