Hi folks,

So I have a *mostly* unusable 500gb Western Digital disk sitting in front of me. I have some data on it I'd rather not lose, but if it turn into a paperweight it's not the end of the world. I figure it'd make a fun project to try and learn on and see if I can't recover any data from it. So without further ado here's my situation...

OS: Ubuntu Linux 11.04/Windows 7 Ultimate
Disk Model: WD5000AAKS-00YGA0
Date Manu: 16 OCT 2007
DCM: DHNNNT2AAB

Part 1 - A few weeks ago I'd be streaming episodes of Star Trek (DS9, I know, very relevant) off of my hard disk when it would stop streaming. I'd have to reboot the system to get it to start streaming again. Eventually after this happened 4-5 times I checked /var/log/syslog and saw lots of errors like the following...

Jun 9 01:39:04 rhea kernel: [ 4749.440219] ata3.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6
Jun 9 01:39:04 rhea kernel: [ 4749.440223] ata3.00: BMDMA stat 0x26
Jun 9 01:39:04 rhea kernel: [ 4749.440226] ata3.00: failed command: WRITE DMA EXT
Jun 9 01:39:04 rhea kernel: [ 4749.440233] ata3.00: cmd 35/00:00:00:18:e3/00:04:03:00:00/e0 tag 0 dma 524288 out
Jun 9 01:39:04 rhea kernel: [ 4749.440234] res 51/84:01:00:00:00/84:00:00:00:00/e0 Emask 0x30 (host bus error)

Jun 9 05:32:01 rhea kernel: [18726.882555] ata3.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6
Jun 9 05:32:01 rhea kernel: [18726.882559] ata3.00: BMDMA stat 0x26
Jun 9 05:32:01 rhea kernel: [18726.882562] ata3.00: failed command: WRITE DMA EXT
Jun 9 05:32:01 rhea kernel: [18726.882569] ata3.00: cmd 35/00:48:08:2e:52/00:01:07:00:00/e0 tag 0 dma 167936 out
Jun 9 05:32:01 rhea kernel: [18726.882570] res 51/84:01:00:00:00/84:00:00:00:00/e0 Emask 0x30 (host bus error)
Jun 9 05:32:01 rhea kernel: [18726.882573] ata3.00: status: { DRDY ERR }
Jun 9 05:32:01 rhea kernel: [18726.882575] ata3.00: error: { ICRC ABRT }
Jun 9 05:32:01 rhea kernel: [18726.882584] ata3: soft resetting link
Jun 9 05:32:01 rhea kernel: [18727.100148] ata3.00: configured for UDMA/133
Jun 9 05:32:01 rhea kernel: [18727.100168] ata3: EH complete

After seeing these I knew there was a serious issue going on. However, being the idiot I am I continued using the disk despite knowing it was failing. It would stop working and I'd just restart the system to get things going again.

Part 2 - At some point I could no longer stream Star Trek episodes off the disk anymore even after system restarts. At this point I considered the disk dead. Then I proceeded to do some stupid and probably harmful things to the disk in an effort to maybe/hopefully easily retrieve my data, including...

Freeze the disk (not in a moisture proof container)
Plug the disk in and attempt to copy data off it despite noticing a great deal of moisture on it
Take the disk apart on my desk (not a clean room)
Take the PCB board off and try running it without the insulating gasket between it and the rest of the drive
Run the drive without the back plate attached

Part 3 - So here we are. Surprisingly the disk somehow still functions even after everything I've done to it. I can still plug it in and turn on my PC and read *some* data off of it. The drive makes no noticeable noises, no click, whirring, etc. Seems to spin just fine. Once I boot up I can still mount an encrypted volume stored on the damaged device and copy *small* files off of it. However my episodes of star trek at ~120GB cause it to puke I/O errors and freeze my PC sometimes.

Questions
Would a platter swap be a good idea?
Would a PCB swap be a good idea?
Would a head swap be a good idea?
Any other clever ideas?

Wow dude =) you are harsh on ur disks =)
Most likely after this abuse the drive has multiple bads and chances are a big file of 120gb will be affected. Id say u should have tryed imaging the drive onto another one as your 1st step but it is to late for now =)) Just for fun you can still give it a go. Get ur hands on by sector imager ( like HDDCLONE for example ) and let her run =)
Report here the imaging result ( especially if there is any funny stuff like you being able to image untill certain point and then nothing at all )

PS Try MHDD as a surface scanner , see if there are any patterns to sector access.

@rkeiii:

You say you're streaming data from the disk, yet the failed commands you show are writes to the disk - so there may be something you're not telling us, as you don't need to write to a disk, to stream data from it. Perhaps your swap is also on that disk? Or ...?


This reported ICRC error suggests a possible interface issue, and not necessarily a disk problem. Further detailed analysis of those errors may give further insight. Collecting and reviewing the raw SMART data from the disk would also give more info.

Only after investigating those areas of concern above, I might consider using ddrescue to attempt to clone the disk (similar to what Alexii suggested), if you're going to take the risk of DIY.

Wild educated guess: PCB issue (well, was/still is, but with all the other tricks performed on the drive, there may be additional issues)
Pretty common issue for Tornado drives

_________________
Hard Disk Drive (HDD), Solid State Drive (SSD, SATA, NVMe, etc), USB Flash Drive and RAID Data Recovery Specialist in Massachusetts

What did you do exactly here?

Update--

So I've completed bricked the drive now. It won't even recognize at POST. I got a bit of saliva on the platters then attempted to clean it off. Apparently that was a bad call.

Also there was never a problem with the drive. My PSU was dying and undervoltoing everything in my machine. I noticed my 3GHz CPU running at 2002mhz in /proc/cpuinfo. I also started getting similar HDD errors out of my SSD. So truth be told I destroyed this poor 500gb WD disk for no good reason.

Sorry to waste everyones time... Thanks for the replys.