Backblaze Hard Drive Reliability 2014 Figures

Backblaze Storage Pod v4 Power Supply
Backblaze Storage Pod v4 Power Supply

Backblaze made some ripples in the industry when they released their reliability figures for hard drives a year ago. Many companies were caught off-guard and were quick to say that Backblaze was using drives not for their intended purpose. Suffice to say, it was a bombshell of a post. Today, Backblaze released their updated figures one year later.

Backblaze ended 2014 with 41,213 disk drives active in their data center. This is an increase of 14,079 or 51% from 2013 year end. Most of these new drives were 4TB and some 6TB drives. The 6TB drives did not fall down the affordability curve far enough to be cost effective during the majority of the year. The Backblaze post above has detailed data on AFR’s by drive, but here is some summary data.

4TB better than 3TB

Probably the biggest takeaway from the data is that the 4TB drives are faring much better than 3TB drives. Backblaze had this chart showing various hard drive failure rates by manufacturer and size. It should be noted that Backblaze does use different models from a given brand at a capacity point so this is an aggregation and there is certainly a large variance in different models.

Backblaze Reliability 2015 by Drive Size
Backblaze Reliability 2015 by Drive Size

The interesting fact here is that the 3TB drives from all vendors had a rough 2014. Each spiked in terms of failure rates. The Seagate 3TB drives were weighed down by the Seagate 7200.14 3TB drives (the Barracuda XT drives were much better.) In terms of the 4TB drives, Backblaze had about 10x the number of Seagate Desktop HDD.15 4TB than they did of the 7200.14 so the much lower 2014 failure rate is impressive.

Backblaze Reliability 2015 4TB Failure Rate
Backblaze Reliability 2015 4TB Failure Rate

The bottom line is that the HGST drives continue to be AFR leaders, as they were in last year’s round-up. Folks on the STH forums have been confirming this believe for years even with their (oftentimes) smaller sample sizes. The Backblaze piece above is certainly worth a read if you are interested in the topic.

Previous articleSupermicro X10DRG-Q Review – GPU Compute Server Motherboard
Next articlepfsense 2.2 Released! Major new version
Patrick has been running STH since 2009 and covers a wide variety of SME, SMB, and SOHO IT topics. Patrick is a consultant in the technology industry and has worked with numerous large hardware and storage vendors in the Silicon Valley. The goal of STH is simply to help users find some information about server, storage and networking, building blocks. If you have any helpful information please feel free to post on the forums.


  1. My only problem with that theory is the drives in the external enclosures are often the same desktop drives they would buy normally. While I think it is important to know that these are shucked drives… I don’t think that should allow us to ignore the failures. Might as well ignore the Red failures since they are only designed for up to 8 drives enclosures.

  2. The drives in external enclosures aren’t always the same as their desktop equivalent drive. Sometimes they have special firmware.

    For example, sometimes they go into suspend after X minutes, regardless of idle times… I think the SAMSUNG D3 Station enclosures did this.


Please enter your comment!
Please enter your name here