Comment 1 for bug 324166

Yeah, I had the same issue with the ntfsresize that is currently in elive (debian lenny) ntfsprogs-2.0.0-1 -- I was not sure of the original disk geometry, and /fixboot was not helping, so my solution was to resize the partition again, and put it at the start of the disk.

It worked! I would speculate that the partition start was not on a cylinder boundary. No combination of chkdsk and ntfsresize --force --info would cause the fixboot task to succeed, until I tried this.

Unfortunately this is not enough. It's my friend's computer, and she really wanted to keep that recovery partition. Does anyone know more about this problem than we've exposed in this thread? It only vaguely resembles the issue from the FAQ, posted in 2004 about libparted-1.6. I was actually afraid this might happen, so I refrained from moving the starting boundary of the partition at first.

Can you tell about what you did to copy the boot sector (dd? what parameters) and where you put it so as to be sure that a) chainloader +1 will find it, and b) you're not writing somewhere that you shouldn't?