Which Server Components Are Most Likely To Fail?

2 Men trying to repair a server while standing on a question mark

Posted on Wednesday, April 3, 2019

Despite overall quality and resilience of server hardware from all the major vendors improving, components can, will, and do fail eventually. In order to maintain system availability, you need to know which components are likely to fail so that you can schedule replacements in advance.

Our OEM-trained engineers closely monitor the lifespan of every component so we can accurately predict which will need replacing first. These are the top three you need to be aware of:

Disk drives

The parts most likely to fail in any server or storage array are the disk drives themselves. Each platter spins hundreds of thousands of times every day, and the bearings will only tolerate so much load. Similarly, the mechanisms controlling disk heads will move millions of times over years of service, creating another potential point of failure.

SSDs have a distinct advantage over magnetic hard disk drives, but they are not immune to failure, either. Each memory cell can sustain a finite number of write operations, after which it will be marked as unusable. When a sufficient number of cells have reached that point, the SSD needs to be replaced.

Power supplies

There’s a reason most servers come with a redundant power supply option: they all need replacing eventually. In many cases, power supply units (PSU) break down because the bearings in the integrated cooling fan fail. Sometimes the motor that powers the fan burns out, but the result is always the same: the PSU overheats and malfunctions. 

Fans

Heat is actually a major cause of component failure, particularly in dense racks in the data center. Any cooling fan failure will impede air flow through the server or array chassis, greatly increasing the likelihood of CPU or RAM burnout.

It sounds crazy, but prompt fan replacement is essential to preventing server downtime.

You need a trusted spares partner

Here at CDS we not only know when components are likely to fail, but we also maintain a complete inventory of OEM-approved spares. With international parts warehouses holding every part for every system we support, our customers know that we have them covered when something does go wrong.

To learn more about our spare parts replacement service and how it can keep your servers running optimally, please give us a call.

Download article as a PDF - Which Server Components Are Most Likely To Fail?

More Articles

Server surrounded by recycling symbol (three arrows in a triangle)

The Ultimate Server Recycling Option

Recycling unwanted server hardware is an established principle, but there’s a less well-known way to claw back any residual value.

DNA double helix with binary code as rungs

DNA Storage Edges Closer To Production Readiness

The concepts behind DNA data storage are proven. Now Microsoft is demonstrating a potential production-ready device for your data center.