New Proxmox Cluster with AMD?
- Thread starter AlfredP
- Start date
What's your goal here, replace 1 or all servers? I'd try to 'better' the processors. You mention you have a 'dual-core' Xeon 12/24 so potentially 24 core/48 threads and are asking about a AMD (single) with 16/32. Not really the same type of 'hardware'. I don't pve will care which processor you want to use, unless you have some special - or specified, the cpu type as long as you have the QEMU agent installed and 'ticked' in the config, they should work if you have to migrate/move them from 1 platform type to another.
Thank you for your reply
The warranty of all servers has expired and therefore the entire system will be replaced. The current system would be used for other purposes that are not critical.
According to CPU Benckmark, the AMD Epyc 9174F CPU would already provide a lot of performance. In order to have resources for the future, the decision could fall on the CPU AMD Epyc 9274F (24/32) with several cores (single CPU motherboard - Supermicro H13SSL-N).
The warranty of all servers has expired and therefore the entire system will be replaced. The current system would be used for other purposes that are not critical.
According to CPU Benckmark, the AMD Epyc 9174F CPU would already provide a lot of performance. In order to have resources for the future, the decision could fall on the CPU AMD Epyc 9274F (24/32) with several cores (single CPU motherboard - Supermicro H13SSL-N).
Attachments
One other thing you should probably consider… if you intend on running the AMD machines while you still have the Intel-based machine operating, live migration may not work, because of the different processor architecture. (Assuming you're running Proxmox on all of them.)
We have already thought about that. If it is enough to stop the VMS, move it to the new system and then restart it, we would be satisfied.
However, if changes need to be made to the configuration of the individual VMs, then this will be quite time-consuming.
Thanks for the information
However, if changes need to be made to the configuration of the individual VMs, then this will be quite time-consuming.
Thanks for the information
Get the new servers and add them to the cluster. Do not reuse names or IPs to avoid some ssh pitfalls. To migrate the VMs: stop them, move to the new AMD node and start. If changes are needed depend on how the processors are configured in the VMs. If all hosts have the same CPU and microcode revision, you can use "host" to get the best performance and live migration.
