Backblaze report: 8-terabyte HDDs lasting a long time - HGST Most Reliable

Published 2017-02-01 09:02 by Hilbert Hagedoorn

Click here to post a comment for Backblaze report: 8-terabyte HDDs lasting a long time - HGST Most Reliable on our message forum

#5387250

Razer82

2017-02-01 09:16

If I am not mistaken, I believe that Toshiba bought the HGST 3.5-inch business from WD because of requirements of regulatory agencies. So I think reliability is on the side of Toshiba.

#5387297

slyphnier

2017-02-01 11:01

If I am not mistaken, I believe that Toshiba bought the HGST 3.5-inch business from WD because of requirements of regulatory agencies. So I think reliability is on the side of Toshiba.

its quite complex but its safe to say HGST = WD based http://www.enhancedonlinenews.com/portal/site/eon/permalink/?ndmViewId=news_view&newsId=20120228007152&newsLang=en&permalinkExtra= OR https://en.wikipedia.org/wiki/HGST In May 2012, WD divested to Toshiba assets that enabled Toshiba to manufacture and sell 3.5-inch hard drives for the desktop and consumer electronics markets to address the requirements of regulatory agencies so toshiba making their own 3.5 HDD, thats why on early take over there lots rebranded hitachi hdd, there no longer now we can say toshiba hdd = hitachi hdd, as its using hgst asset to manufacture their hdd but hitachi hdd is not toshiba hdd so cant say hitachi hdd same to toshiba hdd / reliability, as toshiba not producing hitachi hdd

#5387301

coth

2017-02-01 11:09

45 drives isn't really statistics

#5387348

Kaarme

2017-02-01 13:33

its quite complex but its safe to say HGST = WD

Considering WD seems to be falling behind in reliability, it seems like all of their efforts go to the king HGST, while the consumer WD branded drives are an afterthought. Or maybe this is a purposeful action, to remain both in the more expensive (profitable) business, but also to cut costs mercilessly.

#5387369

Ma Lau

2017-02-01 14:05

Red herrings? The numbers are difficult to use and draw out conclusions. It seems the report does not show failure rates of drives at 1-year-old, 2-year-old etc. E.g. The numbers are showing that older drives have higher failure rates. However we are more interested to know which brand/model has more reliable drives, all else being equal including age. Or have I misunderstood something?

#5387388

Reardan

2017-02-01 14:50

The numbers are difficult to use and draw out conclusions. It seems the report does not show failure rates of drives at 1-year-old, 2-year-old etc. E.g. The numbers are showing that older drives have higher failure rates. However we are more interested to know which brand/model has more reliable drives, all else being equal including age. Or have I misunderstood something?

I don't know that that's true. The HGST drives are the oldest drives in the list and have the highest reliability.

#5387439

alanm

2017-02-01 16:10

Quite an improvement for Seagate, and worse showing for WD. Earlier stats (2013-2015): https://www.backblaze.com/blog/wp-content/uploads/2015/10/blog_q3stats_manufacturer-e1444680042365.jpg

#5387538

tunaphish6

2017-02-01 19:56

Every statistician is collectively rolling their eyes at these charts. I'm not beyond saying Seagate has improved their brand, but as long as you're not throwing your drives across the room, I've yet to have a Western Digital drive fail (given, I'd had to disable head parking on all my Green/Blue drives).

#5387600

alanm

2017-02-01 22:40

Every statistician is collectively rolling their eyes at these charts. I'm not beyond saying Seagate has improved their brand, but as long as you're not throwing your drives across the room, I've yet to have a Western Digital drive fail (given, I'd had to disable head parking on all my Green/Blue drives).

Doesnt make sense. Statisticians would have less issue with large data points (thousands of drives) than they would with any single users anecdotal evidence of a few drives.

#5387735

HeavyHemi

2017-02-02 08:09

Doesnt make sense. Statisticians would have less issue with large data points (thousands of drives) than they would with any single users anecdotal evidence of a few drives.

For example, that chart you posted lack the metrics necessary to make any sort of valid comparison. At the minimum we need age, hours, install base, capacity a few more come to mind. So he's not talking about the data, he's talking about the presentation.