Complex recover of Proxmox

netvolt

New Member
Jun 30, 2022
5
1
3
Hello all, we´ve got in a hard situation where we need the help of this community to solve a hudge problem, or guidance at least. Our knowledge of Proxmox is close to newbie! This is the situation:

- Original SSD with root/installation of Proxmox installation has failed large blocks. The old "IT guys" installed it at a home SSD inside a server... /dev/root partition of this SSD is readable, /dev/data partiton is not readable. The system does not boot as it suppose to do anymore (a bunch of I/O errors) and no HTTPS available as well. But we can SSH it.

- The server have a second HDD (2TB) split into 1.8TB storage of some thin LVMs, where is the most part of data (but we cant access them directly).

We need your guidance and help to find ways to:
- Recover the meta data of broken SSD: VM´s, storage pools, and any other important meta data or config, to a new fresh Proxmox installation. Where to look, what need to copy so the new installation can find and start up the 2VMs inside second HDD.

Is there any light you can give us in this issue?

Below are some detailed information of our discs setup and vms.

Thanks in advance

1.jpg

2.jpg
3.jpg
4_1.jpg
4_2.jpg
4_3.jpg
4_4.jpg
 
PVE stores most of its configs in "/etc/pve" (and some few files outside of that like "/etc/network/interfaces" or "/etc/vzdump.conf") but "/etc/pve" will be empty if the pve-cluster service isn't running as this service will mount the contents of a SQLite DB as files in "/etc/pve". See here for more infos about the pmxcfs filesystem: https://pve.proxmox.com/wiki/Proxmox_Cluster_File_System_(pmxcfs)
 
  • Like
Reactions: netvolt
Thank you very much for your reply @Dunuin! Appreciate that!

We imported the old database and restarted the service that loaded the /etc/pve.
It shown the old "hyper" but we cant select it due certificate error. Is there more steps to fix it, in order proxmox to read Vms and discs properly?

Below is the image of what happened. Is there any recommendation to fix it into just one node (hyper)?
Already updated /etc/hosts, but it didnt change anything.

Thanks

new_error.jpg