[SOLVED] Update 7 to 8 issue with cloud-init

[Batto]

Hi,

I updated a server from v7 to v8 (for now it's only for tests purpose ).
Everything seems to goes OK but when I restart the server, no web GUI
SSH is ok so I dig and it's nearly the same problem as here : https://forum.proxmox.com/threads/127-0-1-1-problem-start-pve-cluster.100642/

My /etc/hosts file is automatically populated with a 127.0.1.1 entry which cause a lot of trouble.
Removing the 127.0.1.1 line allows to succesfully start pve processes (like pvestatd and pveproxy) but /etc/pve folder is still empty. And when I reboot, this annoying 127.0.1.1 line comes back.
The cure is to remove cloud-init package, fix /etc/hosts and reboot.

Some precisions :

• v7 is completely fresh/new
• Hosting is OVHCloud
• If you need more investigations, this is a test server so I can do anything for now on it (reinstall, upgrade etc ...).

I don't think it's a normal behaviour and as I have some old Proxmox instance to upgrade and new to create, I'm not veery confident ^^'
So if someone has a "real" fix or a logical explanation, it would be great

Regards

 

[bbgeek17]

So if someone has a "real" fix or a logical explanation, it would be great

the fix was mentioned in the thread you referenced:

yes your host shouldn't be managed by cloud-init.

Its possible in theory to use it for some customizations, but you really need to understand why its there and how every component of the system can interact with it. Proxmox is not tested or qualified by developers to run with cloud-init, you are on your own.

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[Batto]

the fix was mentioned in the thread you referenced:

Yes I didn't said the contrary xD

Its possible in theory to use it for some customizations, but you really need to understand why its there and how every component of the system can interact with it. Proxmox is not tested or qualified by developers to run with cloud-init, you are on your own.

But I don't use cloud-init, I discovered it with this issue xD
It's a dedicated server

 

[Batto]

Ok I succesfully upgraded to v8, I found the point which changes things : during apt dist-upgrade, there is some classic questions about keeping or dropping cfg files. At a moment it ask for /etc/cloud/cloud.cfg if I want to use the maintainer version : if I don't, upgrade goes well, if I do, cloud-init is messing things.
So I think documentation might mention that point (if this behaviour is normal) and I also noticed that openssh dropped support for ssh-rsa so you have to modify your sshd_config to force the support or regenerate a safer key on a supported protocol.

 

[spirit]

I don't think it's a normal behaviour and as I have some old Proxmox instance to upgrade and new to create, I'm not veery confident ^^'
So if someone has a "real" fix or a logical explanation, it would be great

Regards

The reason is ovh ^_^ It seem that they use cloudinit to provision their physical server.

 

[Neobin]

See also:
https://forum.proxmox.com/threads/strange-proxmox-errors.129419/post-567511

 

[Batto]

Thanks guy it's clear now