I am thinking about putting together this Proxmox box based on Christian Lempa’s “My Proxmox Server Build.” My use case is running Docker or Podman images and VM’s for network monitoring, home automation, and learning about operating systems.
I have to find a different ECC RAM because the one that Christian recommends has only one stick available on Amazon. I know that I have to be careful to pick ECC RAM that is compatible with my system. Does anyone have any recommendations on type and brand?
I also welcome any comments on the other components.
I always used Kingston memory for all my builds and never head any issues with it. They have a search tool on their site with which you you can list all the compatible memory modules for your specific mainboard… Kingston Memory Search - Kingston Technology
Kingston, Crucial, SK Hynix, whatever Supermicro sells for their servers.
I am questioning if your 6 cores will be enough. You can get used servers on ebay with lots more processor and ram for less money and maybe something to consider. Unless noise is going to be a problem, and I do understand that can be a huge factor in a home lab. My wife would kill me if I brought my servers home, luckily I have space at work and a VPN so I can get in and do stuff from home.
Ryzons are solid but yes you are limited as to what you can run simultaneously but then this is a home lab. As a first iteration I think he is in good shape.
All this talk about cores got me thinking if my build is strong enough. Here is my use case:
Docker with 20 containers running simultaneously (don’t know what I’m going to run)
4 virtual machines
Given this use case, do you think I should go with the original build or something like the 16 core Ryzen 9 3950X? If I go with the Ryzen 9 3950X, can anyone recommend a compatible micro ATX board that does not require a bios update?
Sorry, but the pure number of containers/VMs is useless to determine whether a certain hardware is powerful enough. You could run something lightweight like a Unifi controller, or a heavily used RDBMS server. The latter would require multiple times the capacity of the former.