Make a Proxmox Cluster and Delete a Proxmox Cluster

We make a proxmox cluster, transfer all the VM's and containers, and finally delete that cluster.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Remove Node from cluster
turn off node you want to remove
pvecm nodes
pvecm delnode (name of the node)
reboot
DELETE CLUSTER
pvecm expected 1
Stop the running cluster
systemctl stop pve-cluster
Force the nodes to run in local mode
pmxcfs -l
Delete all the cluster configuration files
rm -f /etc/pve/cluster.conf /etc/pve/corosync.conf
rm -f /etc/cluster/cluster.conf /etc/corosync/corosync.conf
rm /var/lib/pve-cluster/.pmxcfs.lockfile
Stop cluster from running
systemctl stop pve-cluster
Reboot and cluster should be gone
rm /etc/corosync/authkey
pve.proxmox.com/wiki/Cluster_...
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#LoResDIY
My Channel: / loresdiy

[@santaclaus1291]

No backups? No mercy! Thanks for your efforts. Appreciate it!

[@paulbunion584]

This saved my bacon. Had to do this at work and wasn't completely sure how to proceed w/o blowing something up. Worked like a charm first go-around. Cheers to you sir!

[@LoResDIY]

Cheers

[@original2k]

this was a great/concise rundown on how to delete a node but more specifically delete a cluster. i had tried it out for giggles and it turned out to be more of a hassle than anything due to my lack of good hardware, the HA stuff is nice but not necessary until i get some more hardware stood up, so this worked out perfectly for me. thank you sir!

[@LebedotNaPero]

Thank you LoRes, your guide helped me to bring back cluusted that wont turn on VMs.

[@higorribeiro5368]

Thank You, I did the step by step and it worked in my environment.

[@yasaralzakout7691]

Great work, thank you a lot; I was struggling after changing my node IP address while it was connected to a cluster; it took a lot of research without getting a solution; your video is a complete solution for this issue. I appreciate your hard work.

[@LoResDIY]

Glad it helped!

[@noisynerdman]

Thanks sooo much for this. I was compiling knowledge from various blogs after i learned I didn’t need a cluster. Good way to fix without having to redo a hypervisor from scratch. :) yes agreed always back up and probably have a backup of that backup

[@LoResDIY]

Great to hear!

[@sparkie5571]

perfect walkthrough! Thank you for the expected 1 command too! Working on migration from an old node to a new one with more storage etc., has been very helpful thus far.

[@sparkie5571]

Biggest takeaway was that the backup method was a far easier and faster approach. As always though, I learned that AFTER I had made a cluster!

[@LoResDIY]

Glad it helped!

[@marifhossain]

Thanks! Saved days of work my friend. That's how we are evolving exponentially with the help of other's experience.

[@LoResDIY]

Great to hear!

[@user-wv7mx2ni2e]

thank you so much. His method was surgical. excellent

[@amammes]

Thanks so much, those steps worked fine for me on 7.4-3.

[@LoResDIY]

Good to hear

[@Skippygamer]

I disliked the video at first but you pulled me and helped at the end!!!

[@LoResDIY]

Glad you stuck around for it. Cheers

[@danieldewindt3919]

Grate, finally a way to safely add and delete. I'm going to reinstall proxmox i'm stuck with some pev servers in my datacenter tree All vpe servers are deleted from the node still all systems work fine

[@ushanslife4580]

Thank you !

[@LoResDIY]

You're welcome!

[@ierosgr]

An interesting follow up to this video would be.....transfering VMs and Containers to another node without the two nodes being members of the cluster. Maybe backup the VMs to an attached external drive or an remote share ( that would be an extra video of toy do it in parts) and then attach that hdd or remote share to the other node and restore (no you can t restore since this node is with no VMs) so scp that VM file to the appropriate storage and path and start it up. Probably it will work

[@pealock]

Look into Proxmox Backup Server (assuming thats the platform you are using). It is really easy to set up and allows VMs and CTs to be easily passed back and forth between non-clustered PVEs

[@philippef.4590]

Thank you

[@LoResDIY]

You're welcome

[@fortedexe8273]

Is it possible to join those 2 clusters again after delete it ? I want to change the host name, so I need to separate the cluster first, change hostname, then join it again. I dont know if it possible with your delete method ?

[@MarkConstable]

So how did you enable that lovely dark theme for Proxmox?

[@LoResDIY]

Here you go github.com/Weilbyte/PVEDiscordDark follow the instructions it's super easy.

[@fio_mak]

just wondering what happened to Lo-Res-Temp node? does it needs re installation or it just works as an independent node on its own?

[@LoResDIY]

It works on it's own or you can repurpose the server for another task.

[@Tr1pke]

Looks like I’m back in school. Blindly copy past and read what you copy paste 😂😂

[@jeffherdzina6716]

I must try your way of waiting. Does Import or domestic work better? Great Video.

[@LoResDIY]

Sadly I can't get to many imports in the area I live; so the majority of my waiting time is spent with Goose Island IPA and Coors Light. Luckily though there are a handful of micro brews within an hour of my house that I can visit if I'm feeling adventurous. Cheers 🍻

[@brunocabrera765]

Hi there, I followed your video to destroy the cluster, because one of them isn't allways on. All it0s ok, but the cluster continues showing me the 2ond node. What I had to do to make it disapear is to delete folder on /etc/pve/nodes/ This was on a proxmox 7.3-3 version. Now all VM starts ok and the cluster has gone. Thanks!

[@vng]

After you remove the temp node from the cluster, will you still be able to use that node as a standalone? Or do we need to reinstall Proxmox on that temp node after it has been removed?

[@LoResDIY]

You should still be able to use it as a stand alone, but you won't be able to readd it to the cluster.

[@konney236]

I made the same procedure as in the video ...removing files... now the temp node is able to be added to a new cluster. Just hat to delete "rm -f /etc/corosync/authkey"

[@fmj_556]

Dude your awesome!

[@LoResDIY]

Right on! Thank you

[@snowballeffects]

The moment @2m25 where the page looks like it had hung - in fact what has happened is that the node has now joined the cluster and the page you are looking at is now expired as the session is effectively lost as the node is now part of the cluster and by refreshing you iinstigated a new login a with a new session. HTH

[@LoResDIY]

It does help. Thank you for sharing. Cheers

[@mariamhtsiou9175]

It seems pretty strange the fact that gui gives you the opportunity doing things but not undoing them ... so you can del/remove a cluster only from cli? Also, when you migrate VMs to another node, shouldn t be advised (even if you think its obvious) that user must keep CPU type the same between two machines, like KVM64->KVM64 or Qemu->Qemu ... but Host -> Host there problems may rise if both CPUs arent of the same CPU family. Imagine moving a licensed SQL Server 2019 to another Host and when you start it, the no license has been acquired message to be displayed. PS A nice video would be a comparison test between CPU selection (upon VM creation) and the impact it has inside the VM because it has and an important one. Spoiler kvm64

[@LoResDIY]

Agree and thanks for the info

[@brentmobley6968]

@@LoResDIY I would assume removing nodes from a cluster in the gui is to prevent folks from accidently doing it. Kind of like going to the more button to remove a vm instead of being able to right click on the VM to remove.

[@ierosgr]

3:04 you choose target node ok. ... shouldn t there be an option to select in which storage of the target server the VM is going to placed on? At lest for VMs it gives you the option of target storage as well. Some questions though rise up -in my environment each VM has for nic a virtual bridge which is based upon 2-4 ports with LACP mode. When it will be migrated it will have as a nic the targets server's nic which will be different in many aspects of the source one. Imaging now running your Active Directory Server there and upon restarting it it will see different nic (it's mac address at least) ..... I think for this migration to be achieved successfully both servers need to be exactly identical in storage , network configuration. Not to mention if you have attached to the VM in the source server an extra storage which is being kept in the local server. I wouldnt want to know wha twill happen after migration. Then again I need to because its kinda like what my current configuration is.,... Sorry for the long post ..... but none comes up close of explaining or showing these things even in forums.

[@LoResDIY]

You are correct if the nic set-up is different on the two nodes the VM or container will not be accessible after migration. I have had this happen and was able to resolve it by changing the nic settings to match those of the machine I migrated too. Same with the storage. Hopefully that made since.

[@thegreyfuzz]

PM 7.0-2 - Joining cluster hangs in the GUI, but the action succeeds... did it to me add each of the 4 nodes. weird bug I guess.

[@LoResDIY]

10-4 thanks for the feed back

[@sullenfps]

life saver

[@LoResDIY]

Cheers

[@Snp88tag]

HI in my case i also at the end need to remove NODE folder from /etc/pve/nodes before that deleted node still show in GUI

[@LoResDIY]

Thanks for sharing

[@mihalcimihalci4768]

Hello, try do it on diferent vlans on diferent network this its comlicateed

[@zubrizavr]

1. backup all vm\ct ...

[@zacp6339]

You may also have to delete the node here: /etc/pve/nodes/ or it might still show in the GUI. Great video though!

[@LoResDIY]

Thank you for sharing the info. Much appreciated!

[@merlingt1]

Thank you!

[@LoResDIY]

You're welcome!