Increasing rate of hard drive failures

[corkyg]

When a drive has to work harder, it can be more subject to failure. Failure rate equates to reliability. Further, if a drive does not do what you want it to do in a timely fashion, it is not something you can rely on.

 

[DrMrLordX]

Microsoft claims to have incorporated disk error correction into its in-development "Windows Home Server V2" that's currently out in Beta form. I haven't heard of any conflict with NetApp's patents.

My guess is that Microsoft's lawyers will do a better job than Sun's of guiding product development to end-run NetApp's patents where necessary (10 percent+ differentiation, etc). I sometimes get the feeling that people who develop large open-source products grow a hole in their head when it comes to issues of intellectual property.

Don't get me wrong, I think ZFS is awesome . . . I'd just hate to see it die because NetApp killed it.

 

[Voo]

When a drive has to work harder, it can be more subject to failure. Failure rate equates to reliability. Further, if a drive does not do what you want it to do in a timely fashion, it is not something you can rely on.

Well first of all "full" doesn't mean the drive has to work harder, something which every backup disk that's only used once a week can easily prove

So while useage may make it more probable that a drive fails, that has nothing to do with how full the drive is.

Also the performance is hardware dependent (i.e. density, rpm,..) and from the position where you read/write - but again, that has nothing to do with how full the drive is.. wouldn't make any sense if it did. Reading from sector X should always take the same time, independent if sector X+100 has meaningful data or only trash on it (as long as you start from the same position and ignore the cache)
If you don't defrag your drive (on a FS that doesn't do that automatically), you can also easily end up in a fragmented state, even if the drive is half empty (you only need a bit spare space so you can fix that)


@DrMrLordX: Yeah ZFS would be great without the whole patent and license issues..

 

[corkyg]

OK - that's cool!

 

[Seven]

I haven't had a hd failure in about 5 years. One big problem with drives currently is people moving them around when they are running. They see the portable label on external drives and think that means they can pick it up, toss it on a shelf or slide it around the desk while it is running.

The other thing is people moving their pc to get to cables, or something they drop behind it. That would be fine if the pc is off but doing it while the pc is running is a bad idea.

Same goes for laptops, they were not intended to toss around like a book while running.

Indeed. When I visit my cousin I always see his external maxtor laying on the floor and the kids knock it off from time to time. The damn thing is still working though..

 

[Chapbass]

Okay so heres a slightly different take on it that I just thought of.

Could it possibly have to do with Microsofts alternate way of doing drive caching (superfetch) on Vista and 7? Now, I don't feel like this alone would be enough to impact anything, but the drives are more sensitive when moving all over the place, and something like caching would do that. Combine that with people moving computers around when theyre booting up, and if you include the caching time as "booting up"...that could potentially make for more overall damage to drives.


Obviously that only applies to laptops, but just a thought.

 

[Emulex]

there is system write back cache os
and drive write back cache

disable or enable either.

 

[somethingsketchy]

Every external drive I've bought in the last three years has failed (WD and Seagate), only one out of five internals has died over the last five years. I just find external HDDs to be rubbish tbh.

+1

I personally don't use external drives for this reason. More often than not, the "airflow" designs aren't the greatest.

Which is probably why the OP is seeing so many dead hdds lately...

 

[thescreensavers]

^ Yea

My External HD, its a 2.5 inch HD in there, and it started to click bad, got my data off of it and got a new drive.

 

[sdifox]

It started when the Perpendicular Recording method was adopted.

 

[Blain]

It started when the Perpendicular Recording method was adopted.

Very true...
That's when our capacities grew and our failures increased.

 

[Emulex]

i disagree. the last generation of non-perp was pushing the areal density too high and too many platters.

but you guys use enterprise storage? out of 1000 hard drives, 1 failure per month at 100% duty cycle is about what i see. That is not bad at all. There is no SATA in enterprise

i've got 4gb and 9gb quantum scsi servers - 1999'ish maybe 2000 - rocking out on original drives. caked with so much dust i dont dare sneeze near the machines. that's 10 years of service. the last 5 being qmail/clamav/spamass and other tasks - very disk intensive when you are fending off 100K email/spam attacks a day. alot of swapping (1gb of ram for all of that no less).

Am i impressed? no. that is what i call a normal hp proliant.

 

[Idontcare]

Very true...
That's when our capacities grew and our failures increased.

Is perpendicular recording really to blame here though? I don't see the cause-and-effect connection. I see a coincidence in timing though. Hard-drive prices collapsed, ASP tiers collapsed dramatically.

You no longer had $800 drives, $600 drives, $400 drives...if you were a drive maker you had no choice but to find a way to make a hard-drive that could be sold for at most $300 retail.

So perpendicular recording may have heralded an era of such excess capacity that mainstream users no longer felt capacity constrained and thus no longer felt compelled to buy an $800 hard-drive anymore, but the reason we see drives failing ever more commonly has probably got more to do with drive makers cutting costs/quality and cutting costs/quality and cutting costs/quality so they can sell you that 7200rpm 2TB drive for $110.

They could prolly slap a vintage 2002 platter in one of these 7200rpm drives today and it would still fail more quickly than a drive from 2002 because of the lower quality components that go into everything that is not the platter.

 

[ShawnD1]

Every external drive I've bought in the last three years has failed (WD and Seagate), only one out of five internals has died over the last five years. I just find external HDDs to be rubbish tbh.

It's the controller that sucks. I've had lots of external drives fail, but the actual hard drive itself is usually fine. 2 of my "dead" external hard drives are still being used as internal drives in my file server. They work great.

I haven't had any problems with drives aside from the ones I drop on hard floors. HD Tune reading the SMART profile says my 120gb backup hard drive has just under 50,000 hours of power on. That's 5.7 years. I still check the drive with Bart's Stuff Test and it passes every time.

 

[RebateMonger]

It started when the Perpendicular Recording method was adopted.

I haven't seen any evidence for that (based only on my own experience). I don't know when others started using perpendicular recording, but I believe the first "mass use" was the Seagate 7200.10 drives. I use maybe twenty of these and, so far, I've only had one drive that even developed a single "reallocated sector".

My 7200.10 drives are all 320 GB and were purchased in the 2006 time frame. They are used for nightly backups and are swapped out and carried offsite on a three-week cycle. So they are on around a 33% duty cycle, but get very hard use while they are in their servers (six-hour backups each night). Plus they have to survive transport every three weeks in a car between the server owners' homes and businesses.

On the other hand, I've had about half of my Seagate 7200.7 disks (mostly 200 GB) develop bad sectors. These were used in the same manner as the 7200.10 disks. Admittedly, these disks are a couple years older, too. But I'd be hard-pressed to say that perpendicular recording was worse.

All-in-all, I'm pretty pleased with the Seagate 7200.10 320 GB perpendicular recording disks so far (and, no, I'm not a big Seagate fan).

 

[sdifox]

I haven't seen any evidence for that (based only on my own experience). I don't know when others started using perpendicular recording, but I believe the first "mass use" was the Seagate 7200.10 drives. I use maybe twenty of these and, so far, I've only had one drive that even developed a single "reallocated sector".

My 7200.10 drives are all 320 GB and were purchased in the 2006 time frame. They are used for nightly backups and are swapped out and carried offsite on a three-week cycle. So they are on around a 33% duty cycle, but get very hard use while they are in their servers (six-hour backups each night). Plus they have to survive transport every three weeks in a car between the server owners' homes and businesses.

On the other hand, I've had about half of my Seagate 7200.7 disks (mostly 200 GB) develop bad sectors. These were used in the same manner as the 7200.10 disks. Admittedly, these disks are a couple years older, too. But I'd be hard-pressed to say that perpendicular recording was worse.

All-in-all, I'm pretty pleased with the Seagate 7200.10 320 GB perpendicular recording disks so far (and, no, I'm not a big Seagate fan).

For me the 320s were the beginning of the problems with seagates. I have had at least 5 distinct Seagate drives from 320 up to 1TB forget its partition table and do that stupid seek all over the drive thing. Same with a 1TB WD green drive.

Meanwhile the 40GB WD on my Athlon 700 is still going strong. Original drive :awe:

 

[Juddog]

I've seen a lot of bad laptop drives lately, but then I realized that it was typically the people that go out in the field. When I asked them about how they actually used the laptops, I realized that many would simply close the lid and toss it in the bag and go travel with it, instead of shutting off first.

One of the reasons I have been liking the Lenovo series is their hard-drive locking software; basically if you move the laptop, the hard drive locks. Other laptop manufacturers have it, but the Lenovo software seems to be the most effective for what it does.

 

[Bill Brasky]

One of the reasons I have been liking the Lenovo series is their hard-drive locking software; basically if you move the laptop, the hard drive locks. Other laptop manufacturers have it, but the Lenovo software seems to be the most effective for what it does.

I thought it was pretty slick that Lenovo detects different kinds of movement as well. Obviously there is vibration, but it also detects the orientation of the laptop and locks the drive if it's too far from horizontal.

 

[Russwinters]

Manufacturers are pushing the limits of the technology now, and they are trying everything they can to keep increasing the storage capacity and speeds, but unless there is a significant breakthrough we are running close to the limits of hard drive technology after 4K sector sizes are adopted.

Heat assisted recording is the next mechanical upgrade to come, that will be interesting.

 

[Juddog]

I thought it was pretty slick that Lenovo detects different kinds of movement as well. Obviously there is vibration, but it also detects the orientation of the laptop and locks the drive if it's too far from horizontal.

Yup - it detects the pitch and angle of the laptop, and you can set the rate of parking based upon a number of preferences, so you can be more or less aggressive with the parking if you want. It's pretty sweet in terms of preventing failures, and the amount of failures I have seen in my company have dropped quite a bit because of it.

 

[YBS1]

It would be hard to tell without having access to the failure rate data the manufacturers have. I'm hoping I don't hex myself here but I've personally not had a hdd failure since the IBM Deathstar era, in which 3 of the 4 Deathstars I owed died. So in my eyes at least, they are as good or better than ever. However, I do own a 750GB Seagate that has always made "click of death" noises since it was brand new. I was certain it was going to die an early death but it's still going strong two years later.