A freelance client called me in about a CCTV system that had been misbehaving — recordings dropping out, playback that wouldn't load reliably. The assumption going in, from everyone who'd looked at it before me, was a failing hard drive. It's the obvious suspect, and it's usually also the most expensive one to replace unnecessarily.
Isolate before you replace
Instead of swapping the drive first and hoping, I worked through the system in order: power, then the compute box running the recording software, then the drive itself, then the connections between them. Each piece tested independently rather than assumed guilty by association.
The drive, when checked on its own, was fine. The actual problem was the compute box — an ageing NVR unit that was struggling to keep up and quietly corrupting what it handed to the drive. From the outside, that looks identical to a drive failing. From the inside, it's a completely different fix.
The expensive-sounding fault and the correct fault are not always the same fault.
Why it mattered
Replacing the hard drive would have cost the client money and changed nothing — the corruption would have continued on a brand new disk. Identifying the compute box as the real cause meant the actual fix was smaller than everyone expected, including me at the start of the job.
Alongside that, I planned and installed some structured network cabling for the same client — the less glamorous half of the job, but the part where clean cable management and a verified end-to-end connection quietly prevent the next fault before it happens.