Cluster reliability when removing nodes

[Pablo Alcaraz]

Reading the cluster chapter of the manual, there is this paragraph:

"At this point you must power off hp4 and make sure that it will not power on again (in the network) as it is.

As said above, it is critical to power off the node before removal, and make sure that it will never power on again (in the existing cluster network) as it is. If you power on the node as it is, your cluster will be screwed up and it could be difficult to restore a clean cluster state. ​

After powering off the node hp4, we can safely remove it from the cluster."​

Is it planned some kind of protection in case of a separated node is rebooted again? It could be a problem on organizations sharing machines between dev ops guys...

 

[Magneto]

The way I understand it, is that only applies when you want to permanently remove a host node from the cluster.
What is your usage, that it might be a problem for you?

 

[M-SK]

It would be good to have some info on impact of using the same hostname in the same subnet for something else (we, for instance, recycle the names and name hosts sequentially). Also, what if I want to reinstall Proxmox host for whatever reason and rejoin under same name?

 

[tschanness]

That won't be a problem, you may only have to remove the old SSH keys for the removed servers.

Regards,
Jonas

 

[M-SK]

That won't be a problem, you may only have to remove the old SSH keys for the removed servers.

Regards,
Jonas

That's what I reckoned as well, but why warnings of apocalypse from the documentation?
So, if I want to reinstall or add a new hardware node with previously "blacklisted" hostname, all I have to do is remove keys from /root/.ssh?

 

[tschanness]

Hi,

That's what I reckoned as well, but why warnings of apocalypse from the documentation?

The documentation is about the _same_ server (with the same PVE install). This does not apply when installing PVE again.

So, if I want to reinstall or add a new hardware node with previously "blacklisted" hostname, all I have to do is remove keys from /root/.ssh?

It's in /etc/pve or something like that. The cluster will complain if the key is already there...

Regards,
Jonas

 

[M-SK]

Ok, thanks for clarification.