Comments on: Hard Drive Stats for Q3 2018: Less is More

By: John Kartal

John Kartal — Thu, 15 Aug 2019 21:16:50 +0000

Thank you so much

By: DO.NUT

DO.NUT — Sun, 02 Jun 2019 23:47:04 +0000

ST2000NC001 just failed after 5 years light usage ….

By: Andy Klein

Andy Klein — Mon, 04 Feb 2019 16:17:40 +0000

In reply to Quentin. Good question. We have in the past published failure rates for the different manufacturers. While that's interesting, it is more accurate to look at just the failure rates of a given model, as different models perform differently in our environment.

By: Carlton Whatley

Carlton Whatley — Sat, 02 Feb 2019 08:52:07 +0000

This is redundant but thank you for posting the hard drive stats – I’ve been following since day 1. Recently though, I’ve been wondering, do you do any kind of monitoring the drives for impending failure – or do you just have to wait until they fail before replacement?

By: Tim Newman

Tim Newman — Wed, 23 Jan 2019 17:05:05 +0000

im a reliability engineer and this data is misleading. microsoft have conducted studies on their datacentres and conclude that the AFR increases exponentially over time and as such these short range figures are effectively useless for any practical reasons. its analogous to concluding that your tyres will last 1 million kms because everyone in your neighbourhood’s tyres havent failed after 100km.

By: French Fries 28

French Fries 28 — Fri, 28 Dec 2018 11:17:57 +0000

I have a server at home which runs a small SSD for a boot drive and a 500GB Seagate desktop hard drive. It is 11 years old and has over 80,000 hours of runtime and it still runs 24/7. The drive still works so I’d see it as a waste to replace it but as stated in this article higher density drives are not too expensive.

By: theGreatBren

theGreatBren — Fri, 14 Dec 2018 03:19:24 +0000

So 4TB is the safe spot for conservatives like me. I don’t want any risk of failures.

By: Zen

Zen — Thu, 29 Nov 2018 01:51:51 +0000

Just want to say, thank you so much for sharing this to public. These data has become our go-to site for real-world industrial knowledge reference.

Would you mind include the definition and severity of “Failure” for better referencing in the future?

Thank you so much!

By: Quentin

Quentin — Thu, 08 Nov 2018 00:39:42 +0000

Thank you for continuing to publish these statistics. I wonder, though, if you now have sufficient data to show reliability by manufacturer over all models you’ve used?