Is it worth using Proxmox in cluster mode, or am I overcomplicating things?

[jtecg]

Hello! Thanks to everyone who takes a moment to read and give their opinion.

Currently, at my workplace, we have a Proxmox virtualization environment with several VMs running as SIT (development) environments on about 5 hosts (Proxmox 8) . These environments are not CRITICAL, but the company's developers use them for their work. A few hours of downtime wouldn’t represent a huge loss, but we can't leave the system down for an entire day either—those of you in IT and development will understand. For these hosts, we are using hardware that’s not the latest, but it’s all enterprise-grade (real servers in a data center).

My question is: Is it really worth adding the complexity of creating a cluster with these 5 machines? Recently, we reinstalled all the systems but haven't set up the cluster yet because older Proxmox versions (6) gave us a lot of issues when adding and removing nodes due to hardware replacements.

For this type of environment, are we adding unnecessary complexity? Or, aside from the ability to move virtual machines live, are we missing out on any other benefits for this kind of setup?

Again, thanks for sharing your experiences.

 

[bbgeek17]

Hi @jtecg , welcome to the forum.

For this type of environment, are we adding unnecessary complexity?

Only you and your organization can put a price tag on your data availability. There are always nuances and "it depends" involved.
What does it mean for you "down for a few hours ok"? Are you going to recover from backup? Do a fresh install? Have a golden template easily deployable? Or, are you just hoping you can fix it and recover quickly?
There is not a big difference between "few hours" and "down for a day". Are you prepared to spend a weekend recovering when the proverbial "stuff" hits the fan on Friday night?

For this type of environment, are we adding unnecessary complexity?

Do you have redundant network? Power? Storage? If not, then adding a slice of redundancy to your Compute may not buy you much.

Or, aside from the ability to move virtual machines live

Just having a PVE cluster does not immediately give you live VM migration. You need appropriate network and storage in place.

are we missing out on any other benefits for this kind of setup?

There is no other major benefit from having a PVE cluster beyond VM HA. That's the primary purpose of PVE cluster.
You will get a side benefit of centralized node management.

Best of luck

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[jtecg]

First of all, thank you for the warm welcome!
Yes, indeed, there isn't much difference between a couple of hours and a day... it clearly shows that these are non-critical environments, and I realized I answered my own question when I reread that, haha.
But yes, I think the main point is where you mention high availability for Internet and storage. Currently, we have two redundant internet links, but no redundancy in storage or any load balancer that would allow us to distribute the load in case of a sudden equipment failure.
With the migration of VMs between servers, I think we have good network and storage availability to handle it, but again, a VM could be down for a few hours and the cost to the company would be $0.

Again, thank you very much for your time and response, it really helped me think about it. With this type of environment, I think I will stick with a stand-alone setup and write some scripts to quickly migrate the VMs if necessary (this doesn't happen very often).