My production cluster is made of five physical computers connected via a 10GbE (10 gigabit) copper connection using RJ45.
Compute:
Each of the three compute nodes has the following hardware:
- AMD Ryzen 9 5900X CPU (12 Cored / 24 Threads)
- 128GB DDR4 3200 RAM (Non-ECC)
- 512GB NVME SSD
- Generic GPU for text display.
- 10GbE Network Card
- 500W Power Supply.
- Proxmox VE 9.x Operating System
Storage:
Storage is provided by another computer and has the following hardware:
- AMD Ryzen 5 5600G CPU (6 Cores / 12 Threads )
- Integrated Graphics
- 64GB DDR4 3200 RAM (Non-ECC)
- 512 GB NVME SSD for Operating System
- 8 x 2TB SATA SSD with TLC Flash
- 10GbE Network Card
- 500W Power Supply
- TrueNAS-Scale Operating System
- 10.45TB Uncompressed Usable Storage
- Fault Tolerance : Raid-Z2. Can lose 2 disks and continue to run.
Backup:
Backup is provided by another computer with the following hardware:
- AMD Ryzen 7 5700G CPU (8 Cores / 16Threads)
- Integrated Graphics
- 64GB DDR4 3200 RAM (Non-ECC)
- 512GB NVME SSD for Operating System
- 4 x 4TB SATA SSD with QLC Flash
- 10GbE Network Card
- 500W Power Supply
- Proxmox Backup Server OS
- 11.76TiB Raw Usable Storage
- Fault Tolerance : Raid-Z1 Can lose 1 disk and continue to run.
Network:
My network setup is simple. I have a single 10GbE Un-managed switch that all 5 computers connect to. While this is not redundant, it is efficient for a home lab like mine.
Power:
All of the servers are connected to a UPS to allow everything to run during brown outs, power spikes, or up to 10 minutes in the event of a power failure. This is enough to give me time to shut everything down cleanly.
Overall Production Cluster Specifications:
Compute: 36 Zen 3 CPU Cores / 72 Threads
Memory: 384GB DDR4 3200 (Non-ECC)
Storage: 10.45TiB of SSD Storage
Backup Storage : 11.76TiB