Discussion Forums > Technology
Looking for hardisk format software
nstgc:
--- Quote from: Pentium100 on August 16, 2010, 07:24:55 PM ---
--- Quote from: nstgc on August 16, 2010, 07:15:06 PM ---Packing data more closely shouldn't have any effect since read write heads must become more sensitive in response to the increased density.
--- End quote ---
It does. There is a minimum noise level, just like with a cassette tape. Packing the data more closely reduces signal to noise ratio. Old hard drives (and floppy drives), recorded a strong signal and were able to read it back using simple peak detection method. The data was usually read without the use of ECC, when ECC was used it was a sign of a weak sector. Modern drives use EPRML and ECC is almost always used during during normal operation of the drive, that is - the drive can no longer reliably read what it has written "on top" without the use of ECC (which in the past was only used in case of a problem). And you want it to be able to read what it has written before.
--- Quote --- As for zero-filling that drive I was mentioning, no I didn't. Instead I filled it with all sorts of stuff and tried, fruitlessly, to recover the data after two drives in a RAID 5 died. The data should not have been there, yet the drive was almost filled with financial records.
--- End quote ---
Interesting. Maybe that's why your RAID failed - the drive could no longer record reliably (as evidenced by the previous data on it).
--- End quote ---
Still you could recover the ECC data as well.
As for what I typed in regards to my drives, that was incorrect. I had two drives drop out. The raid controller was what failed (two drives don't fail at the exact same time). What destroyed my data was my attempt to recover the data.
Pentium100:
--- Quote from: nstgc on August 17, 2010, 06:24:37 PM ---Still you could recover the ECC data as well.
...(two drives don't fail at the exact same time)...
--- End quote ---
I was saying that even when reading the data that was written last (as it is designed to do), the hard drive cannot read it back 100% correctly without applying ECC. If you wanted to read the data that was overwritten, the error rate would be too high.
And for RAID5 to fail, the second drive does not have to fail at the exact time as the first one (though it can do that, for example if the power supply fries both drives at once). If just has to fail before you replace the bad drive and rebuild the array.
nstgc:
Given that it was my RAID, and I was using it at the time, I'm certain the it was the controller because they did drop out at the same time.
You are making the assumption that FBI would be using the same equipment that we would be using. They have special hardware. They don't have the read error problems that we have.
Pentium100:
Well, I cannot say whether the FBI, CIA or any other similar organization can recover data that was overwritten once (if they can, then it shouldn't matter whether it was overwritten with zeros or random).
I also do not know how expensive is that recovery to them. For example, I read somewhere that if you shred a CD and the resulting pieces were above a certain size (a few microns) then it is possible to put it back together like a jigsaw puzzle with an electron microscope. Though unless the stored information was extremely important (let's say war plans of an enemy country or terrorist attack plans or something like that), it would be too difficult and expensive o do, therefore shredding to smaller pieces is only useful for spies and the like.
Overwriting the data in a hard drive a few times should destroy the data or at least make it to expensive to recover (probably a single pass would make it too expensive).
nstgc:
Think about reading a bit on a hard drive as checking for a certain limit. Lets say that if 75% of the particles in that bit are up, then its a 1. Writing a zero will bring that below 25%, however if you write zeros to zeros, the the zero'ed zeros will still be lower than the zeroed ones. As a result there is a contrast. Writing random bits reduces this contrast. Ultimately it would be best to write zeros to ones and level existing zeros alone, but I've never seen such a program.
Writing zeros to a drive may make it difficult for someone without special equipment to read the disk (such as the case of getting a refurbished drive), but it wouldn't effect someone who can read files that have been over written.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version