[SOLVED] Dying Seagate 3TB HDD

NateBeast

Honorable
Jul 3, 2015
210
1
10,695
So like 4 days ago my HDD wasn't showing up in Windows one time when I turned it on, so I shutdown, checked the cables and turned it on again and it showed up.

So I thought I was all good. Then it happened again, and was really slow to use. So I checked SMART with cryataldiskinfo and it said caution, for pending, uncorrectable and reallocated sectors. One was at like 700 and the other 2 were the exact same number, around 10k, don't remember which were which tho, sorry.

So I started copying my importabt files off the drive, and the 700 went up to 900 something during that time. Then, I used Macrium reflect to image 2 partitions on my HDD, F: and I:.

However, when I tried E:, I got error 6 MFT corrupt and it said to run chkdsk. I tried but it said I need to unmount it, which I did but at stage 4 it was insanely slow, like estimating 100 hours, so I cancelled it. But I did choose to do it on reboot. I rebooted once, and the drive didn't show up. Again and no drive in Windows. But the third time it showed up, so I tried a different file copier that works for broken drives. During this process the drive disappeared again.

I rebooted, and now it's in stage 4 of CHKDSK before it's in Windows, 338 of 331760 estimating 30 hours. Is there a point of letting this run, or should I just reboot. And what's the best course of action to save some of my data?

Seagate Seatools did say the drive passed SMART and Drive Self Test, but idk how that's possible...

Thanks for all the help!!
 
Solution
Oxidisation on Western Digital (and Samsung) PCBs:
http://www.hddoracle.com/viewtopic.php?f=86&t=649
Yup, my head stack pads looked pretty close to the 3rd one. I decided to try the drive in an external enclosure after giving it a light scrub, I'm getting 1-4MB/s out of the drive due to the heads parking every second... will need to do more PCB-level investigation tomorrow. Worst case, at least I got to confirm the drive contains nothing of importance.

Edit: stuffed the drive in the enclosure (instead of leaving it in open air) and it started clicking less often after warming up a bit. Must have a bad solder joint somewhere. Might be salvageable as a temp drive.

Edit2: dying drive is still going. Got a new external SATA...

USAFRet

Titan
Moderator
Step 1 - Invent a Time machine. Go back 5 or more days, and copy ALL the data off that drive.

It is likely beyond the point of recovering any data from it.
The more you mess with this, the more likely it is to actually and completely die.

Is it still under warranty?
 

NateBeast

Honorable
Jul 3, 2015
210
1
10,695
Step 1 - Invent a Time machine. Go back 5 or more days, and copy ALL the data off that drive.

It is likely beyond the point of recovering any data from it.
The more you mess with this, the more likely it is to actually and completely die.

Is it still under warranty?
Yea I should've been copying data since the first time it didn't show up in Windows... Don't think it's still in warranty, but I'll check the serial. So for any remaining hope shut off the PC for now?

Nope, warranty expired May last year. :(
 

NateBeast

Honorable
Jul 3, 2015
210
1
10,695
Power OFF, and disconnect that drive.
Is this the OS drive?

And you should have been doing backups since day 1 of turning it on. Drives die, sometimes suddenly and with no warning.
Your data is the only real thing that counts.
Agreed, definitly stupid for not making backups. It's not the OS drive, its a secondary data drive. Turned my PC off and disconnected it.
 
Clone your drive with a tool that understands how to work with bad heads or bad media, eg HDDSuperTool or ddrescue. Then run data recovery software against the clone. Avoid CHKDSK -- it will only damage your data and thrash your drive.

As for SeaTools, your experience shows why this software should not be trusted. Instead you should use SeaChest or just about any other SMART reporting tool. Use SeaTools if you need a Test Code for RMA purposes.
 

NateBeast

Honorable
Jul 3, 2015
210
1
10,695
Clone your drive with a tool that understands how to work with bad heads or bad media, eg HDDSuperTool or ddrescue. Then run data recovery software against the clone. Avoid CHKDSK -- it will only damage your data and thrash your drive.

As for SeaTools, your experience shows why this software should not be trusted. Instead you should use SeaChest or just about any other SMART reporting tool. Use SeaTools if you need a Test Code for RMA purposes.
Ok, thanks for the reccomendations, I assume I can run them off a live cd or usb? They don't need to be installed somewhere in an already existing linux installation?
 

USAFRet

Titan
Moderator
Ok, thanks for the reccomendations, I assume I can run them off a live cd or usb? They don't need to be installed somewhere in an already existing linux installation?
How to:
 

USAFRet

Titan
Moderator
And whatever happens, please take this as a wakeup to start a comprehensive backup plan.
Drive die. All of them, sometimes suddenly.

The physical drives are inconsequential...your data is what matters.

 

NateBeast

Honorable
Jul 3, 2015
210
1
10,695
Yep, definitely. Well, I have my good HDD in for copying and my bad HDD as well, and my PC freezes at the "press f2 f11 etc" screen with MB logo. unplugging sata of bad drive gets me to boot options, but choosing UEFI:sandisk (my usb) just refreshes the boot menu screen.

Disconnected both drives and this time it works. Will keep updated

Other drives don't get detected when I plug them in after, even in the boot systemrescurecd menu

Edit3: Working now with both
 
Last edited:

NateBeast

Honorable
Jul 3, 2015
210
1
10,695
So I'm mounting my 2nd good hard drive so that I can copy an image file to there, but it says it's in an unsafe state from hibernation. However, this drive isn't my C drive, its a secondary data drive, and I dont want to risk rebooting, and having the drive not show up and all. Is there a risk with just forcing it, seeing as its not my C drive?
 
A couple years ago, statistics seemed to indicate that 3 TB drives seemed somewhat more prone to failure than smaller or larger drives, and this tendency (~2x-3x the failure rate per 1,000 drives) seemed to occur with multiple drive makers' products... (go figure)
 

NateBeast

Honorable
Jul 3, 2015
210
1
10,695
A couple years ago, statistics seemed to indicate that 3 TB drives seemed somewhat more prone to failure than smaller or larger drives, and this tendency (~2x-3x the failure rate per 1,000 drives) seemed to occur with multiple drive makers' products... (go figure)
Damn, well my other is a WD Red 3TB which seems more reliable, but hopefully it doesn't fail the same way. At least I'm gonna go backup that drive once I'm done this,
 

InvalidError

Titan
Moderator
Yup, when you see or hear suspicious activity from a HDD, don't waste time on drive diagnostics, start backing stuff up immediately if there are any important files you need to get off of it.

A couple of weeks ago, my WD blue drive (~7 years old) was still at 9 pending sector count (always had 9 from the day I got it) and after I re-installed Windows yesterday, the drive started clicking every few seconds and the pending count bumped up to 10. I removed the drive from my system until I have a spare to work with to make sure I don't run out of space getting stuff off of it. Since the drive mostly contains ISOs of stuff I can just re-download and a bunch of sneakernet movies/TV series, I don't care too much if I lose some, mainly need to do a "dir /s > stuff_I_may_have_lost.txt" to make an inventory.

I decided to inspect the drive's PCB before doing anything else just in case I'd find something repairable like a bad SMD cap to explain the new behavior aside from a mechanical failure. Contact pads are corroded for some reason while the rest of the board looks pristine. Contact points between the PCB and drive body must have still been wet from cleaning during drive assembly. Going to clean those and hope the clicks go away.
 

NateBeast

Honorable
Jul 3, 2015
210
1
10,695
Latest update, around 48% done the partition it started getting a lot of read errors, so I stopped it, set it to go in reverse, and started it again, hopefully trying to get the most amount of data possible.
 

InvalidError

Titan
Moderator
Oxidisation on Western Digital (and Samsung) PCBs:
http://www.hddoracle.com/viewtopic.php?f=86&t=649
Yup, my head stack pads looked pretty close to the 3rd one. I decided to try the drive in an external enclosure after giving it a light scrub, I'm getting 1-4MB/s out of the drive due to the heads parking every second... will need to do more PCB-level investigation tomorrow. Worst case, at least I got to confirm the drive contains nothing of importance.

Edit: stuffed the drive in the enclosure (instead of leaving it in open air) and it started clicking less often after warming up a bit. Must have a bad solder joint somewhere. Might be salvageable as a temp drive.

Edit2: dying drive is still going. Got a new external SATA enclosure with USB3/UASP support to see if I could get the remaining data off any faster and also be able to monitor SMART status along the way. Interestingly enough, all error counters are staying at 0. Whatever is causing the odd seek/park activity does not appear to be related to anything media-related. I also attached oscilloscope probes to three power rails on the drive and can see 20mV transients on the +/-5V rails going to the head assembly coinciding with the anomalous seeks. Whenever the drive works normally for a few seconds, the peak-to-peak supply noise closer to 40mV, so the transients are likely a side-effect of whatever the problem is. I thought the problem may be caused by a defective G-sensor but I'm not seeing the head load/unload count tick up either, so the head rack noise I'm hearing (much louder than a normal seek, though much quieter than heads crashing into end-stops when heads go bad and can't track position anymore) apparently isn't that despite the head flight hiss briefly pausing.
 
Last edited:
Solution