Hello.

Im not in the business, I have no commercial tools. What I do have is a private HGST Ultrastar 7K3000 HDS723030ALA640 with a weak head/preamp. Symptoms are Drive will read fine for couple gigabytes/minutes and then when it gets warm ~10% percent of files become slower and slower to read all the way down to <1KB/s. It was a secondary storage and I managed to read everything off of it no problem using patience and http://copyfile.sourceforge.net

Now I would like to somehow salvage it back to usable state. Im perfectly fine with primitive method of marking sectors under bad head in the filesystem https://github.com/jamersonpro/ntfsmarkbad.

I read http://www.hddoracle.com/viewtopic.php?f=59&t=650 and indeed I can clearly see my HDS723030ALA640 stepping thru 10 heads as stated on https://rml527.blogspot.com/2010/10/hdd ... hi-35.html (5/10)

I used Victoria 4.47 set to 32 block 50ms timeout and was able to identify bad head (7) and manually locate more or less its position:

2092600 2401000
5100000 5381600
53181000 53467000
etc

~300000 LBA per head. Sadly this method is unreliable because it relies on drive being slow at the moment of testing (random, most likely temperature related?), and time consuming (~2000 gaps to locate).

Are there any free tools/methods to do it quicker in an automated way? An MHDD/HDDSuperTool script? some way of reading this map off the drive itself instead of scanning?

If data is saved, why you want to revive a drive that will fail in the near future and risk all your data again?

_________________
http://www.pclab.com.pt facebook.com/PCLAB.A.T
ACELab partner

Its a spare drive with only ~300GB bad. Why throw it away when it can be used for less important stuff?
I wouldnt bother if it was showing growing number of random bad sectors (indicating bad surface/contamination), but its reading/writing/seeking flawlessly when I manually avoid head nr.7.

Since the drive is failing, and will likely fail completely in the somewhat near future, I am assuming you are going to use this drive for non critical data that you are fully willing to lose. So assuming that you know the drive will die, you could use hddsuperclone to get an approximate location of all of the bad sector areas. Just only run phases 1 and 2, with the skip threshold time low enough to skip when it is reading from the bad head. You would then examine the log to find the areas. I won't get into it much more than that as that is not what hddsuperclone is designed for, and you are beating a dead horse.

Also, how will you know if there are critical file system files within the bad areas or not? What if the MFT is partly inside the weak head, or any other structural metadata files? It is just a very bad idea. Good luck.

_________________
http://www.hddsuperclone.com
Home of HDDSuperClone