How to Optimize the Performance of Your Refurbished Server

Server Performance

Before you spend money on a bigger server, make sure the one you have is set up to perform. Many refurbished servers run slower than they should because of a power-saving BIOS profile, a disabled controller cache, unbalanced memory or out-of-date firmware. These fixes cost nothing and often make a noticeable difference.

Quick answer: measure first to find the bottleneck (CPU, memory, disk or network), then set the BIOS power profile for performance, populate memory evenly across all channels, enable the RAID controller write cache when the cache battery is healthy, update firmware and drivers, and right-size virtual machines. Upgrade hardware only where the measurements point.

Find the bottleneck

MetricHealthy signWarning signUsual fix
CPU usageSpikes, but not stuck near 100%Constantly near 100%Spread workloads, faster or more cores
MemoryFree memory availableHeavy paging or swappingMore RAM, or fewer or smaller VMs
Disk latencyLow and steadyHigh latency during normal useSSDs, better RAID level, controller cache
NetworkBelow link capacityLink saturated or errorsFaster ports, teaming, fix cabling
TemperatureWithin normal rangeThrottling or high fan speedImprove cooling and airflow
server upgrades

Free performance fixes

  1. BIOS power profile: set the system profile to performance (Dell) or high performance (HPE), or OS-controlled if your hypervisor manages power. Power-saving profiles can slow responsive workloads.
  2. Balanced memory: populate the same amount of memory on every channel and every CPU. Unbalanced memory reduces bandwidth.
  3. RAID controller cache: enable write-back cache when the controller has a healthy battery or flash cache.
  4. Firmware and drivers: update BIOS, controller, NIC and drive firmware, and use vendor drivers.
  5. Right-size VMs: giving a virtual machine too many vCPUs can make it slower, not faster.
  6. Clean up: remove unused services, old snapshots and scheduled tasks that clash with working hours.
Watch out: enabling write-back cache without a working cache battery risks data loss if power fails mid-write. Check the battery or flash cache status in iDRAC or iLO first.
Pro tip: a performance problem that appears at the same time every day is usually a scheduled job, such as a backup, antivirus scan or database maintenance. Move it out of working hours before buying hardware.

When to upgrade instead

If measurements show the server is genuinely short of resources after tuning, targeted upgrades are the next step. More RAM and SSDs are usually the best value. Read upgrading a refurbished server for compatibility checks, and see our storage configuration guide for disk layouts.

Good to know: good monitoring makes performance tuning much easier, because you can see trends instead of guessing. Compare free options in our server monitoring tools guide.

Frequently asked questions

Why is my server slow?

The most common causes are slow or busy disks, too little memory, a power-saving BIOS profile, scheduled jobs running during the day, or a network bottleneck. Measure each before changing anything.

Does BIOS power profile affect server performance?

Yes. Power-saving profiles reduce CPU clock speeds and can add latency. Performance or OS-controlled profiles are usually better for busy servers.

Will adding RAM speed up my server?

If the server is short of memory and paging to disk, yes, often dramatically. If memory is not the bottleneck, extra RAM will not help much.

How do I check server performance?

Use the operating system’s performance tools, your hypervisor’s graphs, and iDRAC or iLO for hardware health, ideally with monitoring software that records trends.

Need an upgrade? Browse server memory and server SSDs, or ask our team to recommend one.

Server Site