Ceph Licensing and Cost Considerations for Large-Scale Proxmox Deployments

[Rajan_Kumar.Jaiswal@sita.]

Hello Proxmox Team,

We are currently evaluating Proxmox VE with Ceph as part of a large-scale virtualization assessment and would appreciate some clarification from the community regarding licensing and operational costs.

From my understanding:

• Ceph is open-source and does not require a separate software license.
• Proxmox VE subscriptions are licensed per physical CPU socket and provide access to the Enterprise repository and technical support.
• Ceph is tightly integrated with Proxmox VE and can be managed directly through the Proxmox interface.

I would like to confirm the following:

1. Is there any additional licensing cost associated with deploying Ceph in large Proxmox clusters, or is the only commercial cost the Proxmox subscription?
2. For larger production environments (e.g., approximately 100 TB of usable storage), are there any hidden operational, maintenance, or support costs that organizations should consider beyond the cost of hardware, networking infrastructure, and Proxmox subscriptions?
3. Are there any official best practices or sizing recommendations regarding the cost of operating large Ceph clusters (for example, networking requirements, NVMe/WAL/DB devices, monitoring, or maintenance overhead)?

If there are any official Proxmox or Ceph reference documents covering these topics, I would appreciate links to them.


Thank you in advance for your guidance.

 

[LnxBil]

You're totally right with your assessment about the PVE subscription cost per socket as the only cost you have. It's open source, so you could also run everything without any licensing/subscription cost and of course then without any support. I recommend however to work with a Proxmox partner to get everything right and have first hand experience and support, especially if you want 24/7 support.

larger production environments (e.g., approximately 100 TB of usable storage)

IMHO, that's not that large. How many nodes are you talking about?

Are there any official best practices or sizing recommendations regarding the cost of operating large Ceph clusters (for example, networking requirements, NVMe/WAL/DB devices, monitoring, or maintenance overhead)

Depends on your performance requirements, size is not a performance requirement. You can go with 25 GBE, SAS disks with NVMe WAL of with multiple 400 GBE links and only 24+ local PCIe5 NVMe. What is your IOPS requirement?