patching best practices

Binary Bandit

Well-Known Member
Dec 13, 2018
Hi All,

I'm looking for some input on best practices for updating / patching our three node cluster.

As a bit of background, we've been running 5.2 for nearly a year in a three node cluster. This was rock solid stable. I would update it once a month from the free repositories with no issues.

Fast forward to an update that installed version 5.3 a few weeks ago. The day after this we began experiencing random cluster node reboots. I've learned a great deal about logs and troubleshooting. Long story short, we're now on 5.4 and the cluster is stable again. We also bought support for all cluster nodes before upgrading to 5.4 so we're now using the enterprise repositories.

Thoughts? Suggestions? How would you recommend we do our patching, what's worked best for you and your organisation?

We were planning to add a forth node to the cluster purely so that CEPH would have an available OID to move data to should another node go down. Our thinking is now that this forth node should still be duplicate hardware but that it should be stand alone. Specifically the forth server run Proxmox outside of the cluster, will receive patches a week prior to the cluster nodes and will provide backup for cluster VMs. In the event of a cluster issue we'll have the option to spin up backed up VMs on it. Basically the forth node isn't a node, it's a server that's used for testing and disaster recovery.

thanks in advance,


Thoughts? Suggestions? How would you recommend we do our patching, what's worked best for you and your organisation?
this is only true for minor updates like 5.x > 5.y.
For major updates, there is an upgrade how to like this [1].

In a Cluster you should generall do an upgrade one by one node.
1.) shutdown/move all running VM from this node.
2.) Disable HA if it is active (just for safety)
3.) upgrade the node.
4.) If a new kernel version where installed reboot the node.
5.) do the same with the next node

Thanks for the response Wolfgang.

That first Wiki is interesting ... hope that we go after new hardware at that point. I'd much prefer to add the new node, role VMs to it and retire an old node.


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!