Load balancing spreads requests across two or more servers so that no single machine is overloaded and one failure does not take the service down. It is one of the best uses for refurbished servers: instead of one expensive new server, you can run several capable refurbished ones behind a load balancer and gain both capacity and redundancy.
How load balancing works
Clients connect to a single address owned by the load balancer. The balancer forwards each request to one of the backend servers according to an algorithm, and constantly checks that each server is healthy. If one stops responding, traffic goes to the others.
| Algorithm | How it chooses a server | Best for |
|---|---|---|
| Round robin | Each server in turn | Identical servers and similar requests |
| Weighted round robin | In turn, but stronger servers get more | Mixed hardware, such as older and newer refurbished models |
| Least connections | The server with the fewest active connections | Long-lived connections and uneven request times |
| IP hash or sticky sessions | The same client goes to the same server | Applications that keep session state on the server |

Why refurbished servers suit load balancing
- Redundancy for less: two refurbished servers often cost less than one new server of similar power, and you remove a single point of failure.
- Scale out: add another node when demand grows. See our guide to server scalability.
- Mixed generations: weighted algorithms let you use older and newer hardware together.
Setting up load balancing
- Choose the balancer. Software options such as HAProxy and NGINX run on a small server or virtual machine. For critical services, run two balancers with a shared virtual IP (for example with keepalived) so the balancer is not a single point of failure.
- Prepare identical backends. Same OS, application version and configuration on every node.
- Move shared state out. Store sessions, uploads and data in a shared database, cache or storage that every node can reach.
- Configure health checks. Check an application URL, not just whether the server responds to ping.
- Test failure. Switch one node off during testing and confirm traffic moves to the others without errors.
- Monitor. Watch response times and node health. Our monitoring tools guide compares free options.
Common problems
| Problem | Likely cause | Fix |
|---|---|---|
| Users get logged out randomly | Sessions stored on individual servers | Use sticky sessions or move sessions to a shared store |
| One server overloaded | Wrong algorithm or weights | Switch to least connections or adjust weights |
| Failed server still receives traffic | Health check too basic | Check an application URL and lower the check interval |
| Different results on different servers | Configuration drift between nodes | Automate deployment so every node is identical |
Frequently asked questions
What is server load balancing?
It is distributing incoming requests across several servers so that no single server is overloaded and the service keeps running if one server fails.
How many servers do I need for load balancing?
At least two backend servers. For critical services, also run two load balancers so the balancer itself is not a single point of failure.
Can I mix old and new servers behind a load balancer?
Yes. Use weighted round robin or least connections so stronger servers take more of the load.
Is HAProxy free?
Yes. HAProxy is open source and widely used for software load balancing. NGINX also offers load balancing in its open-source version.
Building a load-balanced cluster? Configure matching refurbished servers or ask us for volume pricing.


