A Colossal HDD Failure

[eccerr0r]

Re: A Colossal HDD Failure

FFFFKK.

The drive is very intermittent. Worked perfectly fine overnight, was able to update Linux and build a new kernel, then today after moving it to another machine, it blew chunks trying to write again. Seems it has no problems reading still.

Oddly enough the machine that had no problems last night was the "home fixed special" - an amalgamation of the flaky EP43, the SP300 that fried to a crisp (and fixed with "blue light specials"), and the hot hot hot Core2Duo...nevertheless it worked fine until I transferred the drive back to the P4 with a PSU that hadn't ever given me problems before minus the MB ATX12V connector overheating and jury rig repaired...

Mystery continues though could very well be a PSU issue still...

[eccerr0r]

Re: A Colossal HDD Failure

Look at the fact that I specified -n : a non destructive write... This reads data off the disk, writes test pattern, and then writes the old data back - nondestructive (and is very slow!). Without -n would have meant read-only, but it definitely wrote..

Note that I already know there are a lot of sectors that had problems writing in the past and was hoping to get them ignored by software (NOT HARDWARE) by adding them to the software based bad sector map. It completely ran out of spare sectors already.

And no, I'm not disposing of this drive yet, I'm milking it until the motor dies or head scrapes the rust off...

[TechGeek]

Re: A Colossal HDD Failure

FSCK this...wtf.

badblocks -n reports 0 FAILURES!!! Come on, write the bad block locations to the ext4fs bad block list!!!

The best way to expose bad blocks on an HDD is to either run SpinRite on it at Level 5, DBAN it on the DoD Short wipe, or to run a destructive badblocks test against the drive.

Destructive BadBlock test (1gig ram in machine)
 badblocks -b 4096 -c 98304 -p 0 -w -s /dev/hda

It is not good practice to attempt to test a drive for bad blocks by performing a read-only test. A read only test, however, will expose damaged sectors that cannot be read. These will generally show up in the Current Pending Sectors Count, or CPSC SMART parameter. If a drive has a non-zero CPSC SMART parameter, that likely means that data has been damaged as a result of the drive's failure mode, and it should be backed up immediately (preferably by using ddrescue to maximize the amount of salvageable data), wiped, and disposed of.

[eccerr0r]

Re: A Colossal HDD Failure

FSCK this...wtf.

badblocks -n reports 0 FAILURES!!! Come on, write the bad block locations to the ext4fs bad block list!!!

[eccerr0r]

Re: A Colossal HDD Failure

Ugh looks like the WD205BA just choked on me... cant write new data ...

Still can read existing data however.

Well it said it was bad, it really is!

[momaka]

Re: A Colossal HDD Failure

haha! here's a wd cav blue 500gb ide drive i bought used off junkbay with 56k+ poh. it also has over 50k start stop and load unload cycle counts.

I see only 149 "hard" power cycle counts. The 50k start/stop and load cycle counts may just be head parking jobs, depending on how the HDD is made to report these. That's still a lot... though nowhere near as bad as WD Green drives - now those drives really like to waste time parking heads more than doing anything else.

tho one died on me recently a couple of years ago and totally bricked itself. undetectable by bios and wont spin up. just makes a few clicking noises on power on and then no further noise or activity from it.

That actually may be repairable, as it sounds like the spindle motor is not spinning. This could be either due to stuck head on the platter or perhaps motor driver is bad.

To check for stuck heads, sometimes this is easy to do on Seagate and WD drives by holding the drive in your hand with the spindle motor right above the middle of your palm. Then spin the HDD back and forth. The intertia will make the platters not spin as much as the HDD, so the platters will rotate a little back and forth inside. As such, since WD and Seagate drives don't use a parking ramp but rather park heads on a designated track on the platters, you should be able to hear a faint "wheezing" while spinning the HDD in your hand back and forth. (Note: may listen very carefully in a quite room, since newer drives tend to be pretty quiet with that test.) If you can't hear the heads sliding on the parking track, they may be stuck somewhere else on the platter, hence the HDD not spinning.

*OR*

Since the HDD is not detecting at all, check for blown fuse links and other power components to the controller and motor driver circuits.

OMG! thats like 12 and a half years in total of use and being powered on! thats like a centurion in human years! lol!

Yeah, I know.
I bought it used refurbished, so it was probably working 24/7/365 before I that. I currently have it as a main drive in this PC. It gets very occasional use now, being located in my grandparent's old family house in the countryside - maybe a few 10's of hours a year tops.

i'd bet that hard drive is going to be the last surviving hard drive of humanity some day. lol! probably, thousands of years later, when an alien race discovers earth and the ruins of human civilization, they will find that hard drive. the only surviving record of what humanity is like...

Well, given what I said about it above... yeah, it most likely will be one of hte last HDDs to survive humanity. Moreover, the foundations of that house were built over 100 years ago, and the half-basement has 6-foot-thick rock-and-clay walls. Calling it a bunker is equally appropriate. If I place the PC down there, it would surely be a relic to be discovered many generations after I have passed... provided moisture doesn't destroy the PCB on it (which it probably will, despite this being an old drive from before the lead-free era and using good quality leaded solder.)

yikes! that makes me now think carefully of what i put on my hard drives. what if i put pr0n on the last surviving hard drive of humanity and the aliens find my pr0n collection?! they will get the impression that humans just luv being nakkied and they just luv having sex/mating with each other the whole day?

Sounds about right, actually.

so they surmise that overbreeding killed off humanity?!

Sounds about right x2

maybe i should put science and animal documentaries on the last surviving hard drive instead, so the aliens get a better impression of humanity.

Oh, you mean this, right:
https://www.youtube.com/watch?v=xat1GVnl8-k


Sorry, I realize this is not VIP... but it's on YouTube, after all. So it can't be that bad, can it?

[ChaosLegionnaire]

Re: A Colossal HDD Failure

haha! here's a wd cav blue 500gb ide drive i bought used off junkbay with 56k+ poh. it also has over 50k start stop and load unload cycle counts.

i like those line of drives! they are tough old boats! tho one died on me recently a couple of years ago and totally bricked itself. undetectable by bios and wont spin up. just makes a few clicking noises on power on and then no further noise or activity from it. ah well... no drive or drive line is ever 100% perfect. all hard drives fail... eventually... but other than that, those drives have been pretty reliable for me so far. i have hoarded like 5 or 6 of those drives. hahaha! im such a hoarder too rofl!

I think my highest is on a Hitachi 120 GB IDE HDD that has close to 110k POH now.

OMG! thats like 12 and a half years in total of use and being powered on! thats like a centurion in human years! lol!

i'd bet that hard drive is going to be the last surviving hard drive of humanity some day. lol! probably, thousands of years later, when an alien race discovers earth and the ruins of human civilization, they will find that hard drive. the only surviving record of what humanity is like...

yikes! that makes me now think carefully of what i put on my hard drives. what if i put pr0n on the last surviving hard drive of humanity and the aliens find my pr0n collection?! they will get the impression that humans just luv being nakkied and they just luv having sex/mating with each other the whole day? so they surmise that overbreeding killed off humanity?! wow thats definitely a good deterrant now to make me make better use of my time instead of dloadin and watching obscene films all day. hahahaha!

maybe i should put science and animal documentaries on the last surviving hard drive instead, so the aliens get a better impression of humanity. that we were intellectuals, scientists, great knowledge seekers and great preservers of the natural world...

[momaka]

Re: A Colossal HDD Failure

But I thought most "modern" hdds that had voice coil head actuators (pretty much all hard drives that have under 30ms access time) could not be low level formatted as there's no way to recreate the servo data.

Well, it's a low-level format on the "data" level for modern HDDs - that is, all *data* tracks on the HDD are filled with zeros. But the tracks that contain firmware-specific info or anything relating to track position don't get LLF-ed. Then again, those tracks don't contain user data, so there is no need to LLF those.

But yes, you CAN LLF a modern HDD... and sometimes it does clear out old bad sectors, if those happened to occur from interface or other external error (or power outage?)

Just 50K POH? That's it?

On *average*, yes.

I have a few of those old WD drives that are only in their 20-30k POH, and also a few in the 60-80k range. I think my highest is on a Hitachi 120 GB IDE HDD that has close to 110k POH now.

On that note, most of those hours aren't from my use - all of these HDDs were bought/acquired used with most of the POH already on there. As for me, I don't really add that much POH to my HDDs. I'm more of a heavy "power-cycler" - that is, if a computer is not being used, I keep it OFF and turn it On only as needed. Thus, often times I'll switch off or put the PC in Standby (Sleep) if I'll be away for more than 30 minutes... though it does depend on the season - in the Winter, I do sometimes leave my PC on for longer, even if I won't be using it for an hour or two. But in the summer when it's hot and A/C is running, I keep no computer On for any reason if it's not needed.

With that being said, my main and gaming PCs get the most power cycles - about 1 per day, typically (and sometimes more and sometimes less.) So that's ~360 power cycles / year. And in terms of hours, it can vary anywhere from 0 to 6 hours... so about 2-4 on average / day. As such, my POH to PCC (power cycle count) ratio tends to be about 3:1 - i.e. for every 3 hours of operation, I put about 1 power cycle.

I used to be much worse in high school, though, where I'd power-cycle/standby my PC multiple times a day due to having a loud PC and not wanting the extra noise in the room. So back then, my main PC had only about 4500 POH in the course of 3 years, but over 3000 power cycles.

[eccerr0r]

Re: A Colossal HDD Failure

Think it was a WD205BA or something like that. I got this drive from a dump pile many years ago, so I do not know its history -- it may have been abused in a prior life, so while it may have been solid then, don't know if it's solid now.

Oh hell there's nothing really useful on this hdd. But I thought most "modern" hdds that had voice coil head actuators (pretty much all hard drives that have under 30ms access time) could not be low level formatted as there's no way to recreate the servo data. But indeed those ancient stepper motor drives, LLF is a good thing to try?

Just 50K POH? That's it?

[momaka]

Re: A Colossal HDD Failure

It's an old WD drive, indeed UDMA/PATA. Currently using it as a root disk, definitely not going to store anything critical on it, it's probably on life support right now...

Nah, old WD drives like that were pretty solid. It's not even at 29k POH. Most of the ones I got are over 50k. So you're probably just about breaking it in.
Care to share a model number, though? Not all old WDs were that rock solid. Some of the slightly newer ones indeed developed bad sectors over time. But yours, with those UDMA CRC Errors is telling me there may be something screwy going on with the interface.

If you had a spare HDD to clone the data onto so you could check this one out in more detail... running a LLF (Low Level Format) should tell you if that drive really had bad sectors or not. If drive shows 0 bad sectors after LLF and passes a surface scan, it's probably solid.

[RJARRRPCGP]

Re: A Colossal HDD Failure

I also found an old 40 MB (MegaByte) Seagate ST-251 5.25" in the scrap pile at work. Someone removed its top a long time ago just to look at the inside and never put it back on.

Reminds me of that 240 MB Conner! I took the cover off once to look at the inside, wiped off any dust that I saw on any part of the platter(s) then screwed the cover back on and it was like nothing happened. It still worked fine!

[eccerr0r]

Re: A Colossal HDD Failure

It's an old WD drive, indeed UDMA/PATA. Currently using it as a root disk, definitely not going to store anything critical on it, it's probably on life support right now...

[momaka]

Re: A Colossal HDD Failure

I see some numbers in UDMA CRC Error Count field. Old PATA drive by any chance? If so, maybe check cable or motherboard... or PCB contacts if it's a newer WD HDD.
Still, there are more bad sectors than UDMA CRC's, so the drive likely is failing. Could also be buffer chip issue (a known problem on some Maxtor and IBM HDDs) - usually showed up when the drive cooling was passive.

[eccerr0r]

Re: A Colossal HDD Failure

YAY I'M GOING TO STORE MY BITCOIN ON THIS 20GB DRIVE WITH THIS SMARTCTL REPORT!

Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME     FLAG   VALUE WORST THRESH TYPE   UPDATED WHEN_FAILED RAW_VALUE
 1 Raw_Read_Error_Rate   0x000b  100  100  060  Pre-fail Always    -    0
 2 Throughput_Performance 0x0005  100  100  050  Pre-fail Offline   -    0
 3 Spin_Up_Time      0x0007  129  129  024  Pre-fail Always    -    347 (Average 335)
 4 Start_Stop_Count    0x0012  100  100  000  Old_age  Always    -    1211
 5 Reallocated_Sector_Ct  0x0033  001  001  005  Pre-fail Always  FAILING_NOW 30
 7 Seek_Error_Rate     0x000b  100  100  067  Pre-fail Always    -    0
 8 Seek_Time_Performance  0x0005  100  100  020  Pre-fail Offline   -    0
 9 Power_On_Hours     0x0012  096  096  000  Old_age  Always    -    28780
 10 Spin_Retry_Count    0x0013  100  100  060  Pre-fail Always    -    0
 12 Power_Cycle_Count    0x0032  100  100  000  Old_age  Always    -    850
196 Reallocated_Event_Count 0x0032  100  100  000  Old_age  Always    -    8
197 Current_Pending_Sector 0x0022  100  100  000  Old_age  Always    -    0
198 Offline_Uncorrectable  0x0008  100  100  000  Old_age  Offline   -    0
199 UDMA_CRC_Error_Count  0x000a  200  200  000  Old_age  Always    -    17

*cough*cough*wheeeeze*

(Computer is actually still working fine...for now )

[momaka]

Re: A Colossal HDD Failure

^ Good question. I didn't think to check.

If all the motor driver IC needs is just an enable signal, then it may be possible. But I imagine that most likely the motor driver IC probably talks to the main IC over I^2C bus or similar... and if that's the case, then this may not be so simple.

I guess I'll check the next time I have it out.

[TechGeek]

Re: A Colossal HDD Failure

Nice. Would there be a way to directly wire the motor driver to the spindle motor and power it directly, instead of having other control logic get in the way?

[momaka]

A Colossal HDD Failure - Part 4 / HDD Bench Grinder

And here’s the project I was talking about.


It’s a desktop bench grinder!
(Get it? OK, maybe not so funny… I’ll lay off the excess water drinking. Clearly it’s drowning out my brain cells too much. )

Silly jokes aside, this thing is actually not that useless. The motor in this particular HDD is fairly strong, because it hardly slows down when grinding smaller objects, like sharpening drill bits or grinding away wood and plastic. I’m guessing this probably has something to do with the fact that it’s a 4-platter HDD, so it probably has a slightly more powerful motor than, say, a 1 or 2 -platter drive.

Now, I’d love to claim this idea as original… but it’s not. I saw someone do this to an HDD more than a decade ago (it was a YouTube video, if I remember correctly - back when YT was still a new thing.) So the idea to try this has been sitting in my head all these years, and I finally decided to execute it.

That being said, to make one of these, first make sure the HDD you use doesn’t mind the headstack assembly being gone. Some HDDs will not spin up or will spin up and spin down when they can’t read anything from the heads. Mine actually does that (spins down after a while), but it stays spinning for around a minute or two, which usually gives me enough time to grind whatever it is that I need. And secondly, I think it helps that I happened to pick a 4-platter HDD, as probably the motor and motor driver IC are a little stronger in order to cope with the extra spinning mass, as mentioned earlier. So I think this is probably what gives my HDD grinder a bit more torque. I’ve tried slowing down a few other HDDs by hand before, and they seemed much easier to slow down / stall.

Also, for those who are curious how I attached the sandpaper to the platter: wood glue. It spreads nicely and evenly, so it won't create bumps on the sandpaper sheet. Basically, once I had the sandpaper cut down to the proper shape, I applied the wood glue to it and then put in on the platter (the platter was separate from the spindle, as shown in my previous post.) After that, I flipped the plater with the sandpaper facing to the ground and placed the whole thing on a flat, even surface. Then, I put a heavy object on it (a UPS transformer, in my case) to keep the sandpaper glued flat to the platter. Wait for a day and it was perfect.

Anyways, that’s all I have for this project for now. I might eventually get different types of sandpaper sheets and do the rest of the platters too. That way, if I need finer or coarser grinding, I can just swap the platters to swap sandpaper. And I’m also thinking maybe add a small “table” / shelf in front of the disk, so I have a surface to lean objects against when sanding. IDK, we will see as I get more ideas.

[momaka]

A Colossal HDD Failure - Part 3 / SOUNDS

Perhaps only hardcore geeks may appreciate this post, as I am including an attached file with the *sounds* from the two failed HDDs. Hope you enjoy.
failed HDD sounds (112 kbps mp3) ZIP file

What's interesting is that while the 2nd HDD was in storage, the heads (or whatever was left of them on the actuator arm) got stuck to the platter, so the HDD couldn't spin up and made this beeping noise instead (which I recorded as well. ) I opened the 2nd HDD up and gave it a manual spin, then closed it and recorded the sounds.

And what do you know, it works now!
.
.
.
.
.
Nah, just kidding.
Both HDDs are still dead as a rock.

In any case, I decided to disassemble the 2nd HDD and clean it, as I had a project in mind for it. So the first thing to come out for the cleaning was the headstack assembly. Notice the missing heads. I found a few of them (in pieces) stuck to the actuator arm magnet.


While at it, I noticed the bearing on the actuator arm had some seriously rough spots when turning it by hand (suggesting failing/failed bearing.) I wonder if this is from wear from high # of power On hours. Perhaps this is what caused the heads to touch the platters? Or was this rather a consequence of when the heads got ripped off from the actuator arm and the force from that damaging the bearing? As you may recall, I mentioned in the first post that as I had the 1st HDD running for the first time, it made a loud *whack* noise after spinning for a while. When that happened, the HDD jumped a little from where it was resting (perhaps that was most likely another / the last of the heads getting ripped off?) If this happened, I imagine the actuator arm bearing saw quite a bit of force exerted on it.

Another thing I noticed while removing the actuator arm from the 2nd HDD is that I didn't need to loose the screw on the bottom of the bearing. Once I removed the top screw (that holds the actuator arm bearing through the top cover), the screw holding the bearing from the bottom was already loose. You can actually see it in the pictures shown below that it's still sticking up from the bottom. Normally getting to that screw requires removing the PCB and also a few shiny warranty-like stickers.

This makes me wonder, since this is a refurbished HDD, if someone at Seagate didn't torque the bottom actuator arm bearing screw properly. And if so, perhaps this is what led to the head crash. After all, why did both of these refurbished HDDs crash the same way? Given where the HDDs came from (CCTV recording equipment), they probably had high # of power On hours and probably a good number of reads/writes (though nothing excessive, I imagine, as the footage was low bit rate.) Still, I don't think that kind of wear alone could loosen the bottom bearing screw. Or could it?

Anyways, I gutted all of the “unnecessary” components away from the 2nd HDD (for the project I will show a bit later.) With that said… let me just warn everyone once more - DO NOT get that fine platter dust on anything you care about! It's really fine stuff and really hard to wash out. I imagine it's probably bad if you breathe it in too, given how fine it is. Luckily, it's magnetic and prefers to stick to metal surfaces rather than stay too long in the air. But it can stay up for a little. Regardless, DO NOT breathe that dust in!

Also, this dust was so fine and sticky that I couldn't just brush it off (and for the above reasons, I had no reason to anyways.) Thus, to get the HDD clean, I had to remove all of the platters and wipe everything down with damp paper towels soaked in soapy solution. Here are a few more pictures I snapped while doing that. (BTW, notice the scratching occurred on ALL platters and not just the top one.)




Kind of ironic about the cleaning part, given the project this HDD will be used for.
Stay tuned!

[Dan81]

Re: A Colossal HDD Failure

Try resetting SMART. Saved a 2TB drive that way.

[Xenon-Codex]

Re: A Colossal HDD Failure

This 7200.11 failed at quite high power on hours. It was not recognized by any computer, until one last time it started and I got all data recovered. It came to me from trash, and I used it at least 3 years as my main HD (for photos etc).