Cluster removal

[sysadminarchive]

Hello

I recently was looking into creating a cluster and adding a few servers into it. Just wanted to see how the cluster is made so I made one on a production server. Which was a mistake.

I then realised that the server should join a cluster, not create one and once you create a cluster, you can't join another one. I have multiple VMs running in my Proxmox server.

How would one go about removing the cluster from that server. I didn't add any other servers to the cluster.
I looked up some guides, but those mention a loss of VMs which is very bothersome.

Is there a way to remove a cluster that does not have any other servers added to it?

 

[bit2gb]

Hi
I have same problem in the past but i can't find a solution for this, to remove cluster safe with no error in the future.
What i suggest for you:
1.If you have different storage (for ex you have secondary 4 6 hdd in raid mounted into a folder like home2 where you have all virtual disk for vms into images folder) for os and for vm's virtual disk steps are:
-copy content of the /etc/pve/nodes/pve/qemu-server (conf file for all vm's) into your pc
-reinstall proxmox into same storage (be carefful because you don't want to form your hdd where you have virtual disk's for vm)
- mount again that second hdd where you have images for vm's into same location like in the present
-copy from your pc into /etc/pve/nodes/pve/qemu-server all conf file
-95% all vms work fine
2. if you have 1 storage for all (os and vm images) do same steps like for the first situation but firstly go into /var/lib/vz/images (if you use diractory with ext4 for filesystem) or where you have virtual disk copy all into other server/comp and when you finish all steps from 1 situation copy again into same location all virtual images

Have a nice day,
Alin

 

[sysadminarchive]

Hi
I have same problem in the past but i can't find a solution for this, to remove cluster safe with no error in the future.
What i suggest for you:
1.If you have different storage (for ex you have secondary 4 6 hdd in raid mounted into a folder like home2 where you have all virtual disk for vms into images folder) for os and for vm's virtual disk steps are:
-copy content of the /etc/pve/nodes/pve/qemu-server (conf file for all vm's) into your pc
-reinstall proxmox into same storage (be carefful because you don't want to form your hdd where you have virtual disk's for vm)
- mount again that second hdd where you have images for vm's into same location like in the present
-copy from your pc into /etc/pve/nodes/pve/qemu-server all conf file
-95% all vms work fine
2. if you have 1 storage for all (os and vm images) do same steps like for the first situation but firstly go into /var/lib/vz/images (if you use diractory with ext4 for filesystem) or where you have virtual disk copy all into other server/comp and when you finish all steps from 1 situation copy again into same location all virtual images

Have a nice day,
Alin

Thank you Alin. It's weird how it's so simple to create the cluster, but so hard to remove. I think this is really a fault in Proxmox.
Copying hundreds of gb of data over and reinstalling Proxmox just because you accidentally created a cluster seems like something that shouldn't be possible.

I'll keep the thread open a little bit to see if there are any other options. A staff member reply would be great, but we'll see.

 

[mira]

Just follow our documentation https://pve.proxmox.com/pve-docs/chapter-pvecm.html#_remove_a_cluster_node
This should be enough to remove the cluster.

 

[sysadminarchive]

Just follow our documentation https://pve.proxmox.com/pve-docs/chapter-pvecm.html#_remove_a_cluster_node
This should be enough to remove the cluster.

Can't really use that guide since I have no other machines in the node? I took a look at it before.

 

[mira]

First stop the corosync and the pve-cluster services on the node:

systemctl stop pve-cluster
systemctl stop corosync
Start the cluster filesystem again in local mode:

pmxcfs -l
Delete the corosync configuration files:

rm /etc/pve/corosync.conf
rm /etc/corosync/*
You can now start the filesystem again as normal service:

killall pmxcfs
systemctl start pve-cluster

 

[sysadminarchive]

and this wouldn't in any way affect the VMs/storages right?

 

[bit2gb]

nope because you have conf file for vms into /etc/pve/qemu-servers and virtual disk are located into other location.
For example for 1 directory as storage used for os and for vms location is /var/lib/vz/images