Upgrade 7.4 to 8, not reacheable after reboot

Hello everybody! I have the same problem. But I have KVM Session on my OVH Server. I open SSH via Firewall and now I have console. The Problem is /etc/pve is empty

I have checked the syslog logs and found this:

Code:
/etc/pve/local/pve-ssl.key: failed to load local private key (key_file or key) at /usr/share/perl5/PVE/APIServer/AnyEvent.pm line 2009.

Any ideas? :-( Hole Proxmox is offline !
 
Hello everybody! I have the same problem. But I have KVM Session on my OVH Server. I open SSH via Firewall and now I have console. The Problem is /etc/pve is empty

I have checked the syslog logs and found this:

Code:
/etc/pve/local/pve-ssl.key: failed to load local private key (key_file or key) at /usr/share/perl5/PVE/APIServer/AnyEvent.pm line 2009.

Any ideas? :-( Hole Proxmox is offline !
Ok i confirm that i have the same problem, pve is empty :/
 
Last edited:
This is because pve-cluster service is not starting.
Seem that ovh use cloudnit for provisiong, and it's adding "127.0.1.1 <hostname" in /etc/hosts , at each boot.

and pve-cluster need to resolve the real ip.

can you check your /etc/hosts ?


similar reported bug:
https://forum.proxmox.com/threads/strange-proxmox-errors.129419/#post-568304
Yes, I checked yesterday and the file is the same as version 7.4 of another server I have.

So I can't find the changed file. I also added touch /etc/cloud/cloud-init.disabled to be safe

finally, before restarting post upgrade to version 8, I saved the pve folder and replaced it in rescue mode.

the only thing i see via kvm is

Booting 'Proxmox Ve GNU/Linux'

Load Linux 6.2.16-3-pve...

Loading initial ramdisk...


and nothing else..
 
finally, before restarting post upgrade to version 8, I saved the pve folder and replaced it in rescue mode.
you have replaced it ??? /etc/pve is not a real directory, it's mounted with pve-cluster service. (the real datas in /var/lib/pve-cluster/config.db , and pve-cluster expose it like a directory in /etc/pve).

also, pve-cluster will not start, if a real /etc/pve/ with files exist. (so double check that you don't have restore content in the real filesytem, /etc/pve should be empty until pve-cluster is started.)

Finally, check pve-cluster service status.

# systemctl status pve-cluster
# journalctl -u pve-cluster -f



I don't known about your kvm hanging, it's another bug, not related to pve-cluster.
 
you have replaced it ??? /etc/pve is not a real directory, it's mounted with pve-cluster service. (the real datas in /var/lib/pve-cluster/config.db , and pve-cluster expose it like a directory in /etc/pve).

also, pve-cluster will not start, if a real /etc/pve/ with files exist. (so double check that you don't have restore content in the real filesytem, /etc/pve should be empty until pve-cluster is started.)

Finally, check pve-cluster service status.

# systemctl status pve-cluster
# journalctl -u pve-cluster -f



I don't known about your kvm hanging, it's another bug, not related to pve-cluster.

ok, it's not a problem i redo the upgrade from 7.4 to 8 and then i try to understand this thing through my kvm, thanks.
 
ok, it's not a problem i redo the upgrade from 7.4 to 8 and then i try to understand this thing through my kvm, thanks.
Since you tried man things already. Please check if ovh installs NTP/ntpsec and replace those by chrony. There is a very annoying bug with ntpsec when network bridges are involved. I am still puzzled that this bug made it into the super stable Debian. :-/
 
Since you tried man things already. Please check if ovh installs NTP/ntpsec and replace those by chrony. There is a very annoying bug with ntpsec when network bridges are involved. I am still puzzled that this bug made it into the super stable Debian. :-/

Bingo! thanks the problem was ntp! :)

My mistake I installed in previous tests ntp and ntpdate.

This time I didn't install ntp and the server restarted perfectly.

At this point I wonder if I have to repeat the same upgrade on a production machine do I remove ntp first?

Thank you
 
Bingo! thanks the problem was ntp! :)

My mistake I installed in previous tests ntp and ntpdate.

This time I didn't install ntp and the server restarted perfectly.

At this point I wonder if I have to repeat the same upgrade on a production machine do I remove ntp first?

Thank you
Yes, purge ntp and ntpsec as well as systemd-timesyncd from all systems and deploy e.g. chrony.
After that you can proceed with the upgrade.

I was debugging this issue for about 6h. But it's also mentioned in the upgrade notes since I reported it.
https://pve.proxmox.com/wiki/Upgrade_from_7_to_8#Network_Fails_on_Boot_Due_to_NTPsec_Hook
 
  • Like
Reactions: milk
Good morning everyone.

Over the weekend I tried to upgrade several times from 7.4 to 8 on an empty ovh test machine with this intel(r) xeon(r) cpu e3-1270 v6 @ 3.80ghz and NVME

the upgrade works without errors until the machine is rebooted.

When restarted, the machine is unreachable.

I checked through ovh rescue mode several times, also tried to downgrade the kernel with proxmox-boot-tool kernel, but without success.

I don't find particular errors in the logs through the rescue mode, but after trying and trying again I went back to 7.4

Do you think it's a kernel problem? or to the network configuration? or is the machine too old to host version 8?

Do you have any ideas?

Thank you all
Hi @milk do you install proxmox 7 from the ovh template ?
Now all work fine after upgrading?
Do you think it’s a good idea to upgrade a product environment or reinstall 8 version ?
Thanks
 

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!