Skip to content

Why You Shouldn't Stress Test HDDs Unless You're Trying to Maximize Uptime

Jump to bottom
jdrch edited this page Sep 24, 2019 · 1 revision

Such testing makes the drive more likely to fail anyway. In other words, the "test" just becomes a useless self-fulfilling prophecy. Brand new drives are protected by warranty. If one fails on you, just get a free replacement from the OEM.

If HDD failure is such a huge risk, something is wrong with the workflow. The entire point of datahoarding is to eliminate data risk from hardware failure. So, to me, it doesn't matter when the drive fails, BUT I make my best effort to prolong its life as long as possible.

A drive that passes a torture test is still gonna fail at some point in the future. Modern HDDs rarely die immediately; typically some SMART parameter threshold gets crossed first. This means that you can always do an online, pre-failure replacement at any time said threshold event occurs, whether it's within or outside of warranty.

TL,DR: Torture testing drives doesn't reduce your redundancy/emergency requirements. It may decrease your downtime, but that's an issue for "five 9s" (99.999% uptime) service providers, not most operators.

Sources:

Organized Alphabetically:

Clone this wiki locally