The hard drive failure thread

[eccerr0r]

Re: The hard drive failure thread

Wow that was an old post of mine, must have been my 120G disks for my old RAID... I've since decommissioned that RAID but the drives were still working just fine when I took it apart.

Oh well, I guess it won't hit 100K hours unless I decide to use the disk somewhere else... And I wonder if anyone would take their chances with a 70K POH drive on eBay

Toshiba MK3275GSX, failed from bad sectors. Made on Feb 29th, 2012. Surprised the thing even let me load Windows, with 11,000 reallocated sectors.



Anyone ever had good luck to Toshiba? I've come around to using the name Toshita, I've never had a Toshiba drive last more than 2 years of use, even if it's a stationary laptop.

I've actually had decent luck with Toshiba disks. The old ones are slower than the "usual brands" but haven't had one fail as I simply haven't had many of them. The other brands I've had failures on pretty much all of them as I've went through more of them.

That being said I wonder if the raw number is actually the wrong endian, wrong bit order, or even if there are sub bit fields in that number. You can't necessarily trust the number in "raw value" because it is manufacturer (and model) specific. The only numbers that you can trust is making sure the VALUE number is greater than THRESH; the only other indication is if VALUE has decreased since it was new. Note that this number for VALUE can go up and down so it's really hard to tell what's the disk lying about... Then again there are a lot of pending reallocates...

[fzabkar]

Re: The hard drive failure thread

That being said I wonder if the raw number is actually the wrong endian, wrong bit order, or even if there are sub bit fields in that number. You can't necessarily trust the number in "raw value" because it is manufacturer (and model) specific. The only numbers that you can trust is making sure the VALUE number is greater than THRESH; the only other indication is if VALUE has decreased since it was new. Note that this number for VALUE can go up and down so it's really hard to tell what's the disk lying about... Then again there are a lot of pending reallocates...

I've seen a LOT of SMART bugs. ISTM that the Toshiba drive's raw Reallocated Sector Count is correct, but the normalised values are telling lies. That said, some Advanced Format drives report the Reallocated LBA Count rather than the Reallocated Sector Count, ie a logical sector count instead of a physical sector count. The actual number of reallocated physical sectors is obtained by dividing the logical sectors by 8. The numbers 11056 and 2008 are both divisible by 8 ...

[ChaosLegionnaire]

Re: The hard drive failure thread

last week one of my beloved and no longer being produced hitachi p7k500 500gb ide drives died... literally.

was copying some files over and went afk for an hour. came back and tried to copy more files but the drive started clicking and froze the system. hit the hard reset button. drive could no longer be detected by bios. tried power cycling the drive. bad idea and even worse epic phail! drive wouldnt spin up and still cannot be detected.

i thought it might be a stiction problem as its used infrequently for backup. i tried taking it out and spinning it left and right horizontally while holding it in one hand. no dice. still wont spin up. gonna have to try the exotic freezer trick some time later when i have the time... oh my data... 500gb!

this is the first time ever i got a completely dead undetectable drive. previous failing drives i've had would still be detectable once i leave them running enuff and they give up and come out of error recovery mode and come back online again thus i could proceed with data recovery. not this one... i hope i didnt kill the drive dead when i power cycled it while it was clicking...

also got another two hitachi 7k3000 hds723020bla642 2tb failing. one has 105 reallocated sectors, 1 pending sector and 1 offline uncorrectable. the other has smart failure due to seek error rate failure. that seek errored drive is purchased second hand. i think it must have had a drop in the past based on previous experience in this forum with another guy who dropped his seagate drive.

damn i thought hitachis were the most reliable... i've hoarded like 6 of those drives due to their legacy format 512b sectors as im sticking to winxp. sad to see 2 or one third of them fail after 3 years. now i'm forced to buy into this advanced format bs since none of the drives being sold in my country now are legacy format.

some Advanced Format drives report the Reallocated LBA Count rather than the Reallocated Sector Count, ie a logical sector count instead of a physical sector count. The actual number of reallocated physical sectors is obtained by dividing the logical sectors by 8. The numbers 11056 and 2008 are both divisible by 8 ...

thats why i hate that advanced format bs. if u have a bad sector on an AF drive, i can twist my words and say that advanced format drives cause 8 times as much data loss when a bad sector occurs compared with legacy format.

[lti]

Re: The hard drive failure thread

Anyone ever had good luck to Toshiba? I've come around to using the name Toshita, I've never had a Toshiba drive last more than 2 years of use, even if it's a stationary laptop.

Toshiba drives suck. I'm surprised that my laptop's Toshiba drive hasn't failed yet. It's three years old.

[momaka]

Re: The hard drive failure thread

Anyone ever had good luck to Toshiba? I've come around to using the name Toshita, I've never had a Toshiba drive last more than 2 years of use, even if it's a stationary laptop.

Me neither.
When I used to work in a electronics repair shop, laptops and PS3s with failed Toshita HDDs were common sight. I think I still have a dozen of them in a box somewhere in my closet. Mostly 120 GB to 1 TB. I don't think any of them had more than 5000 power-on hours.

I know Backblaze rate the Toshiba above other drives.

Where did you get that from their article? If anything, HGST (Hitachi) is the one getting praised.

I've also had nothing but good luck with Hitachi myself. Heck, even their old IBM "Deathstars" (prior to the Hitachi-IBM merger) were more reliable than the aforementioned Toshita HDDs. I can probably pinpoint only two specific models/lines that are problematic: the DTLA-3070xx line (IBM) and DK23AA-12B model (Hitachi). That last one should not be mistaken with the DK23EA-20 model, which is actually reliable too. Go figure.

And I wonder if anyone would take their chances with a 70K POH drive on eBay

Maybe not from eBay
I did buy an 120 GB IBM ICA-something model (I think) from MicroCenter. Turns out, it had nearly 100k power-ON hours (995xx to be exact), but ZERO bad and reallocated sectors, and only like 20 or 30 power cycle counts. I checked the time counter to make sure that it counts in hours and not minutes (beware that some older 2.5" Hitachi HDDs did do that) - but no, it does count in hours. I thought I saved the SMART log somewhere on my main PC here, but I guess I didn't. Anyways, the computer in which I put that IBM HDD in is used maybe once a year or so for a few weeks. So it will probably be a few years before it gets to 100k hours. Still, I call that reliable!

[ChaosLegionnaire]

Re: The hard drive failure thread

Toshiba drives suck. I'm surprised that my laptop's Toshiba drive hasn't failed yet. It's three years old.

same. my mom has a three to four year old sony vaio laptop with a toshiba 640gb drive in it. surprised its still running with no reallocated sectors last i checked several months ago.

I did buy an 120 GB IBM ICA-something model (I think) from MicroCenter. Turns out, it had nearly 100k power-ON hours (995xx to be exact), but ZERO bad and reallocated sectors, and only like 20 or 30 power cycle counts. I checked the time counter to make sure that it counts in hours and not minutes (beware that some older 2.5" Hitachi HDDs did do that) - but no, it does count in hours. I thought I saved the SMART log somewhere on my main PC here, but I guess I didn't. Anyways, the computer in which I put that IBM HDD in is used maybe once a year or so for a few weeks. So it will probably be a few years before it gets to 100k hours. Still, I call that reliable!

thats really not very nice esp from an ebay seller! selling off a drive that is about to hit the 100k hour mtbf! u'd think they'd be transparent and post smart screenshots of the drive to inform potential buyers...

[fzabkar]

Re: The hard drive failure thread

Toshiba acquired some of Hitachi's storage technology. In particular, some of Toshiba's 3.5" models are really Hitachis.

[eccerr0r]

Re: The hard drive failure thread

thats really not very nice esp from an ebay seller! selling off a drive that is about to hit the 100k hour mtbf! u'd think they'd be transparent and post smart screenshots of the drive to inform potential buyers...

But if it clearly says "USED" is it really that incorrect of an advertisement? Just so happens to be well-used... Then again 100K hours POH != MTBF... I think most of these drives are in the 300K-500K hour MTBF, but all that means that it's nearing failure time (But anyway, this is worse, Microcenter selling a clearly marked "used" 100K hour drive versus buying a 100K hour drive on eBay?)

I guess my drives are similar, I have two that are around ~70K POH. One has 300 and other one probably has 1100 power cycles on them, both have zero reallocated and zero bad sectors. I have two others from that RAID5 that are a bit younger as I grew my array, but one of them has two reallocated sectors on it.

Should I "trust" this brand of hard drive with 70K POH? Even though I had to RMA one of them (massive bad sectors)? Hard to say. (Brand = Maxtor, bought out by Seagate.)
My current array has HGST drives in it, a 500G. I even had to RMA one of these too (it stopped detecting too!!!) Luckily it was a RAID5 member and no data loss.

I'm still trying to test out the 2T disks that will replace the 500G/disk array. We'll see if these 500G disks hit the 70K hour mark before I swap them out, they still have thousands of hours to go...

[fzabkar]

Re: The hard drive failure thread

But if it clearly says "USED" is it really that incorrect of an advertisement? Just so happens to be well-used... Then again 100K hours POH != MTBF... I think most of these drives are in the 300K-500K hour MTBF, but all that means that it's nearing failure time

MTBF is not what you think it is. It is not telling you how long the manufacturer reckons an individual drive will last on average. Rather, it is telling you the average time between failures in a large population of drives.

For example, let's say the MTBF is 500K hours and that you have a datacentre with 1000 drives. In this case you can expect that there will be, on average, one failure every 500 hours. If you had 10,000 drives, then the failure interval would be 50 hours. Note that this MTBF spec applies only to new drives. This means that the MTBF may fall to 250K hours by the second year, and 50K hours by the 5th year.

BTW I see nothing wrong in selling a drive with 70K hours provided that the buyer is made aware of this. Just attach the SMART report. In any case some buyers might want the drive just for the PCB.

Edit:

Notice that the raw value of Power-On Hours for the Toshiba HDD is 3085, and that the normalised value has dropped from 100 to 93. This means that the drive has lost 7 points in racking up those hours. It may have just recently lost the 7th point, or it may be just about to lose the 8th point.

This means that we can calculate how many hours the drive will have logged by the time it loses all 100 points.

Expected Lifetime is less than or equal to (3085 hours / 7) x 100 = 5.03 years

... and greater than or equal to (3085 hours / 8) x 100 = 4.4 years

[momaka]

Re: The hard drive failure thread

thats really not very nice esp from an ebay seller! selling off a drive that is about to hit the 100k hour mtbf! u'd think they'd be transparent and post smart screenshots of the drive to inform potential buyers...

That wasn't from an eBay seller. I got the HDD from MicroCenter. Either way, it was cheap and it works so I can't complain (I paid $5 or $6 for it).

Toshiba acquired some of Hitachi's storage technology. In particular, some of Toshiba's 3.5" models are really Hitachis.

Good to know.
Probably why their 3.5" models may have less problems then. I still wouldn't trust them, though. If you get a drive that is made by Toshiba and not Hitachi... good luck with your data!

[diif]

Re: The hard drive failure thread

You're right momaka, i got confused between the two. i'm blaming agonising shoulder pain for my oversight.

[mariushm]

Re: The hard drive failure thread

I have four drives in my system right now. I had more drives but I gave them away (a 250 GB Seagate and a 400 GB WD Black) while they were still working fine.

The Seagate is the only one I had back luck with, quite a few of bad sectors but luckily they're isolated to a region of about 20-30 GB, so I just created a partition for that area and left it unformatted.

The WD 1TB Green just recently developed a few sectors according to SMART data, but it works perfectly fine. With the Seagate, the drive was noticeably slower when writing data to that region with bad sectors.

I just love my Hitachi NAS 4TB drive, behaves and works wonderfully so far. I think it's worth paying extra over drives like WD Green but while I wouldn't mind still buying WD Green, I'm staying away from Seagate or WD purple or other non-NAS drives. (even though i don't use them in raid or a nas box).

I keep my computer running 24/7, basically restarting it only when antivirus pushes me to reboot or I accidentally install some Windows updates.









For a drive that ran 42980 hours (1790 days or 4y 330 days), i really don't mind that i may soon have to retire it (or change its role to storing music or tv shows i don't care if they get corrupted or lost)

The Seagate doesn't have the same excuse, yeah it's close to 40k hours but it developed bad sectors months after warranty was over (2 years), not just now like the WD.

[eccerr0r]

Re: The hard drive failure thread

MTBF is not what you think it is. It is not telling you how long the manufacturer reckons an individual drive will last on average. Rather, it is telling you the average time between failures in a large population of drives.

So much for pedantic replies, but the claim is still true for this single 10 year old drive - it is still a craps shoot and could fail any time as it already has likely gone through at least 7 of its 9 lives. If the MTBF rating was 2M hours, then statistical failure is farther away. It still could die tomorrow, after the buyer hooks up +12V to +5V.

BTW I see nothing wrong in selling a drive with 70K hours provided that the buyer is made aware of this. Just attach the SMART report. In any case some buyers might want the drive just for the PCB.

Why do eBayers need to advertise this when Microcenter does not? I'd think that if asked, this data should be supplied (barring "but I don't have the tool, I don't run Linux, I don't trust freeware SMART tools...") - but if not asked, there's no reason to post it, just like Microcenter.

I also got a used laptop disk from Microcenter with a lot of miles on it. This time it's not POH but rather hundreds of thousands of load/unload cycles. The drive apparently was in a machine that was turned on and off frequently and the SMART rating for this has dropped considerably (I think it's in the 30s where threshold is 0). Likewise it is still working, but who knows when it will fail.

[momaka]

Re: The hard drive failure thread

I also got a used laptop disk from Microcenter with a lot of miles on it. This time it's not POH but rather hundreds of thousands of load/unload cycles. The drive apparently was in a machine that was turned on and off frequently and the SMART rating for this has dropped considerably (I think it's in the 30s where threshold is 0). Likewise it is still working, but who knows when it will fail.

What model drive was this? I know the WD Greens will rack up an amazing amount of those. My old (but trusty?) Hitachi DK23EA-20 also does - you can hear it click like every 5 to 10 seconds if the HDD is idle.

Load/unload cycles is not the same as power cycle counts, though.
Power cycle counts is for how many times the power was cycled ON and OFF.
Load/unload cycle count is *usually* for the number of times the heads have parked (which for some drives is the same as the PCC above, if the HDD parks its heads only when OFF). High load/unload cycle count is (again, usually) only bad for drives with head ramps.

[Agent24]

Re: The hard drive failure thread

You can turn the WD Green parking feature off with WDIDLE utility in DOS

[goontron]

Re: The hard drive failure thread

You can turn the WD Green parking feature off with WDIDLE utility in DOS

or HDPARM

[eccerr0r]

Re: The hard drive failure thread

I don't remember what model/manufacturer it was, other than it was a laptop disk. It may have been Toshiba actually (but it could have been a Fujitsu? Not sure. Pretty sure it was not WD, Seagate, or HGST/IBM). The load/unload cycles stopped when I was using the disk so it was purely an OS thing or how it was used in its previous life, and not a function of the disk itself.

I also don't recall if it was load/unload or actual power cycles or something else. It may have been the latter, something was actually stopping the disk from spinning during power down. I suppose when the disks are spun down, the heads might well be unloaded...

[Compgeke]

Re: The hard drive failure thread

You can turn the WD Green parking feature off with WDIDLE utility in DOS

or HDPARM

From my experience with greens you'll want to do do this. Experience has been unless this is disabled they tend to die after a year or so but once it's disabled they last alright.

[eccerr0r]

Re: The hard drive failure thread

Yay:

Model Family: Western Digital Green
Device Model: WDC WD20EARX-00PASB0
5 Reallocated_Sector_Ct 0x0033 200 200 140 Pre-fail Always - 0
9 Power_On_Hours 0x0032 060 060 000 Old_age Always - 29268
193 Load_Cycle_Count 0x0032 001 001 000 Old_age Always - 613370

One more will break the camel's back? Or a hundred more? Or tens of thousands more? ... (I used wdidle3 on this disk long ago, but not soon enough to not let it rack up 600K cycles. Not sure how "accurate" the implementation of wdidle3 is in hdparm yet...trying reverse engineered software on something that holds 2TB of data is risky!)

Oh, and it's the "Start_Stop_Count" on that laptop disk that was very high. The load cycle count was not as high as this disk above... but it was a HGST after all...

Model Family: Hitachi Travelstar 4K120
Device Model: HTS421280H9AT00
4 Start_Stop_Count 0x0012 039 039 000 Old_age Always - 97602
5 Reallocated_Sector_Ct 0x0033 100 100 005 Pre-fail Always - 0
12 Power_Cycle_Count 0x0032 039 039 000 Old_age Always - 96922
193 Load_Cycle_Count 0x0012 073 073 000 Old_age Always - 275886
196 Reallocated_Event_Count 0x0032 100 100 000 Old_age Always - 0

Still good...so far...

[Spork Schivago]

Re: The hard drive failure thread

I believe my wife's hard drive in her laptop is dying. But I want to know if anyone thinks it's fixable. It's a Seagate 7200.4 500GB ST9500420AS. Read rates are very low, maybe 400KB/s. SMART is turned on in the BIOS but doesn't report any failures. I know I should probably run a tool to read the SMART data. Anyone have any suggestions on a good free Linux program that can read SMART data? It's an Asus laptop that cost maybe 800$ a few years back. Anyway, the original hard drive died, this is the second one.

I was wondering if the vent hole could be clogged and if so, could it explain issues like this? If it's fixable and permanent damage hasn't occurred yet, we'd like to fix it to get the data off. I don't know if this relevant or not, but the night we noticed it failing, it was making a weird beeping type of noise. Never heard a computer make a noise like this before. And it wasn't coming from the laptop speakers. It was coming from where the hard drive was so we assumed it was the hard drive. It wasn't loud, more like a dying beep, not a squeak. Any ideas?

Also, in 2005, I discovered Seagate hard drives aren't water proof, however, most of the Western Digital drives that where under water survived. Once the water level subsided from the flood, we would take the drives out of the PCs, dry them off, shake them to see if they had water in them, if they did, we'd trash them, if not, we'd hook them up to a USB enclosure and try to copy the data. Almost all the WD drives worked long enough for us to get the data off. The one Seagate that didn't have water on it....we plugged it in and the thing started smoking (which was really cool). Just thought you'd like to know!