How Seagate Tests Its Hard Drives

[Anders235]

Holy crap.

At this point I'm beginning to wonder if Backblaze's owners have some kind of personal issue with Seagate. All their catamites pop up everywhere on the internet citing backblaze's unscientific horse-sh*t reliability stats to 'prove' Seagate makes terrible products.

Backblaze shite psudo tests + "my friend in IT says" = proof positive that Seagate's products are crap! Erm, no.

 

[Anders235]

Holy crap.

At this point I'm beginning to wonder if Backblaze's owners have some kind of personal issue with Seagate. All their catamites pop up everywhere on the internet citing backblaze's unscientific horse-sh*t reliability stats to 'prove' Seagate makes terrible products.

Backblaze shite psudo tests + "my friend in IT says" = proof positive that Seagate's products are crap! Erm, no.

 

[rubberjohnson]

Seagate drives are terrible! Out of 150 lenovo desktops with 1TB drives we run in house, about 20% of the seagates have died, the previous model used WD and only about 3 of them have died from those 150.

 

[f-14]

i saved this article and that shot of the peeling warranty sticker from that one drive, going to pick a bone with seagate tomorrow.

 

[Ra_V_en]

The article explains the different types of drive/MTBF and why the backblaze test is useless information. Marketing plot to have folks talking about it and re-posting its link. It seems to work as we keep seeing the link over and over... They are not getting my data. They put drives designed for desktop into servers and run them to the ground and call it a "reliability test". Let's test my kids bicycle with training wheels at the Tour de France and complain about its quality....

You are a joker...
From what you say some might understand that they should not even ask for any reliability beacause they bought cheapest drive out there. If so Seagate should name the drives properly so nobody even look at it, instead of Barracudas it should be named Failarcudas.
Now to the testing methods, they can throw them from the moon as far as you get any reasonable conclusions. Who the f... said we are not allowed to test quality of any drive with any methods even if the drive cost $0.01 and even if it's not a workstation/server grade one. A buyer has a right to make decision based on any facts and any methods not the one allowed by the manufacturer because product look good at it.

 

[jimmysmitty]

The article explains the different types of drive/MTBF and why the backblaze test is useless information. Marketing plot to have folks talking about it and re-posting its link. It seems to work as we keep seeing the link over and over... They are not getting my data. They put drives designed for desktop into servers and run them to the ground and call it a "reliability test". Let's test my kids bicycle with training wheels at the Tour de France and complain about its quality....

You are a joker...
From what you say some might understand that they should not even ask for any reliability beacause they bought cheapest drive out there. If so Seagate should name the drives properly so nobody even look at it, instead of Barracudas it should be named Failarcudas.
Now to the testing methods, they can throw them from the moon as far as you get any reasonable conclusions. Who the f... said we are not allowed to test quality of any drive with any methods even if the drive cost $0.01 and even if it's not a workstation/server grade one. A buyer has a right to make decision based on any facts and any methods not the one allowed by the manufacturer because product look good at it.

Would you buy a compact car based on its performance off road? Short of a rally style car but lets say a Ford Focus or a Honda Civic was put through a off road test and it failed horribly. Does that tell you how long it will last in normal road conditions?

Or better yet, imagine testing enterprise grade HDDs in a consumer environment. They will probably last forever but it tells you nothing as to how well the HDDs will last/perform in a enterprise environment.

That is the point. I would not expect a consumer product to last as long in a enterprise environment as an enterprise product much as I would not expect a compact car to survive off road as long as a truck or vehicle designed for it.

 

[Ra_V_en]

Would you buy a compact car based on its performance off road? Short of a rally style car but lets say a Ford Focus or a Honda Civic was put through a off road test and it failed horribly. Does that tell you how long it will last in normal road conditions?

Or better yet, imagine testing enterprise grade HDDs in a consumer environment. They will probably last forever but it tells you nothing as to how well the HDDs will last/perform in a enterprise environment.

That is the point. I would not expect a consumer product to last as long in a enterprise environment as an enterprise product much as I would not expect a compact car to survive off road as long as a truck or vehicle designed for it.

Comparison to the car is a fail but in fact even I could even respond to that. It depend of the test, if they tested performance going over the hill then that might not interest me at all but if they test how long it take for the wheels to fell off or engine to break down then that is some really interesting information.
Testing something in heavy conditions is a good way to find out a general durability. In fact almost most of the tests which are related to wearing which is a long term process are done with smaller time sample but under higher tension over the material because there is no other way to do it.

Then you talk about putting enterprise grade HDD's for a consumer tests.. I'm not sure what kind of silly logic is that. It's darn obvious putting lower load on something which is called heavy duty will tell you nothing, but putting higher load on something not so heavy duty will tell you quite something. And that was is blackblaze about.
They nailed the truth about few Seagate series and that's it, nothing more. If there is any other non sponsored source of information I'll be glad to look at it, but since there isn't and this one actually confirms my experience with some of those series I don't need any further confirmation.

Whatever methods Seagate is using to test their drives they seem to fail in consumer market.... and well enterprise grade is a whole different story, I don't need Seagate testing methods, HP, DELL, and so on are testing drive series themselves, you can choose only from confirmed models and you pay decent pile of cash for that certification sticker. There is no other way than that if you rent or lease servers and/or workstation on large scale and want to keep the warranty.

 

[jimmysmitty]

Would you buy a compact car based on its performance off road? Short of a rally style car but lets say a Ford Focus or a Honda Civic was put through a off road test and it failed horribly. Does that tell you how long it will last in normal road conditions?

Or better yet, imagine testing enterprise grade HDDs in a consumer environment. They will probably last forever but it tells you nothing as to how well the HDDs will last/perform in a enterprise environment.

That is the point. I would not expect a consumer product to last as long in a enterprise environment as an enterprise product much as I would not expect a compact car to survive off road as long as a truck or vehicle designed for it.

Comparison to the car is a fail but in fact even I could even respond to that. It depend of the test, if they tested performance going over the hill then that might not interest me at all but if they test how long it take for the wheels to fell off or engine to break down then that is some really interesting information.
Testing something in heavy conditions is a good way to find out a general durability. In fact almost most of the tests which are related to wearing which is a long term process are done with smaller time sample but under higher tension over the material because there is no other way to do it.

Then you talk about putting enterprise grade HDD's for a consumer tests.. I'm not sure what kind of silly logic is that. It's darn obvious putting lower load on something which is called heavy duty will tell you nothing, but putting higher load on something not so heavy duty will tell you quite something. And that was is blackblaze about.
They nailed the truth about few Seagate series and that's it, nothing more. If there is any other non sponsored source of information I'll be glad to look at it, but since there isn't and this one actually confirms my experience with some of those series I don't need any further confirmation.

Whatever methods Seagate is using to test their drives they seem to fail in consumer market.... and well enterprise grade is a whole different story, I don't need Seagate testing methods, HP, DELL, and so on are testing drive series themselves, you can choose only from confirmed models and you pay decent pile of cash for that certification sticker. There is no other way than that if you rent or lease servers and/or workstation on large scale and want to keep the warranty.

The point is unintended situations. A compact car is not designed to handle the rough off-road conditions, it is designed to handle normal streets. It will fail horribly due to it not being designed for it.

And the point, again you are missing it, is if you put something into an environment that it was not designed for you wont get the results you would if you look at the real environment.

I could make a test where I power on then shake HDDs like crazy and put together the results but that data is flawed since the drives are not meant to be moved while running. These "tests" that backblaze do are flawed because they use consumer grade products for non consumer grade applications and use a "pod" design that doesn't even mount the drives properly and even they have had to redesign the mounting to reduce the vibrations of the drives, which is a big killer of drives.

Look at a SAN or server rack. The drives are screwed tightly into a holder then slid and locked tightly into a bay. There is very little wiggle room. Then look at backblazes "Pod". The drives in some of the designs have nothing locking them into place.

Another note is that a lot of consumer drives try to fix I/O errors when a RAID controller will and this can cause a lot of issues, such as using a WD Green drive in a redundant RAID array can be declared dead by the RAID controller in this situation. I know as I have seen it happen when a guy had us, against our suggestion, a server using 8 WD Greens. At least once a week he would come in with multiple "bad" HDDs that were fine but the RAID controller said they were bad. Enterprise grade drives will not do this.

Again, the application of the drive matters immensely.

 

[Tzn]

Damage control, too late everyone knows Seagate by now.

 

[littleleo]

I haven't gone to the Blaze website, but the problems about the high failure with the Seagate 3 TB are well known. On the bright side WD discontinued the green drives so you won't have to deal with those in the future Jimmy. They were replaced in many cases by the New WD Blue 5400RPM drives. Just went you thought desktop 5400RPM drives were gone forever..... THEY ARE BACK!

 

[kiruwa]

The article explains the different types of drive/MTBF and why the backblaze test is useless information. Marketing plot to have folks talking about it and re-posting its link. It seems to work as we keep seeing the link over and over... They are not getting my data. They put drives designed for desktop into servers and run them to the ground and call it a "reliability test". Let's test my kids bicycle with training wheels at the Tour de France and complain about its quality....

Sure, the backblaze pdf is backblaze marketing, but that's completely irrelevant to the fact that it DOES reflect badly on Seagate's quality.

If your kid's bike fell apart on the Tour de France roads, I'd hope you'd consider a different brand next time.

 

[jimmysmitty]

I haven't gone to the Blaze website, but the problems about the high failure with the Seagate 3 TB are well known. On the bright side WD discontinued the green drives so you won't have to deal with those in the future Jimmy. They were replaced in many cases by the New WD Blue 5400RPM drives. Just went you thought desktop 5400RPM drives were gone forever..... THEY ARE BACK!

The WD Greens were combined into the WD Blue line. They are the same drives just under the Blue label. They would still have the same issue though, most consumer drives are not designed to be put into a redundant array.

I don't have to deal with it luckily as I no longer work in a repair shop and all of our servers use enterprise grade drives.

 

[Arbie]

To all of you bad-mouthing Seagate: Their drives ARE tested. The customers do it, and even pay for the privilege. I myself tested several 3TB drives for them. Those drives failed after about one year of light use, on average.

So there.

 

[littleleo]

To all of you bad-mouthing Seagate: Their drives ARE tested. The customers do it, and even pay for the privilege. I myself tested several 3TB drives for them. Those drives failed after about one year of light use, on average.

So there.

Wow thanks for letting us know.

 

[Ra_V_en]

The WD Greens were combined into the WD Blue line. They are the same drives just under the Blue label. They would still have the same issue though, most consumer drives are not designed to be put into a redundant array.

I don't have to deal with it luckily as I no longer work in a repair shop and all of our servers use enterprise grade drives.

I'm not gonna argue anymore about BlackBlaze... you seem to have own opinion I do mine. As far as I never blindly believe web tests if they are not confirmed by any additional sources, but fortunately here I can also confirm some similarities based on my own observation on returning to service drives.

Now the "Blue" series from WD is a weird thing, I might consider this move being a trick as Kingson did with V300 series... bite and switch. We all know EZEX being quite successful, single platter with firmware design well above average. I started to use those drives well before a hype arrived and BlackBlaze even started to test them, just reading about design was enough to convince me that might be a deal breaker on that price range.
I'm not being a conspiracy theorist but there is few clues that WD is getting greedy and I'm about to be more careful about advising them and certainly I'll start putting a disclaimer notes about the specific models.

There are still 7200 RMP models on the market, even on WD website you can see it:
http://www.wdc.com/en/products/products.aspx?id=770#Tab3

So at least those who know the model will choose more carefully, the others, well save them God...

When you look more closely WD sells both Green and Blue with different labels, example WD10EZRZ.
Blue advertised as 5400 model Green as IntelliPower
Now that might mean 2 things:
- both are exactly the same models and WD is not being a consequent in the tech naming
- both have different firmware to control them thus Blue are not called IntelliPower
If they changed the firmware both might not have the same performance and/or reliability which is a subject to further testing...

I do know WD representatives hanging around this forum so they might clear this up, because it smells fishy as hell and all that good reputation might be wasted very soon with such actions.

 

[coyote2]

I also want the fastest drives. Seagate's aren't.

And my experience with Seagate drives match stats like https://www.backblaze.com/blog/hard-drive-reliability-stats-for-q2-2015/ that show Seagates as LESS reliable.

So I buy WD drives.

 

[Bossyfins]

^ We pretty much declared, in our minds, whether Blackblaze was reliable or not. Personally, no

"These "tests" that backblaze do are flawed because they use consumer grade products for non consumer grade applications and use a "pod" design that doesn't even mount the drives properly and even they have had to redesign the mounting to reduce the vibrations of the drives, which is a big killer of drives." - jimmy

 

[JimmiG]

Seagate are still taking a lot of crap for the ST3000DM001 fiasco. That drive definitely had real problems - we used a bunch of them in our NAS at work and most of them have failed at this point. The older WD 2GB drives are still working fine.
It would probably be better for Seagate to admit that those drives had a manufacturing defect, at least then maybe people would begin to trust their other drives which seem to be reasonably reliable.

 

[Bossyfins]

^ You can't tell people their is a defect when they can't really find a defect.
1. If they had a defect, they probably would have gotten to fixin it,
2. If they had a defect, and couldn't solve it, then they would probably remove that specific technology.

Plus, if they did say there is a defect, lawsuits could go endless for data recovery services.

 

[JackNaylorPE]

How many times a month does the backblaze data have to be dubunked here on THG ? Ya would think by now peeps would stop quoting it. THG just finished an article on the faulty study.

http://www.tomshardware.com/news/seagate-hdd-failure-lawsuit-3tb,31118.html



Brilliant setup there over at backblaze....lets take some hard drives, put them in cheap storage rack held in with rubber bands. Then lets buy a bunch of drives whose firmware includes a head parking feature, which while essential in a consumer environment is a death knell for server applications.

Head parking "parks the drives" when not in active use. This feature saves many a consumer drive saving the heads from crashing into the platters on bouncy wood floors ad against that bump into your desk by an e3mployee or ya dog a home. Now take a drive equipped with this protection feature and stick it in a server environment so the drive arm swings back and forth more times in 10 minutes than Chris Davis struck out all of last year (208 strikeouts on 573 at bats).

If backblaze had any clue at what they were doing, they would by the exact same drives mechanically but utilize a model in which head parking was disabled...problem solved.

If you want to see how a consumer drive does in a consumer environment, just look at what % of sales are returned under warranty. The information is readily available and remarkably consistent:

http://www.hardware.fr/articles/934-6/disques-durs.html
http://www.hardware.fr/articles/927-6/disques-durs.html
http://www.hardware.fr/articles/920-6/disques-durs.html
http://www.hardware.fr/articles/911-6/disques-durs.html

 

[littleleo]

All I can tell you is what my customer's are experiencing and that is higher then normal failure rates on Seagate 3TB drives, and lately 1TB & 2TB drives have started running higher RMAs than normal. I think this is an anomaly, but it is something that is affecting customer's purchasing patterns. I'm getting more of the "any brand but Seagate" right now. When I ask them why they tell me they are having failure issues with them.

Using desktop drives for server type applications is always been an issue and now it is happening with SSDs. I've had many arguments with SI about it but they still rationalize and do it. If a company like Backblaze is doing it that is just plain stupid. Then again I wouldn't use a backup service like that in the 1st place. I have a NAS I back-up to, and real important files I also burn on to a disk for a hard copy.

 

[Jeff Wiles]

My name is Jeff Wiles and I setup and supervised Seagate Technology's first Reliability Engineering Testing Laboratories : (AKA) Environmental Testing, Shock, Vibration, Acoustics and Thermal Testing Areas for the Design & Reliability Engineering Departments starting in 1982 until 1990 and then 1993 until 2001. The testing we performed consisted of Operating & Non-Operating Shock & Vibration and was performed to all Design Engineering specifications as most of our drive designs at the time had sensitive environmental specifications that were nowhere near the newer design spec's of today. Our ( DMT ) Design Maturity Testing & (ORT) Ongoing Reliability Testing ( which was equivalent to our products 5 year warranty at the time passed & exceeded this testing criteria ). Unfortunately, the majority of the failures submitted to our Seagate Repair Centers were directly related to ( ESD ) Electro-Static-Discharge damage to the Printed Circuit Boards which held several ESD sensitive electronic components that were sold directly to the public. I have first hand knowledge of these issues because I flew around the entire country designing, building and setting up the Board Swap Repair Centers for 3 years at all of our strategic disc drive distribution facilities all over the United States. A lot of individuals have come into this forum with complaints because they do not understand the true nature of the problems or the testing procedures of our environmental labs at the time or even now. However, I can personally assure you that the company has also devoted millions upon millions of dollars and 1,000's of square feet to Failure Analysis Labs to address the problems that we have performed failure analysis on and then resolved the issues created by them just as fast as they were discovered. ( JW )

 

[Jeff Wiles]

If you are still determined to find a reason as to why the company now has design problems, look into their 2 year warranty instead of the 5 year warranty they once had to understand the current reliability issues. I suggest that you lookup the manufacturing facilities that are in China that are now producing the majority of Seagate Disc Drives and endless amounts of other American designed products that so many of us have lost our jobs to. I have been over there and I have seen lots of people driving around in very nice $60-70,000 dollar American cars and trucks. It kind of makes a person wonder where our prosperity and economy went all of a sudden.......

 

[kwa-e]

All's good in capitalism they say...

 

[Jeff Wiles]

To all of you bad-mouthing Seagate: Their drives ARE tested. The customers do it, and even pay for the privilege. I myself tested several 3TB drives for them. Those drives failed after about one year of light use, on average.

So there.

Wow thanks for letting us know.