bug-ddrescue
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bug-ddrescue] Feature Request: Introduce parameter to skip "slow" (


From: Burkart Lingner
Subject: Re: [Bug-ddrescue] Feature Request: Introduce parameter to skip "slow" (still readable) sectors during first pass, if reading speed is below x kByte/s
Date: Wed, 28 Sep 2011 10:29:10 +0200
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.21) Gecko/20110831 Thunderbird/3.1.13

Hi Felix!

3) Disable relocation of bad sectors?
[...]

[...] it might be quite bad if a
[temporarily] unreadable sector got replaced by
a space and therefore is not accessible anymore.
Maybe there is a way to tell the drive not to relocate sectors on error?

I have no first-hand information on how relocation works but my understanding of it is this: If a sector can't be read successfully due to ECC errors the drive returns an error and saves that information in the SMART logs. If a sector is ECC-correctable but only barely so or maybe if the sector couldn't be read previously and was internally marked as such then the drive's firmware will relocate the restored (i.e. good) data to a spare sector and internally decommission the original sector. A relocation also takes place if a known bad sector is written to.

In essence, relocations only happen if the drive firmware is sure to know what data the sector should contain. If a sector remains unreadable for all eternity and isn't written to it is also never relocated.

However, sectors can also be relocated withouth user interaction by a hard disk's offline data collection. When a hard drive is idle for a certain period of time it automatically tries to read all sectors and relocates them if the data can be read successfully but if the sector is known to be unreliable. As sectors are only relocated when read successfully I don't see anything wrong with this feature in general, but if you want to you can turn it off using i.e. smartmontools. If you do that, don't write to a damaged disk, and don't successfully(!) read bad sectors then no relocations should happen.

So if my information is correct and there are no implementation bugs then you should be happy for every relocated sector as it now holds the data the bad sector was supposed to contain.

Bye, Burkart



reply via email to

[Prev in Thread] Current Thread [Next in Thread]