Hard Drive Reliability - Study Inc.

[MetalLobster]

I thought this info. might be useful for anyone looking to build a PC or server.

EDIT: Now that I'm home, I can post some more key points.

"Only 10% of the hard drives in Backblaze's storage pods come from WD, and they're strictly low-power Green and Red models"

"In Backblaze's storage pods, Seagate's Barracuda 1.5TB has an annual failure rate of over 25%.The 5,400-RPM version of that drive fares better—its failure rate is only 10%—but that's still pretty high compared to the competition"

"two drives proved to be so unreliable in Backblaze's storage pods that they were left out of the totals completely. Seagate's Barracuda LP 2TB and WD's Green 3TB "start accumulating errors as soon as they are put into production,"

" Backblaze says the Seagate drives are also more prone to dropping out of RAID arrays prematurely. The company uses consumer-grade drives that aren't designed explicitly for RAID environments, of course, but that doesn't seem to bother the Hitachis."

http://techreport.com/news/25940/hard-drive-reliability-study-names-names

tldr:

 

[OG buckshot jr]

Iiiiiiiiiiiiiiiiiiiiinternets! I always wondered why my Macbooks came with Hitachi drives...

 

[zackychuu]

Heh...
I'm guessing this only goes for drives that will be used quite a lot?

My dad got me a 1 TB Ext Seagate drive for Christmas.
But I'm not pushing it's limits or anything..just using it for Film/TV show storage.

 

[Locate☣]

LOLS! those hitachi numbers are total bs XD, those things are bricks! Every single hitachi drive Iv'e ever ownd has died with in a month or 2. Also customers at my repair shop who dought shitty comps all had hatachis and they all died. those this fail way more then 4%

 

[OG buckshot jr]

LOLS! those hitachi numbers are total bs XD, those things are bricks! Every single hitachi drive Iv'e ever ownd has died with in a month or 2. Also customers at my repair shop who dought shitty comps all had hatachis and they all died. those this fail way more then 4%

"
In November, online backup provider Backblaze published some interesting statistics on hard drive mortality based on over 25,000 units in active service. It found that failure rates were higher in the first 18 months and after three years. Those conclusions matched the findings of other studies on the subject, but frustratingly, they didn't include information on specific makes and models." - Link provided in OP.

I dunno bud, this study sounds much more sound than your opinion.

 

[Brades]

I think I put Seagate HDD's in my media server... I will have to check when I get back to Canadastan

 

[Cock]

Fuuuuuuuck

Edit:Also
Füüüüüüüüüüüüüüüüüüüü


I bought 3TB Seagates for my media center

 

[Shotgun Jesus]

SeaAIDS

 

[Cock]

LOLS! those hitachi numbers are total bs XD, those things are bricks! Every single hitachi drive Iv'e ever ownd has died with in a month or 2. Also customers at my repair shop who dought shitty comps all had hatachis and they all died. those this fail way more then 4%

This has been my experience also and to be fair they aren't testing 2.5" drives.

 

[MetalLobster]

Heh...
I'm guessing this only goes for drives that will be used quite a lot?

My dad got me a 1 TB Ext Seagate drive for Christmas.
But I'm not pushing it's limits or anything..just using it for Film/TV show storage.

These are numbers come from Backblaze, so the hard drives are running 24/7.

Unless you're running servers, chances are your hard drives will last longer than what Backblaze experienced. However, these numbers do indicate who makes more quality drives.

 

[OG buckshot jr]

There's far too many variables for this to be perceived in any way - model of drives, SAS drives?, etc. are missing from the report (purposely). I strongly believe there's many renditions of the same model as well (i.e. a 1TB [brand-name-here] might have several versions as per requested by retailers) that might perform differently based on price-points.

A very general study, when considering the terms above, but nonetheless interesting...

 

[MetalLobster]

There's far too many variables for this to be perceived in any way - model of drives, SAS drives?, etc. are missing from the report (purposely). I strongly believe there's many renditions of the same model as well (i.e. a 1TB [brand-name-here] might have several versions as per requested by retailers) that might perform differently based on price-points.

A very general study, when considering the terms above, but nonetheless interesting...

Hard drives are (usually) manufactured in one place, since the majority of parts of similar. The differences between hard drives within a company tends to just be the interface, cache size, disk size (if making 2.5 and 3.5 drives) and maybe (big maybe) motor, but the majority of failure comes from the mechanical, like spindle loosening up, motor dying, or the read/write heads dinging up the disk. An article like this would not make sense if say, they were researching failure rates in Toyota cars, since one model of Toyota, say the Corolla, are manufactured in many places for logistic reasons.

You are right that there are many different models, which can have various failure rates, and this article does address that (see edit above). Since this is just from Backblaze, we can assume they buy all their drives annually in bulk, from one retailer, and store them in the backroom.

The take-away is that Seagate is needs to up their QA and QC (at least for the drives researched) and Toshiba's (HDD models used by Backblaze) got their shit together.

 

[OG buckshot jr]

Hard drives are (usually) manufactured in one place, since the majority of parts of similar. The differences between hard drives within a company tends to just be the interface, cache size, disk size (if making 2.5 and 3.5 drives) and maybe (big maybe) motor...

Uhh, nope. Speaking from experience (both retail and corporate) regarding situations like these, when a developer such as Seagate, for example, develops a hard-drive (let's call it a Barracuda 1TB), they will sell the product to a bunch of different retailers and e-tailers. Now, depending on price-lines, different retailers and e-tailers will want to pay different prices. Some retailers/e-tailers regard themselves as 'high-end' and only prefer the best (let's use Apple for this example) and other retailers/e-tailers prefer to pay much less as they market themselves more towards an affordable demographic (let's use Staples for this example.

Apple wants the best, and thus Seagate will oblige, and will use the best firmware it has (i.e. Seagate Barracuda 1TB 1000.15 -> The '.15' denotes the firmware), and it will also use better quality components such as a rubber gromets, dampening pieces, better metal composites on the head and actuator arm etc. This version will have something like a 'Revision 1A' and a different Model #, but same name as mentioned above.

Staples wants the best it can get at the price it intends to pay, and as such Seagate will augment its cost by supplying the very same Seagate Barracuda 1TB name under a slightly less quality drive with lesser/older firmware [i.e 1000.10] (perhaps with even more less-capacity platters), slightly less quality components such as ... read the paragraph above, instead of re-writing everything. This version will have something like a 'Revision 1B' and will carry a different Model #, but the same name.

This is the way most products will be built and sold. This is also why, as you pointed out yourself by referring to the assumption that the test contains products bought at bulk from one place, it matters very much from where you buy something, as you're not always getting what you think you're getting and why I said it's a very general study.

 

[MetalLobster]

Uhh, nope. Speaking from experience (both retail and corporate) regarding situations like these, when a developer such as Seagate, for example, develops a hard-drive (let's call it a Barracuda 1TB), they will sell the product to a bunch of different retailers and e-tailers. Now, depending on price-lines, different retailers and e-tailers will want to pay different prices. Some retailers/e-tailers regard themselves as 'high-end' and only prefer the best (let's use Apple for this example) and other retailers/e-tailers prefer to pay much less as they market themselves more towards an affordable demographic (let's use Staples for this example.

Apple wants the best, and thus Seagate will oblige, and will use the best firmware it has (i.e. Seagate Barracuda 1TB 1000.15 -> The '.15' denotes the firmware), and it will also use better quality components such as a rubber gromets, dampening pieces, better metal composites on the head and actuator arm etc. This version will have something like a 'Revision 1A' and a different Model #, but same name as mentioned above.

Staples wants the best it can get at the price it intends to pay, and as such Seagate will augment its cost by supplying the very same Seagate Barracuda 1TB name under a slightly less quality drive with lesser/older firmware [i.e 1000.10] (perhaps with even more less-capacity platters), slightly less quality components such as ... read the paragraph above, instead of re-writing everything. This version will have something like a 'Revision 1B' and will carry a different Model #, but the same name.

This is the way most products will be built and sold. This is also why, as you pointed out yourself by referring to the assumption that the test contains products bought at bulk from one place, it matters very much from where you buy something, as you're not always getting what you think you're getting and why I said it's a very general study.

The drives, regardless of who buys them, still leaves the same door. The drives are made, binned (as you noted), and sold, but what's to say their QA and QC processes changes between when the drives are binned and then shipped? The QC/QA process between models may be different, but they are the same department nonetheless and are usually done by the same people. I can speak from experience that while different products may be QA/QC differently, the guidelines of what is acceptable or not are the same. Thus, different models within the same product line go through the same QA/QC processes, done by the same people.

That's not to say that the failure rate of model X will be the same as model Y (as you pointed out), but a loose, sloppy and/or lazy QA/QC standard and guidelines are sure ways to lose the confidence of informed consumers.

 

[OG buckshot jr]

The drives, regardless of who buys them, still leaves the same door.

Irrelevant.

I can speak from experience that while different products may be QA/QC differently, the guidelines of what is acceptable or not are the same.

This is contradictory. Guidelines of what is acceptable is, in effect, QA/QC. You are right, that the QA/QC for an entire product line will be the same. This is the very process that allows shit versions of the same product out of the same door you talk about. You're validating my point.

That's not to say that the failure rate of model X will be the same as model Y (as you pointed out), but a loose, sloppy and/or lazy QA/QC standard and guidelines are sure ways to lose the confidence of informed consumers.

I can see where you're going with this, but I don't think it's QA/QC to blame for this. It's the product itself. It all matters what revision/version you're buying, and as a consumer, that's your responsibility. Do you go TigerDirect and buy the drive at $40 or do you go somewhere else and buy the drive for $69? There's reasons for pricing, always.

 

[MetalLobster]

Irrelevant.


This is contradictory. Guidelines of what is acceptable is, in effect, QA/QC. You are right, that the QA/QC for an entire product line will be the same. This is the very process that allows shit versions of the same product out of the same door you talk about. You're validating my point.


I can see where you're going with this, but I don't think it's QA/QC to blame for this. It's the product itself. It all matters what revision/version you're buying, and as a consumer, that's your responsibility. Do you go TigerDirect and buy the drive at $40 or do you go somewhere else and buy the drive for $69? There's reasons for pricing, always.

First point, VERY relevant for product quality. Two manu. plants will have it's own QA/QA dept., which are run and managed with it's own people. Hard drive manu. only have one plant.

Second point, ditto. I was singing to the choir.

Third, QC/QA are the "last lines of defense", as my company call it, for ensuring quality. They get first blame for a defective product and take the majority of the responsibility.

I appreciate the convo. but I must fly (class...)!

 

[OG buckshot jr]

Likewise, thanks :)

 

[$alvador]

I've gone through a shitload of drives myself and all but one of all the Seagate drive I've ever owned going back to the late 90s has died. So much micronesian goat porn lost forever.

 

[zackychuu]

I have no clue what other people do to their drives that kills them?
Myself and my friend both bought some HDDs around the same time 2 years ago now, in that time frame, he has killed 2-3 drives, while mine is still going strong.
Mine was second hand from Amazon too...

 

[Glocky]

These are the HDDs in question... Draw your own conclusions about the sample sizes for each drive, and the age of the drives, and the weight of comparing a $300 drives to $125-$200 drives and expecting the same lifespan.

Model - ----------------------------Size - Quantity of HDD - Average Age (years) - Annualized Failure Rate