|
From: | Antonio Diaz Diaz |
Subject: | Re: [Bug-ddrescue] Disk blocks |
Date: | Mon, 22 Jul 2019 12:00:55 +0200 |
User-agent: | Mozilla/5.0 (X11; U; Linux i586; en-US; rv:1.9.1.19) Gecko/20110420 SeaMonkey/2.0.14 |
David Morrison wrote:
https://twocanoes.com/winclone-image-compatibility-with-512-and-4k-block/ I came across this article which describes some in-depth behaviour of disks, and how it has changed in 2015 and later Macs. The same may well apply to other operating systems. I wonder if this will affect how ddrescue operates?
Ddrescue operates with byte positions, not sector numbers. Therefore sector size can't make ddrescue operate incorrectly, but an incorrect --sector-size may result in inefficient operation or even make direct mode refuse to work.
If there is a bad spot in that 4096, presumably the read will fail every time a 512 byte sector is read or written. So if ddrescue tries to read each 512 byte sector, it will be hitting that bad spot eight times. Apart from wear on the failing mechanism, it will presumably take eight times as long. :-( Now the challenge might be to work out what the physical sector size of the disk is when it is being disguised by the operating system.
If you find that all bad-sector areas in the mapfile have sizes multiple of 4096, most probably the sector size is 4096.
Best regards, Antonio.
[Prev in Thread] | Current Thread | [Next in Thread] |