failing hard drives after sitting for years...

[eccerr0r]

With the discussion about water pumps failing when cars sit for a while, I wonder what people have seen with hard drives. Do people see hard drives failing after being sat on the shelf for a long time unpowered versus a hard drive that's at least powered up at least once in a while, say, weekly?

Wonder if it's worth spinning up the spare spinners once in a while to keep the bearings lubricated?

Wonder how I should best keep my cold spares for my RAID "fresh" for eventual failures of the working drives. Was thinking about keeping a hot spare but that's another question whether it will have a higher failure rate than mostly keeping it unpowered.

[Topcat]

Re: failing hard drives after sitting for years...

I think a lot of that depends on how they're stored. I keep all my spare drives (atleast the rarer ones) in antistatic bags in cool - low humidity places, and they come out working fine when I need them. If it's sitting in a garage or a storage shed, I wouldn't give odds on them working if stored there any length of time.

[eccerr0r]

Re: failing hard drives after sitting for years...

I had one computer in my garage for a few years unpowered, and the drive turned out fine...120GB. It was an old drive as well - it was taken out of service and then installed in that garage computer.

Alas this is only one sample, need more sample data...

[stj]

Re: failing hard drives after sitting for years...

it's the data,
the fucking drive makers store the cal data on the drive surface instead of an eeprom these days
i think weekly is overkill, i have had drives not used for many years that work, booting and doing a surface-scan yearly should be more than enough to get the drive to do a read-erase-write to any weak areas.

the exception is seagate.
those are total shit if they are sata.
the company went downhill around that time, the scsi drives have firmware bugs that admitadly have updates to fix,
the sata drives have more serious issues - like controllers slowly burning up on the pcb.

[eccerr0r]

Re: failing hard drives after sitting for years...

I think some cal data should be stored on the disk because it's specific to the disk -- and the bad sector map probably should also be stored on the disk too.

I forgot if that 120G disk was a seagate or a maxtor, was one of them... old disk. PATA.

Yeah seems a lot of disks can just sit there without damage, but I wonder if it's beneficial to spin them up once in a while or not. Sometimes I wonder if a surface scan is sufficient or perhaps you should get two disks (and a third one to hold onto data temporarily) and swap/rewrite the whole disk instead of just reading the data...

Oh and my garage is not climate controlled, usually around 20°F to 115°F, which does mirror outside temperatures from -5°F to 100°F.

[Topcat]

Re: failing hard drives after sitting for years...

I've had people drop systems off here that have been in basements, garages, and sheds for years. Most all of which I was told 'worked' when they were stored, they just upgraded. One of the common denominators was failed HDD's. This area is miserably humid in the summer; which in my opinion (based on observation) is one of the factors as to why...leads to condensation....that can't be kind to platters & data.

[eccerr0r]

Re: failing hard drives after sitting for years...

I hope to also discount hard drives moved around in a pile. That I've found out is detrimental to life of disks. At minimum they need to be separated by something, antistatic bag or whatnot, but if they're in a pile, aluminum to aluminum, they die...

[Kougar]

Re: failing hard drives after sitting for years...

Wonder if it's worth spinning up the spare spinners once in a while to keep the bearings lubricated?

I'd say no. Or at least, not nearly so frequently. Maybe every couple years? It's actually possible for drives to get over-lubricated too, if you remember the IBM Deathstars...

Wear on the motors/bearings is the worst when a drive spins up, and so if it does it constantly due to an overaggressive power profile or firmware bug then it can wear out the drive. If the motor can't maintain a specific platter RPM or RPM range then the drive will never be able to read the data and it's equally as dead as if it had an actual head crash.

It's been my experience I get a much longer lifespan out of drives by letting them spin 24/7 than to use power save features. It takes hardly any work to maintain the spin once it's going, so at that point controlling heat will help prolong the bearings themselves.

There's been Windows versions where accessing Explorer caused sleeping drives to spin up even though the drive directory itself was never accessed. And Western Digital had that firmware fiasco with Green HDDs some years ago where the spin down & head parking was erroneously configured too aggressively which caused the drives to wear out within months.

I've never had an issue with old HDDs dying on the shelf myself, but it probably depends on the model. Some HDDs use a glass platter with a magnetic coating applied and I've heard of those coatings failing due to some combination of heat or humidity. I wouldn't know if modern drives still use them or not, google gave mixed results. But once that coating begins to delaminate spinning up the platters will have only one inevitable result.

[redwire]

Re: failing hard drives after sitting for years...

I powered up a like new hard drive, sitting on the shelf for years and... click click click. It's dead. It was used only to do a few backups.
Seagate ST3500320AS Barracuda 7200.11 500GB

I thought I could try a firmware update on it but it's a hassle so I put it aside. If it can't find its servo track, no fixing that I guess.

Common failures I know are the flex-pcb connections to the head preamps/motor oxidizing and making poor contact. You have take out the controller board and clean and DeOxit.

[Dan81]

Re: failing hard drives after sitting for years...

I think I might have a bunch of drives that show horribly bad health yet run forever. Recent one is a ST1000DM003 that I suspect of racking up bad sectors at FW level and not physically - getting it to give terminal access is a complete BDSM torture. No, there isn't any sensitive info on it. (it came out off a upgraded iMac very likely - there wasn't any partition on it that I could see, nor was the drive initialized in any way that Windows could access.)

It did sit a long time unused after being taken out of service (around 4 years or so at a minimum.)

[ChaosLegionnaire]

Re: failing hard drives after sitting for years...

in 2019, i bought a brand new hitachi enterprise grade hard drive, still sealed in its factory anti-static bag and manufactured in 2014. it was shipped with so-so padding. when it got here, i tried powering it on but all that happened was click-click-click. mobo wont recognise or detect the drive.

not sure if the crappy padding killed the drive and it didnt survive the international shipping or the drive was doa due to not being powered on for a long time. needless to say, i was not amused that a hitachi enterprise grade hard drive turned out like this! when hitachi is supposed to be the most reliable hard drive brand according to hard drive stats by backblaze and google and it was an enterprise drive to boot! not a consumer grade desktop hard drive!

i also have like several hitachi 2tb drives from 2009 which i used for archival of my data which i havent powered on in a long time. dont know if they still work. i might be afraid to learn the truth if they dont power on or work anymore. so i might just have to bury my head in the sand like an ostrich on that one... lol! because i dont think i can handle the truth like what gene hackman said to tom cruise in the movie "a few good men".

[eccerr0r]

Re: failing hard drives after sitting for years...

I'm still not convinced enterprise grade drives are really more reliable than any other.

Right now I've collected three (one of which was kind of flaky but seems to still work) 2TB disks as cold spares for replacing disks in my RAID array and my PVR. Both these are being backed up to another set of RAIDed hard drives on two different computers as the machines don't have enough sata or psu plugs to get a large number of hard drives connected to them... at least these machines are being powered up and written to occasionally, but I wonder about the disconnected cold spares that don't get powered up.

One thing about original unopened hard drives is that they are sealed with desiccant. Perhaps I need to seal open-box spares in desiccant for longevity?

[momaka]

Re: failing hard drives after sitting for years...

With the myth about water pumps failing when cars sit for a while, I wonder what people have seen with hard drives.

FIFY
Water pumps (or rather their bearings) don't fail from just sitting around. They fail if the bearing seals have been compromised prior to the car being stored for a long time. With a compromised seal, the coolant enters the bearings and starts displacing the lubricant. Same happens in front-loader washers (but with water entering the bearing.) If you use it often enough, the bearing just starts to wear out and becomes super-noisy... and may eventually seize if it gets too worn to the point a bearing splits or shears. (Our front loader washer went like this and it sounded like a freight train before I finally replaced its bearings.) But otherwise, with the car sitting for a long time, a water pump will only seize if the seals on its bearings were compromised and the coolant had time to sit around and rust / seize the bearings. That's when you try to start the car the next time and... *nothing*

Do people see hard drives failing after being sat on the shelf for a long time unpowered versus a hard drive that's at least powered up at least once in a while, say, weekly?

From a stock of about 50-70 mixed HDDs (2.5" and 3.5") for various hobby PCs, I've never had one fail from sitting on the shelf. Some have sat only for a few years, others for more than a decade. I've only had a few develop a few more bad sectors... but they already had bad sectors to begin with. So I suspect the issue with those is just down to marginal / less-than-perfect magnetic coating from the factory. And more than likely, those HDDs will (and have) gotten worse over time regardless of how much I've used them.

Wonder if it's worth spinning up the spare spinners once in a while to keep the bearings lubricated?

If they are stored in reasonable storage conditions, then no.
If not, then I don't think it will matter.

Wonder how I should best keep my cold spares for my RAID "fresh" for eventual failures of the working drives.

As per the most drives' datasheets, just store between the allowable temperatures and humidity conditions... and MOST importantly, in an environment where the temperature does NOT change suddenly.

I've had people drop systems off here that have been in basements, garages, and sheds for years. Most all of which I was told 'worked' when they were stored, they just upgraded. One of the common denominators was failed HDD's. This area is miserably humid in the summer; which in my opinion (based on observation) is one of the factors as to why...leads to condensation....that can't be kind to platters & data.

THIS ^
It's actually not only high humidity, but also sudden changes in the temperature - basically the same as when you leave a car outside in the summer overnight. When you get to it in the morning, there is condensation on its windows. This happens both due to the high air humidity and the sudden changes in the environmental temperatures. And like you said, condensation, especially on the platters, is just not a good thing. In fact, this is what kills most electronics and PCBs that sit outside if they don't have some kind of conformal coating. In the case of HDDs, it's the platters that are more sensitive. And that's why if you've ever looked at any HDD's datasheet, it always gives the storage temperatures and humidity % for NON-CONDENSING conditions (i.e. no sudden changes in temperature that would cause the humidity in the air to cause condensation.) Don't follow that, and you're bound to run into issues.

I hope to also discount hard drives moved around in a pile. That I've found out is detrimental to life of disks. At minimum they need to be separated by something, antistatic bag or whatnot, but if they're in a pile, aluminum to aluminum, they die...

Never had this issue, and I've been storing many of my HDDs like this for over a decade.

the exception is seagate.
those are total shit if they are sata.

Depends... but generally I agree with you.
Their 7200.7 series were bullet-proof... and probably their last reliable drives. These did come in both PATA and SATA versions, and both were equally reliable (though the PATA versions were more common, as 7200.7 was more or less from the early-mid 2000's, when SATA was still not necessarily mainstream.)
Then came the 7200.9 line, which was... OK (not great, not terrible either.)
Then came the 7200.10 and 7200.11, which further declined in reliability.
And then there's the 7200.12 and onwards, which I won't trust worth a damn - too many problems with what seems to be due to poor magnetic coating / media. RoHS also took a toll on some in the way of PCB contacts to the head amp getting oxidized. But this one is easy to fix - just take off the PCB and clean the contacts... and if you're really serious about not having the issue come back, re-coat with a thin layer of 60/40 solder. I've done that to many of my WD HDDs, since those are even more prone to PCB contacts oxidizing (not the SATA data and power contacts, though - those are gold-plated and don't oxidize.)

It's actually possible for drives to get over-lubricated too, if you remember the IBM Deathstars...

Those actually died due to the magnetic coating separating from the glass platters.

Some also had issues due to buffer failure. My DTLA307020 is one of these - once it heats up past about 41-42C, it starts to get corrupt data in the buffer and then starts to do the "click of death"... though it's not a "real" click of death where the head can't read the disk. Rather, it's the buffer corrupting data and feeding it to the head amp... which erroneously picks up the wrong data then. I've replicated this issue many times now. If I let the HDD sit down and cool for a few minutes, I can restart it and it works just fine again.

Funny thing is, I have an old 250 GB IDE Maxtor that does the same thing. If I put a fan to blow air specifically on the buffer IC, it never has problems. But with no air cooling, it starts to go haywire. The DTLA307020 is the same too.
Now I also have a DTLA305030, and that one was decently reliable - has racked up a few bad sectors from magnetic media issues, but otherwise doesn't have the buffer issue that the 307020 has.

Wear on the motors/bearings is the worst when a drive spins up

That's a myth - at least for the bearing wearing out.
The bearings don't care if they spin up or down or if they spin slower than their rated speed.
The motor does get a little stressed when spinning up the platters... but it should NEVER fail for the HDD's life due to this. Same with the MOSFETs/motor driver IC, unless they were marginally chosen to begin with.

It's been my experience I get a much longer lifespan out of drives by letting them spin 24/7 than to use power save features.

Not necessarily.
It really depends on the make and model of HDD we're talking about, though.
In my experience while briefly working in IT, Seagate has always had "weaker" bearings than WD, and so Seagate HDDs always seem to do better when being powered down when not in use (for extended periods of time and not the crap that WD was doing with the Green series.) Meanwhile, my experience with WD has been the opposite - they seem to have better quality bearings, and so indeed seem to last longer when left running 24/7 and not power-cycled too often. Power-cycled them too often, and they start to fail. Of course, this was very specific to WD and Seagate HDDs from the early-mid 2000's - i.e. mostly Seagate 7200.7 through 7200.12 series and WD 80-250 GB HDDs from the same years. With that said, it seems that latter WD HDDs improved on this and are OK with frequent power cycles. I'm running several 250 GB 3.5" Blues, and they have been great so far.

It takes hardly any work to maintain the spin once it's going, so at that point controlling heat will help prolong the bearings themselves.

But at the same time, keeping the platters constantly spinning wears out the spindle motor bearings faster. So just saying that keeping HDDs running 24/7 is better for them is also not true.

Actually, the best is to have a balance, where the HDD is not power-cycled too frequently, but also not kept spinning 24/7 for no reason either.

Some HDDs use a glass platter with a magnetic coating applied and I've heard of those coatings failing due to some combination of heat or humidity.

The IBM "Deathstars" (both Deskstar and Travelstar) were one of the earlies (if not the first) HDDs to use glass-platters almost exclusively.
The idea was that the glass platters could tolerate higher heat, so the heads could be brought closer to the platters to get better / faster read/write rates and less erroneous reads. The idea was great on paper, but terrible in practice (at least until the technology was improved on later), because glass and rust (the magnetic coating) have different expansion rates. So even though the glass platters themselves could tolerate higher heat, the magnetic coating couldn't... and often would "ridge over" (again, due to metal having higher expansion rate than glass), get caught by the R/W head (that was now supposed to sit much closer to the platter), and then *BAM* - head meets platter, you get some bad sectors and some magnetic dust inside the HDD. The filter inside the HDD may or may not catch that dust... but eventually, that dust would then make the head crash somewhere else on the platter and ad-infinitum until the HDD completely fails.

So yeah... glass platters were eventually ditched. I think Hitachi did manage to successfully develop the technology, though, and make it reliable. But in the end, the technology was abandoned.

[Kougar]

Re: failing hard drives after sitting for years...

Not necessarily.
It really depends on the make and model of HDD we're talking about, though.
In my experience while briefly working in IT, Seagate has always had "weaker" bearings than WD, and so Seagate HDDs always seem to do better when being powered down when not in use (for extended periods of time and not the crap that WD was doing with the Green series.) Meanwhile, my experience with WD has been the opposite - they seem to have better quality bearings, and so indeed seem to last longer when left running 24/7 and not power-cycled too often. Power-cycled them too often, and they start to fail. Of course, this was very specific to WD and Seagate HDDs from the early-mid 2000's - i.e. mostly Seagate 7200.7 through 7200.12 series and WD 80-250 GB HDDs from the same years. With that said, it seems that latter WD HDDs improved on this and are OK with frequent power cycles. I'm running several 250 GB 3.5" Blues, and they have been great so far.

Well I try to avoid delving into HDD brands because I'm only a sample size of one and everyone has had their own unique experiences. But my experience leans the other way. WDs would just die or crash after a few years often with no warning signs never getting to that 5-year requirement I expect from a quality drive. The really old <2004 WD drives were very good, but only had a single modern WD make it to five years whereas Seagates mostly just keep going as long as I kept them cool. Seagates were always particularly heat sensitive, just running them in an 80-90 degree room would make some models lose the ability to read/write with the platter and start racking up the reallocated sector count, or even disconnect outright from the system. But they would resume working fine once they cooled down and the AC got fixed.

I bought eight "identical" 320GB 7200.10's and none ever died even though half were used for a NAS that I had disabled the spin-down timer on. I eventually sold them off after upgrading the NAS capacity, but I kept the one with the heat-caused reallocated sector count just for fun. It currently has 12 years and 266 days of power-on count. It's the only mechanical HDD in my new desktop. I can tell one of the bearings is beginning to wear, but 13 years of rotation on a 17 year old drive is exemplary. I still use it as a scratch disk even though it's not necessary with SSDs. Incidentally not even the WD Black SSD I have is entirely issue free, it likes to sometimes not initialize at boot which requires a full power off/on cycle to clear as a reboot won't do it. It's a shame SK Hynix doesn't make HDDs, Hitachi was my preferred until WD bought them.

But at the same time, keeping the platters constantly spinning wears out the spindle motor bearings faster. So just saying that keeping HDDs running 24/7 is better for them is also not true.

Actually, the best is to have a balance, where the HDD is not power-cycled too frequently, but also not kept spinning 24/7 for no reason either.

I'd agree on principle if you can trust Windows to keep the balance, but usually it doesn't. It'll sleep drives after just 20 minutes, so the moment you open My Computer or any Explorer window that has the Navigation pane visible it would spool up sleeping drives regardless of if they were directly accessed. Windows 7 was notorious for this. Users were oblivious this was going on for the life of that OS. Maybe Windows 10/11 fixed it, but still I don't trust Windows enough to leave the default power profiles in place regarding hard drives (or much else for that matter).

Same goes for NAS use, if the NAS is always-on and semi-regularly accessed then it can still be counter-productive to constantly have the entire array spinning down just to wake every hour.

So yeah... glass platters were eventually ditched. I think Hitachi did manage to successfully develop the technology, though, and make it reliable. But in the end, the technology was abandoned.

Actually, I did some more digging and they're phasing out aluminum substrates in 2.5" drives for glass. Some 20TB and larger capacity HDDs are also using glass. Apparently its use never went away after the IBM Deskstars.

[RJARRRPCGP]

Re: failing hard drives after sitting for years...

This area is miserably humid in the summer

Is the dew point regularly in the low 80s or very close to 80F?

I know that OTOH, a dew point of near 80F, is rare in Vermont and is also rare in a good chunk of New Hampshire.

IIRC, July, 2018 had an abnormally-high-dewpoint spell, it possibly hit 80F or was most likely 76F+! I remembered sweat pouring off of me when doing yard work, when "weed wacking", IIRC, LOL.

IIRC, 2006 was like that, too! And wouldn't be surprised if the case for 2011 as well. Those were soupy-wet years in the northeast!

[eccerr0r]

Re: failing hard drives after sitting for years...

Speaking of NOS disks that have servo info recorded on the platters, would that be a dead disk if it sat too long, say a 30 year old hard disk? (I suspect 40 year old hard disks are stepper motor and don't need servo marks like VC actuators.)
What's the magnetic lifetime of this data versus user data, wouldn't user data end up being inaccessible if the servo data fades?

[dmill89]

Re: failing hard drives after sitting for years...

Speaking of NOS disks that have servo info recorded on the platters, would that be a dead disk if it sat too long, say a 30 year old hard disk? (I suspect 40 year old hard disks are stepper motor and don't need servo marks like VC actuators.)
What's the magnetic lifetime of this data versus user data, wouldn't user data end up being inaccessible if the servo data fades?

While I'm not sure how long the data will last in storage (I'm not sure anyone is), I wouldn't worry too much unless the drive was exposed to a strong magnetic field. I've got floppy disks from the early/mid-1990s that still have their data (and those are only "officially" supposed to last around 10 years and the magnetic tracks are much weaker than on a HDD), I haven't run across any 30+ year old NOS HDDs yet (I haven't really been looking for anything that old either) but I did recently get a couple NOS circa 2005 Seagates (one 2.5" laptop and one 3.5" desktop both PATA) and both work just fine.

[momaka]

Re: failing hard drives after sitting for years...

Speaking of NOS disks that have servo info recorded on the platters, would that be a dead disk if it sat too long, say a 30 year old hard disk?

Maybe. Maybe not.
Like dmill mentioned, probably no one knows how long these things will last.
That said, I don't think it matters if the HDD sat unused or used occasionally. The servo/track info on the platters is read-only, from what I know. So the HDD doesn't ever "refresh" the data there or correct any errors. So if those tracks get corrupt, that's it, drive done for.

Well I try to avoid delving into HDD brands... and everyone has had their own unique experiences.

Same and agreed.
But working in IT / PC repair for a while, I just couldn't help but notice certain "patterns" in regards to certain brands... though you are right, it's not just down to the brand. The model / era / years the HDD is from matters too.
While I've had older (mid 2000's PATA and early SATA) WD's not like me power-cycling them much, the newer ones from the 2010's an onwards don't seem to care as much.

But my experience leans the other way. WDs would just die or crash after a few years often with no warning signs

Yes, that's been my general experience with WD too - they either work or they don't work at all. I did have one sort-of give me warning signs - had trouble initializing / doing a seek test at power-on a few times, and then the HDD temperature reporting some bogus high number (127C ) after a cold power-up. I suspect the head amp ICs were starting to go and mis-reporting.
But other than that WD, most others have just been either go or no go. Some older ones (again, mid 2000's) did start to rack up a few bad sectors, but did so silently and still continue to work to this day.

... whereas Seagates mostly just keep going as long as I kept them cool.

Yeah, I agree about Seagate being heat-sensitive. Maxtors too.
IBM / Hitachi - hit or miss, depending on the model and years made.
Had a 2.5" 20 GB PATA run in an old Pentium 3 for close to 10 years. Poor ventilation and heat from the laptop often would have it run close to 60C! It finally corrupted the Windows install and I decided to retire it. Still works OK, though. Has an aggressive head ramp parking schedule, though... and glass platters .

I bought eight "identical" 320GB 7200.10's and none ever died even though half were used for a NAS that I had disabled the spin-down timer on. I eventually sold them off after upgrading the NAS capacity, but I kept the one with the heat-caused reallocated sector count just for fun. It currently has 12 years and 266 days of power-on count. It's the only mechanical HDD in my new desktop. I can tell one of the bearings is beginning to wear, but 13 years of rotation on a 17 year old drive is exemplary.

Nice!
Yeah, when Seagate makes a good drive, they DO last for quite a while.
Though I guess the same can be said about any brand.
I consider the 7200.9 and 7200.10 series a bit "random" when it comes to reliability. Most do seem to last OK, but a good number also didn't. Yet, a few proved to be bullet-proof too.

I'd agree on principle if you can trust Windows to keep the balance, but usually it doesn't. It'll sleep drives after just 20 minutes

Yes, I never leave the default profiles.
I usually turn HDD power down to OFF and instead just set the OS to go to sleep after a certain period of time (typically 1-3 hours, depending on the machine.)

Maybe Windows 10/11 fixed it, but still I don't trust Windows enough to leave the default power profiles in place regarding hard drives (or much else for that matter).

Actually, I did some more digging and they're phasing out aluminum substrates in 2.5" drives for glass. Some 20TB and larger capacity HDDs are also using glass. Apparently its use never went away after the IBM Deskstars.

Interesting. I guess I haven't kept up with the times. Then again, most of my IT "experience" is from a decade+ ago. I don't do much PC repair for people these days. It's more of a hobby now.

[TechGeek]

Re: failing hard drives after sitting for years...

Quantum 1.2GB SCSI drive in a Power Macintosh 7200/120, sat powered down for 15 years and woke up after a couple of attempts. Probably landed in the computer shop that it was abandoned at because it's failing. Many missing or corrupted files, and the drive sometimes gets uppity and the Mac crashes with a "bus error" message. Will be replacing it and rebuilding the system software.

(Just to further disprove that the data on a hard disk fades over time. SSDs, OTOH, are the opposite. They suffer from data retention issues long-term, and will self-erase if left powered down long enough due to the charge in the NAND's memory cells fading away.)

[eccerr0r]

Re: failing hard drives after sitting for years...

I don't remember which quantum models, but there was a whole fiasco with quantum disks (which apple uses profusely) whose heads would stick to the platter, affectionately called "sticktion" and the disk would simply not spin until a few power cycles or perhaps heating. As far as I know this is specifically quantum branded disks, but only specific models. I've had a 30G quantum disk but that failed for a bearing failure - it was bad out of the box!