Removal of Drive, PVE stuck on boot

[kingosoul]

Hi everyone,

We are having a wierd issue, we have proxmox set up on a seperate ssd which works fine, another drive (m2) that contained a VM, has run out of buffer threshold and has gone to read only. That is all fine, we have recoved the data, and have the VM's running again from a new ssd.

However if I physically remove the readonly m2 drive that is now unmounted from the GUI and config files in the shell.

The device boot gets stuck at

Found volume group "pve" using metadata type lvm2
3 logical volume group "pve" now active
/dev/mapper/pve-root: clean 1014506291456 files, 7511043/25165824 blocks


With the m2 still in, this message is slightly different

Found volume group "pve" using metadata type lvm2
3 logical volume group "pve" now active
/dev/mapper/pve-root: recovering journal
/dev/mapper/pve-root: clean 1014506291456 files, 7511043/25165824 blocks


The only difference between the boots is the recovering journal line. Even after waiting 30 minutes with m2 removed, we are stuck at this message.

Any help or insight would be kindly appreciated.

Regards,

James

 

[leesteken]

Removing (or adding) a PCI(e) device can cause other devices to change their PCI ID (as they are numbered sequentially by the motherboard BIOS). This can cause the network device(s) to change their names and then your network configuration won't be correct anymore.
Can you access the Proxmox host console and check the output of ip a and edit the network configuration with nano /etc/network/intwerfaces accordingly?
Do you use VMs with PCI(e) passthrough that automatically start? That can cause all kind of trouble when the PCI IDs don't match anymore.

 

[kingosoul]

Removing (or adding) a PCI(e) device can cause other devices to change their PCI ID (as they are numbered sequentially by the motherboard BIOS). This can cause the network device(s) to change their names and then your network configuration won't be correct anymore.
Can you access the Proxmox host console and check the output of ip a and edit the network configuration with nano /etc/network/intwerfaces accordingly?
Do you use VMs with PCI(e) passthrough that automatically start? That can cause all kind of trouble when the PCI IDs don't match anymore.

Awesome thank you leesteken for your prompt help.

I had a suspision it might be related to that. I currently have two graphics cards and one HDD passed directly through to two seperate windows vm's. Sounds like I will need to:

1. shutdown vm's
2. unmount all of those PCI devices
3. remove auto boot for vms
4. shutdown proxmox machine
5. remove the dead m2 drive
6. Due to unmount the gfx card should now show the shell again, use this to access shell and update the interfaces file so I can regain access on my network.

After which I should be able to use the web gui and remount my devices.

Please correct me if I am wrong