Install Proxmox 7 on OVH server from ISO

[cobuy]

Hello,

I have succesfully intalled Proxmox 6 from the OVH installation template but the upgrade to version 7 breaks my connection to the server after having rebooted.

Now, I'm trying to install directly Proxmox 7 from the ISO using the "Bring Your Own Image (BYOI)" OVH setup:


First attempt

URL: https://www.proxmox.com/en/download...load&args[0]=3d9337b3e1d40ffcbfebd63fbc7db8ef

Format : raw
SHA256
ae38bcb5ecc9aa97f6b13b89689fc4e876f9535f738bc0be4ffa4924274f25d9

On step 5/7:
can't write raw on disk
Burning image on hard drives ( 5 / 7 )


Second attempt

URL: https://www.proxmox.com/en/download...load&args[0]=3d9337b3e1d40ffcbfebd63fbc7db8ef

Format : cqow2
SHA256
ae38bcb5ecc9aa97f6b13b89689fc4e876f9535f738bc0be4ffa4924274f25d9

On step 5/7:
Burning image on hard drives ( 5 / 7 )
can't write raw on disk



3rd attempt

I have dowloaded the ISO and put it on of my server via FTP

URL : https://myserver.net/proxmox-ve_7.0-1.iso
Format : raw
SHA256
ae38bcb5ecc9aa97f6b13b89689fc4e876f9535f738bc0be4ffa4924274f25d9

Install seems OK, no error messages.
Server rebooted
no response to ping


Attempt 4
1 - Debian Buster install from OVH template
2 - Proxmox 7 install : https://pve.proxmox.com/wiki/Install_Proxmox_VE_on_Debian_Buster
don't work


Attempt 5
Proxmox 6.4 installed via OVH template
update to 7 https://dannyda.com/2021/06/25/how-to-upgrade-from-proxmox-ve-pve-6-4-9-to-7-0-4-beta/
6.4 installation is ok
upgrade to 7 is ok
when I reload Proxmox I can see the version 7
But when I reboot the server, connection is lost, no response to ping

I guess the updgrade chas changed something in hosts or network interface


Atempt 5 bis
After upgrade :
systemctl enable networking; reboot

ping: ok
ssh: lost access
web interface: no

 

[lps90]

Same problem here!
Still no solution?

 

[Spektrum]

Just rented a brand new server, installed Proxmox 6.4-13 from their template. After the upgrade, network interfaces aren't reconnecting to the network. I'm using IPMI from OVH's web management interface to access the server right now. I'm no expert in this field, not at all. I'm currently trying to understand where are the "active" configuration files. No matter what I've tried until now, it seems like my config changes have no effect whatsoever.

What is sure at this point: before the upgrade, I had two network interfaces: enp5s0f0 and enp5s0f1. The first one had a public IP address and the second, a private one.

After the upgrade, both interfaces are gone. I now only have eth0, which has the same MAC address enp5s0f0 had before the upgrade. So if the MAC address matches, I don't understand why it's not given a public IP address by OVH's network infrastructure.

I'll keep you guys updated if/when I find a solution to this. I have the feeling that someone relatively competent with Debian networking would solve this in minutes. He he!

 

[Corwin]

I'm interested about this topic as I will do the same stuff in the coming days. As I did not fight yet, I can only *think*...
You should notice bridge configuration has changed, as stated in the documentation.

Even for a new PVE7 installation, before rebooting you need to edit /etc/network/interfaces and add the hwaddress setting with the MAC assigned by OVH, this a manual action I believe.
Check the nic predictable name, it should not be eth0 anymore except if you changed the kernel line in grub.
Have a look also in /etc/hosts for the public IP.

Hope it helps...

 

[Spektrum]

Got it! As of now, I only have the public IP / network adapter working, but I suppose it won't be too complicated to duplicate this config for the private network... So here's my solution

First, connect to your server through OVH's IPMI web interface.

Open /etc/network/interfaces, comment everything out, and add the following (make sure to replace: a.b.c.d with your server's public IP) :

auto lo
iface lo inet loopback

auto vmbr0
iface vmbr0 inet static
    address a.b.c.d/24
    gateway a.b.c.254
    bridge_ports eth0
    bridge_stp off
    bridge_fd 0

Restart networking: /etc/init.d/networking restart. I read everywhere that this is deprecated, but I suppose my current setup hasn't switched to ifupdown2, I'm not sure.

If I understand correctly, vmbr0 is assigned an IP by OVH's infra, and eth0 is bridged through it. Why is that? I really don't know. I adapted this configuration from and older server I have running at OVH.

For the second adapter (in my case at least), I suppose I'll make a similar config using vmbr1 and eth1.

I hope this will help solve your own problems!