What's up with high hard drive failure rates?

[Compman55]

I am really nervous now about purchasing a new hard drive. I am looking in the 500GB ~ 2TB range at many different models in the SATA 7200PRM catagory. Almost 13% or more of newegg's sales are DOA or death within a short time. What is going on? I never remeber that many failures back in the day. Now with load/unload ramps, fluid bearing motors, and 2 bearing supports for the spindle motors, you woudl think failures would be absolutly minimal.

 

[Red Squirrel]

They don't make em like they used to. Everything is all made cheaply in china now, by kids pretty much working slave wages. It's sad that we've put ourselves in such an economic crisis that we can't even afford to manufacture our own stuff.

That said, I seem to have very good luck with hard drives. I've only had a few failures in my 10 or so years of experience.

 

[nickbits]

newegg has terrible packaging for oem hard disks

 

[Emulex]

unsure what you are talking about. $69 1TB - sure. $49 250gb no problems

 

[RebateMonger]

Disk makers are having a hard time building reliable disks with the high data densites needed for these large-capacity disks. Seagate seemed to be having the worst time, but I'm not sure WD is doing any better.

Any additional problems like poor shipping or handling will aggravate the situation. Things you could get away with two years ago are no longer safe.

A primary reason for the new 4K sectors (called "Advanced Format" by Western Digital) is to allow additional error correction code, which is supposed to reduce the uncorrected error rate by a factor of 1/100.

 

[Jd007]

You can't use Newegg reviews to judge the failure rates of hard drives, as most people (whose drives are operating normally) don't write Newegg reviews. People who get failed/DOA parts are more likely to write a review than those with working parts, so obviously the percent of bad reviews is biased towards failed hard drives.

 

[Ken90630]

I am really nervous now about purchasing a new hard drive. I am looking in the 500GB ~ 2TB range at many different models in the SATA 7200PRM catagory. Almost 13% or more of newegg's sales are DOA or death within a short time. What is going on? I never remeber that many failures back in the day. Now with load/unload ramps, fluid bearing motors, and 2 bearing supports for the spindle motors, you woudl think failures would be absolutly minimal.

I've wondered about that myself, and I'm anxious too about a pending HD purchase I need to make.

I wonder if it's a matter of shipping guys just tossing the boxes around when loading & unloading the trucks. That can't be good for HDs. Plus I'd guess some of the Newegg failures are by teenage gamer geeks who are ultra-hyper (aren't they all?) & in a hurry & maybe they dropped it before installing it or banged it around the case during installation or something. Who knows ....

I always wonder how these "bulk" OEM drives arrive at Newegg & other vendors in the first place. Are they in any kind of protective packaging, or are they just tossed all together in a big box like the FDDs at Fry's? I know Newegg wraps 'em in bubble wrap for us before shipping, but did they arrive at Newegg in bubble wrap or other protective packaging? Anyone know?

The only explanation I've read from the mfgrs goes something like this: "We have to keep our parts at low prices consumers will pay. To do that, we can't afford to implement a level of production line QC that would result in ultra-low failure rates of products shipped. So we figure a 10% (or whatever) failure rate into the mix, and the people who get one of the DOA parts are basically screwed & have to hassle with an RMA. It's the luck of the draw, so good luck."

Not in those words, of course, but that was the gist. Ridiculous -- I'd rather pay a few dollars more for better QC, but I guess the masses don't agree.

All that said, I've only bought one OEM HD in my life, and it was fine. The other 3 or so have been retail boxed & all worked perfectly (2 WDs and 1 Seagate). But I'm with 'ya -- I read the Newegg comments & shudder. :|

 

[ExcaliburMM]

I have yet to have an HDD fail on me, in any machine I've owned.. I guess I'm just lucky.

 

[RebateMonger]

I always wonder how these "bulk" OEM drives arrive at Newegg & other vendors in the first place. Are they in any kind of protective packaging, or are they just tossed all together in a big box like the FDDs at Fry's? I know Newegg wraps 'em in bubble wrap for us before shipping, but did they arrive at Newegg in bubble wrap or other protective packaging? Anyone know?

A couple of times, when ordering several identical disks at once (mostly Seagate 320 GB 7200.10 disks), I received them packed in what I assumed was the Seagate OEM packaging. Disks were in their sealed ESD bags and were held in styrofoam crating, with slots for each disk to keep disks apart.

The only explanation I've read from the mfgrs goes something like this: "We have to keep our parts at low prices consumers will pay. To do that, we can't afford to implement a level of production line QC that would result in ultra-low failure rates of products shipped. So we figure a 10% (or whatever) failure rate into the mix, and the people who get one of the DOA parts are basically screwed & have to hassle with an RMA. It's the luck of the draw, so good luck."

It's really a question of design and process control and process capability rather than QC per se. You can't afford to "test" quality into something like hard disks. If your standard processes have too many failures it'll cost you too much money to build and repair the defective ones.

Problems come up when your design and process capability can't build a reliable product. The problem with the highest density disks is that the data density is so high that the raw error rate that used to be acceptable now creates an unacceptable quantity of uncorrectable errors.

I have yet to have an HDD fail on me, in any machine I've owned.. I guess I'm just lucky.

Yup.

It's just a question of how many disks you run, for how many years, and the average failure rate for the disks. More like probability than luck, though.

 

[C1]

Man, Ive got tons of HDDs & havent had any failures. In the case of NewEgg ratings, if you look carefully you will see that dissatisfied customers submit multiple reviews/complaints for the same product purchase. No doubt in the hope of discrediting the manufacturer.

Again, looking closely, you'll see that a good number of people abuse or misuse the product (eg, let a 2TB HDD flop onto its side while its running), incorrectly connect it or connect it to substandard equipment (eg, buy a cheap enclosure multi-platter external notebook HDD & expect it to run using a cheap excessively long [6'] USB cable hooked to a substandard underpowered computer/port.

Finally, dont buy bottom end equipments. You will pay a good price, but the product will work well & last. I pay premium price for IOMEGA external notebook drives. They are $20 more than the lower end competitors. Read the fine print. The units come in black & silver. The black cost $10 more not because of the color, but because they have the extra specification that "THEY REQUIRE NO EXTERNAL POWER SUPPLY". So if you want to use a long cable this drive is more likely to work on a 6' high quality USB cable without issue (all of mine do).

http://go.iomega.com/en-us/products...ortable-series/prestige-compact/?partner=4760

 

[RebateMonger]

In the case of NewEgg ratings, if you look carefully you will see that dissatisfied customers submit multiple reviews/complaints for the same product purchase. No doubt in the hope of discrediting the manufacturer.

Go through all Seagate drives from 80 GB to 2 TB and count the number of one-star reviews for each. Notice that the older designs have ten percent or less "one-star" reviews and the newer (higher-density) disks have up to thirty percent "one-star".

If folks were trying to discredit Seagate, or if the cause was careless users, why wouldn't they give bad reviews to ALL Seagates?

It's the same with Western Digital drives. The older ones have ten percent or less "one-star", while the newer (higher density) ones have around twenty percent "one-star".

Sure, people will "over-report" early failures. That just means that the percent of early failures is less than twenty or thirty percent. It's lower than that. But early failures with high-density disks seem to be 2x or 3x as high as with older, lower-density designs.

I pay premium price for IOMEGA external notebook drives.

Ah. Iomega. The creator of the term "click of death":
http://en.wikipedia.org/wiki/Click_of_death

http://en.wikipedia.org/wiki/Iomega
"Iomega has been widely criticized for customer service issues dating back to 1998, when Wired covered the company's $100 million "advertising blitz" at a time of plummeting stocks and two class-action lawsuits brought to the company by frustrated customers. Warned that increased customer demand would overload Iomega's small in-house resources, the company quadrupled the number of customer support representatives in a thirty-day push to meet demand for its popular products. Fee-based technical support (which continues still today) for questions already found in product documentation was seen as the only feasible option in controlling spiralling support costs. Customers were not charged for support incidents in which Iomega was responsible (e.g., defective products). Nonetheless, some customers took exception.

In one lawsuit customers complained that Iomega's warranty promised free technical support, but once customers called, they were told they would be charged. In a second suit the company was charged with not delivering on cash rebates promised on Zip and Jaz drives sold under promotion.

Iomega's faulty zip drives became known for an anomaly known as "the click of death", where — after a time — zip drives would no longer be able to read disks and would instead produce a loud clicking sound. Many users reported damaged disks and data loss.

Despite the impact on the company's stock price, and negative media coverage about the company's legal, business, and customer service issues, its sales continued to grow."

 

[richierich1212]

Newegg's shipping on OEM drives kills the majority of them. Luckily my OEM drives came with enough peanuts to avoid arriving DOA.

 

[DrMrLordX]

Seagate seems to have had problems ever since they acquired Maxtor. Wonder why.

 

[Duwelon]

You can't take the number of 1 star reviews on face value anyway. I for one take a 1 star review with a giant dose of salt because I find the most ignorant and/or just plain tarded comments in a 1 star review on there. DOA should be it's own category instead of using that rating system.

I've seen Newegg's horrible packaging for OEM drives as well. I purchased 3 1TB enterprise WD RE3 hard drives about 4 months ago. 2 of them were wrapped in bubble wrap and the 3rd was just laying at the bottom of the box in the anti static bag, but it was clearly just flopping around in the box with very loose packing.

 

[fuzzymath10]

Other than a drive from 1996 which failed in 2002, the only drive that has failed on me was an infamous Seagate 7200.11 1.5TB which was DOA from Newegg. The RMA took literally a day; I sent the drive to seagate on a monday and got a drive back on wednesday. It has been working ever since, which is a year so far and I don't expect it to die anytime soon.

I'm highly suspicious of Newegg's packaging so anything with substantial moving parts I buy local. I have a mix of Seagates and WDs and some older Maxtors from 40GB-1.5TB and they're all fine. Even my 200MB Caviar and a 500MB Maxtor SCSI drive still boot.

 

[taltamir]

They don't make em like they used to. Everything is all made cheaply in china now, by kids pretty much working slave wages. It's sad that we've put ourselves in such an economic crisis that we can't even afford to manufacture our own stuff.

no, just no...

in fact, they make them better then ever before.
the problem is miniaturization. As you shrink mechanical components down as well as magnetic disks they get a lot more sensitive and prone to error and failure.
However, advances in technology have been constantly improving the reliability of drives to compensate for it. However, in current state its getting more difficult to compensate as we shrink further and further down resulting a net decrease in reliability in the most miniaturized of drives.
There is also an economic factor of very severe competition forcing them to push miniaturization to the max. If company X has cheaper, faster, larger drives that are 10% less reliable then company Y, people are going to buy company X. (frankly, as long as I back up my data and have good warranty, I would do just that).

Finally, poor packaging, storage, and handling by major e-tailers like newegg causes more drives to break.

 

[Elixer]

For what it is worth, these days you can't go with the old adage of you get what you pay for.
I have seen very expensive HDs fail, and some bargin bin priced units keep on running.

Same goes for SSD units, price alone isn't the only thing you should be concerned with.

For the egg, they still are using bubble wrap + a few foam peanuts in the majority of their warehouses. Some are actually packing things better, but that is the exception, not the norm. (pretty bad packaging depending on where it ships from)
Dell ships a box within a box with some foam peanuts. (They are the best packers)
ZZF ships with a HD foam "case" and foam peanuts. (2nd best)
TigerDirect. (STAY AWAY!--no peanuts or anything... they are nuts.)
Amazon (if Amazon ships it, then I guess it is good, but if not, then ick!)
Buy.com (meh. Paper wadding.)
Frys.com (1 report of tons of bubble wrap, + paper wadding)

Everything will fail sooner or later, there is no stopping that, which means, you need to have a good backup plan to mitigate the failure of the HD/SSD.

 

[Mr. Pedantic]

Some people are just lucky. Some just aren't. I've never had a drive fail on me, but I've heard of people who have had multiple drives fail within weeks of purchase.

 

[pjkenned]

To be honest, look at what major OEM's use or just look at marketshare since that's a decent proxy. Any drive manufacturer that ships too many defective drives will get hammered by their OEM supply agreements (I've read them at the top two drive manufacturers).

You will see more negative reviews of Seagate and Western Digital drives because they ship something like 30M drives each per quarter and each hold 25%+ market share, and have a huge installed base. Samsung, Hitachi, Fujitsu, and etc are smaller players at this point.

Remember that all drives fail, but 99% of home users (and a lot of enthusiasts) don't have proper backup/ redundancy. So when the inevitable happens, people lose their data (really their own fault) and get irate, turning to the internet to vent. If you ship tens of millions of drives per quarter, some people are going to be pissed off. Then again, there are people that run a single TB disk, or raid 0 expensive drives (like raptors/ wd blacks) that store a $2,000 music collection and all of their important documents and won't spend $70 to have a second 1TB disk for redundancy or backup.

One thing you notice is that most people who have lots of drives and therefore have redundancy, don't complain much about drive failures. WD and Seagate have easy RMA processes and hotswap bays are great!

 

[KeithP]

The bulk of early drive failures have nothing to do with packaging. Some drives leave the factory with a defect that causes a failure early in its life cycle. It is as simple as that. This is why whenever I get a new drive I do a multiple pass wipe to put some mileage on the drive before I start using for data.

The reviews at Newegg offer no useful statics in regard to drive failures. If you want something useful, look at the Google white paper on HD failures:
http://static.googleusercontent.com...abs.google.com/en/us/papers/disk_failures.pdf

The downside of that paper is it leaves manufacturer names out. However, if I recall correctly, Google's findings were they were more or less all the same.

-KeithP

 

[Ken90630]

WD and Seagate have easy RMA processes ....

Only problem is that when you send a defective drive to them, they send you a REFURBISHED one for the exchange. Even if it was brand new & DOA outta the box, you get a REFURB'd drive. That is utterly pathetic and unacceptable in my opinion. The drive mfgrs should have a policy that, say, if a drive fails in the first three years for any reason other than customer abuse/negligence, they will send you a NEW one as a replacement. Anything after 3 years and they should give you a choice between a refurb'd drive or a pro-rated discount on a new drive. THAT would be customer service and goodwill that would engender long-term loyalty and appreciation (and thus profits over the long run). If they can't produce drives at a high enough quality to make such an arrangement sustainable, profit-wise, then they need to improve their mfg quality to cut down on defective drives. Period.

I've RMA'd two HDs for friends in the last couple years, and both refurbs they got back failed within a month or two. So they had to go buy new, retail drives off the shelf 'cuz they had to get up & running & couldn't continue the RMA foolishness. Each time, it was $50-$60 down the drain for them.

I agree the legitimate failure rates are probably fairly low (excluding shipping damage or customer abuse), but this glaring fault in how failures are dealt with REALLY needs to be fixed by the mfgrs.

 

[tcsenter]

If you want something useful, look at the Google white paper on HD failures:
http://static.googleusercontent.com...abs.google.com/en/us/papers/disk_failures.pdf

And even the usefulness of it is questionable. The Google study certainly over-represents or exaggerates the actual rate of drive failure by its excessively permissive definition:

Definition of Failure. Narrowly defining what constitutes a failure is a difficult task in such a large operation. Manufacturers and end-users often see different statistics when computing failures since they use different definitions for it. While drive manufacturers often quote yearly failure rates below 2% [2], user studies have seen rates as high as 6% [9]. Elerath and Shah [7] report between 15-60% of drives considered to have failed at the user site are found to have no defect by the manufacturers upon returning the unit. Hughes et al. [11] observe between 20-30% “no problem found” cases after analyzing failed drives from their study of 3477 disks. From an end-user’s perspective, a defective drive is one that misbehaves in a serious or consistent enough manner in the user’s specific deployment scenario that it is no longer suitable for service. Since failures are sometimes the result of a combination of components (i.e., a particular drive with a particular controller or cable, etc), it is no surprise that a good number of drives that fail for a given user could be still considered operational in a different test harness. We have observed that phenomenon ourselves, including situations where a drive tester consistently “green lights” a unit that invariably fails in the field. Therefore, the most accurate definition we can present of a failure event for our study is: a drive is considered to have failed if it was replaced as part of a repairs procedure."

IOW, after [correctly] establishing that a relatively high percentage of drives which are deemed to have "failed" by the user (or repair technician) are subsequently found to be in perfect working order, but do not operate properly or "misbehave" in the user's particular configuration due to some interoperability issue between the firmware, BIOS, controller, driver, or due to bad cabling (or fluctuating/inadequate power source), the authors then use a definition that certainly includes those very false positives/failures.

Sift through the BIOS, firmware, and driver release notes for any enterprise class RAID or storage controller. Its not difficult to find change notes like "solved spurious fault/warnings/errors/behavior with [insert particular make and model hard disk here]" or "when in this specific configuration/under these particular conditions" (e.g. during staggered spin-up, when NCQ is enabled, so on and so forth, sometimes even referencing a particular firmware revision for a particular drive). From time to time, you'll even see a remark along these lines in a motherboard BIOS release note.

In a server environment like Google's, the potential for these kinds of interactions to be exposed is vastly greater than any personal computer. e.g. two to four storage controllers, 16 ~ 32+ hard drives (mixed makes/models) creates hundreds if not thousands of different configuration possibilities that even the most loaded personal computer/gaming/workstation will never encounter. The more complex the configuration, the more complex the interactions become between system BIOS, storage controller/firmware, disk drive firmware/controller, drivers, and even the OS.

The study made no attempt whatsoever to narrow the definition of "failure" to any kind of problem with the drive itself or to eliminate these interoperability issues owing to some BIOS, firmware, or controller glitch/bug. If the drive was replaced because someone (subjective) deemed it to be defective or problematic, it gets counted as a "failed" drive.

Overall, many miss the general conclusion or statement of the authors that disk drives are "generally very reliable" and "rarely fail."

 

[Modelworks]

Shipping is going to have little to do with hard drive failure. Drives park heads off the platter space when powered down so even a very hard bump isn't going to effect the platter. You can cause the heads to un-park if you hit a drive really hard but normal shipping methods shouldn't cause that.

I think a lot of current failures are user error. The closer you pack sectors on the drive the easier it becomes to damage the data. I have seen people doing things with their pc that can cause data damage especially with higher density drives like the 2TB models.


Moving a pc around while it is running. NEVER do this. I have seen more than one person slide around a pc to get to the cables in the back or tip one over to reach something while it was still running. The reply I usually get when I tell them not to is "It never caused problems before" .

People moving around cabling inside the case while it is running. All it takes is one knock of your hand on a hard drive while it is running to damage sectors.

Not securing a hard drive securely in the case. Some people like to just put a single screw in the drive and since it seems solid enough leave it at that. That causes vibrations between the mounting and the drive case. If you can put your hand on the drive when it is running and feel it vibrating that is bad for the drive. The heads are less than the thickness of a hair above the platter, vibration is bad.

Fans mounted in front of the hard drive. Good if done right, terrible if done wrong. This one goes back to vibration. If the fan you use is vibrating the casing you are better to not even use the fan and take your chances with the heat. Drives usually don't require a wind tunnel to keep cool.

Drives in external cases. Don't move them around when they are running. I have seen people moving them around like they are SSD drives. Unplug them and wait about 30 seconds before you move one to give time for the head to park and the platter to stop.

Following this I have only had 2 drives fail in the past 15 years. One was a conner 850MB and the second was a western digital 80GB. Drives are built pretty good internally and will last a long time if cared for. Just last week a drive I use for videos, Seagate 250GB barracuda about 3 years old, in an external enclosure ,was knocked off a desk and fell on a tile floor while it was copying data. I thought it was broken for sure . I looked at the desktop and it had frozen in the copy process. I quickly unplugged it and after making sure it was still intact, plugged it back up and ran a check on all the data and sectors. All 100% good which really surprised me, so drives are made pretty good, just don't press your luck

 

[DonInKansas]

People that have something go wrong are more likely to voice their opinion than those that are happy with their product and it works properly. That's why you always have to take customer reviews with a grain of salt.

 

[pjkenned]

Only problem is that when you send a defective drive to them, they send you a REFURBISHED one for the exchange.

A few months ago I had 3/8 WD Green's fail (within the first two weeks). The first one was in the advance RMA queue for 2 weeks because they had no inventory. One call and I managed to get new drives shipped not refurbs.

The OCZ Vertex I just got back I'm not 100% sure if it is a refurb. Came in full retail packaging (sealed, but then again I'm assuming OCZ could just make a refurb look new). Usually drive manufacturers will but something like "refurbished" on the drive labels to denote that. I'm assuming it is a refurb, but they were easy to deal with and I have great backup.

To ModelWork's point about moving a case while a drive is in operation, that is a really good reason I moved to SSD's on everything but storage servers. I think they are generally quite a bit more durable sans moving parts.