Video summary

One Server vs Cluster: What Your Homelab Actually Needs

Main summary

Key takeaways

Technology

Overview

The video compares running a homelab on one powerful server with building a cluster of smaller machines. The discussion draws on Christian Lempa’s experience with both a large enterprise server and a three-node Proxmox/Ceph cluster. It is an analysis rather than a step-by-step tutorial.

One Capable Server

Advantages

  • Often the simplest way to expand: add memory, storage, a faster CPU, or expansion cards—or replace the server when it reaches its limits. This is vertical scaling.
  • Enterprise platforms can offer room to grow through additional drive bays, memory slots, and PCIe lanes, as well as features such as ECC memory and redundant power supplies.
  • A single host can run many learning projects using virtual machines and containers, including Docker, Kubernetes, VLANs, virtual networks, and virtualized firewalls.
  • Workloads that need substantial CPU and memory in one place may be a better fit for a single server. Cluster resources are not automatically pooled for an individual VM.
  • A simpler setup can mean less infrastructure maintenance and more time for the projects running on it.

Disadvantages

  • The server is a single point of failure: a hardware fault, host update, or required reboot can affect all workloads.
  • Redundancy within a server can help, but its workloads still share core components such as the motherboard, storage hardware, and hypervisor.
  • Expansion is limited by the platform’s physical capacity.

Cluster of Smaller Machines

Advantages

  • Adding machines is horizontal scaling, allowing the homelab to grow incrementally without replacing one large server.
  • Separate nodes create independent failure domains. With shared storage and a functioning quorum, Proxmox can restart important VMs on another node after a node failure.
  • Live migration can move a running VM between nodes, supporting maintenance without taking the whole platform offline.
  • Building and operating a cluster provides hands-on experience with technologies such as quorum, shared storage, Ceph, migration, monitoring, and failure recovery.

Disadvantages

  • Multiple nodes require additional hardware, networking, power, storage, and ongoing maintenance. Costs can rise even when the individual machines are small and efficient.
  • A cluster does not automatically provide high availability; that requires suitable configuration and supporting infrastructure.
  • In the described three-node Ceph setup, each NVMe drive is an OSD, and Ceph keeps three replicas of data across the nodes. This provides resilience but leaves roughly one-third of the raw storage capacity usable for VM storage.
  • Ceph replication is not a substitute for backups. The setup still needs an independent backup system.
  • The example cluster uses 10 GbE networking, and a separate fast Ceph backend network would be preferable. Ceph and Proxmox also add services and states to monitor and troubleshoot.
  • Cluster maintenance can become a burden after the learning goals have been met.

Recommendation

A single server is a good fit when simplicity, consolidated CPU and memory, and time for application-level projects matter most—and occasional downtime is acceptable. A cluster is more appropriate when incremental expansion, independent failure domains, live migration, and cluster-specific learning are priorities, and the extra cost and operational complexity are worthwhile.

The speaker’s broader advice is to start with real use cases and learning goals rather than building a setup just because it looks impressive. A capable single server may be enough; a cluster is valuable when its capabilities are needed or when operating one is itself a goal.

Guides and Related Projects

The speaker refers viewers to existing videos about Proxmox clusters and enterprise servers, including a FreeNAS/Proxmox cluster project. The video itself does not provide setup instructions.

Main Speaker and Sources

  • Christian Lempa — host and source of the homelab comparisons and firsthand project experience.
  • AMD — acknowledged as the video sponsor; the speaker discusses AMD EPYC server processors and their enterprise features.

Rate this summary

Your feedback will help improve summaries.

Improve this summary

Reprocess with a stronger model when the summary feels incomplete or inaccurate.

Pro

Translate summary in another language

Pro

Ask questions to this video

Chat for follow-up questions, clarifications, and source-backed answers.

Coming soon

Share this summary

Original video