[NewsPoster]

Cloud-based backup service Backblaze is releasing the data it used to create its recent hard drive reliability study. The data collection covers over 41,000 hard drives, used by the company in its custom Storage Pods to hold customer data, and is said to be the "largest data set on disk drive performance" to have ever been provided to the public.

The data, covering both 2013 and 2014 in their entirety, stems from a daily snapshot of drive states Backblaze performs every day in its data center. The data includes the model and serial numbers, along with SMART data, itself including the number of hours the drive was active for, drive temperature, bad sector details, and other information.

"There are lots of smart people out there who like working with data, and you may be one of them," the company writes in a blog post. "Now it's your turn to pore over the data and find hidden treasures of insight. All we ask is that if you find something interesting, that you post it publicly for the benefit of the computing community as a whole."

In its most-recent reliability report, Backblaze discovered that HGST drives continued to be the most reliable producer of its drives, closely followed by Western Digital. One 3TB Seagate drive was found to have a massive 43.1 percent annual failure rate, though its relatively reliable 4TB drive was more favorable compared to HGST's version due to its lower cost.

Another similar report based on data from 34,000 hard drives concluded that there was no overall correlation between the temperature of a drive and its chance of failure.

[prl99]

Doesn't surprise me one bit about Seagate drives. This study only deals with full size drives, no laptop drives, but it confirms my position on never buying a Seagate drive again. I've had early death problems with Seagate drives before. I've bookmarked the Backblaze page and will consult it next time I need drives. The last drive I purchased was a replacement laptop drive from HGST so I believe I made the right decision even though it was a desktop drive.

[JackWebb]

This completely confirms what I've seen. Every Seagate I've ever owned no longer works except for a couple I removed before they had a chance. None of my IBM/Hitachi/HGST have ever failed. I've observed this since my first seagate failure of a 2TB drive in 1995. I've never knowingly bought a Seagate for this reason but they have come in certain Macs and needed to be replaced. When purchasing external HDDs or Macs I've tried to find out if it had a Seagate to avoid that. I was already firmly of this opinion by 2000. It's amazing the brand reliability has remained the same all these years. Why is Seagate still in business?

[JackWebb]

I meant 2GB in 1995.

[climacs]

Seagate drives are usually the least expensive... there's a reason for that. Don't go for cheapest on your HD's.

[bjojade]

They don't detail which models of drives they are using. I use a lot of WD drives, and I can say for certain that the drive life between the lines is dramatically different. The WD blacks will outlive the WD greens by a long margin. Of course, they cost more as well.

[Mike Wuerthele]

All they're using is the straight cheap consumer-level drives. No Blacks, or purples.

[waffffffle]

I had a 3TB Seagate drive in an iMac that failed recently. Looks like I'm not alone.

[Cap'n Tightpants]

Since the drives aren't all subjected to the same loads, temps, or other controllable variables, this doesn't mean very much.

[OreoCookie]

Backblaze's blog posts on the topic were the reasons I chose two HGST hard drives for my new Synology NAS. The failure rates of the Segate drives scared the living sh*t out of me -- two of my three external hard drives use Segate hard disks.

[Mike Wuerthele]

Originally Posted by Cap'n Tightpants 

Since the drives aren't all subjected to the same loads, temps, or other controllable variables, this doesn't mean very much.

This isn't accurate. While I agree that all drives aren't subject to the same loads, etc, the environments are close enough with a large enough population that it can be essentially disregarded. Its not like some of the drives are in the back of a pickup truck bouncing down the road in Arizona in the sun while taking data, while the rest are babied.

If anything, the failure rates are slightly exaggerated, giving a de facto time compression to the results, which I'm fine with. The staggering failure rates on that 3TB model can't be ignored.

[Cap'n Tightpants]

I thought we were using Seagate in our storage arrays (80TB and 120TB worth), but it turns out they're Hitachi (HGST) Ultrastars (which have been faultless, not a single error). Funny how tables turn, I guess, at one time IBM Deskstar drives were so bad, the service depot I worked for called them Deathstars. Seems the change of ownership did them some good.