On 04/23/2017 05:42 PM, Andreas Klauer wrote: > On Sun, Apr 23, 2017 at 05:11:08PM +0200, Patrik Dahlström wrote: >> * I create a 5 and a 6 drive raid set and try to find an offset where >> they both carry the same raw data. With some overlays, I should be able >> to create both these raids at the same time, correct? > > Yes, two sets of overlays. > > So overlay A is your 5 disk raid5, overlay B is your 6 disk raid5, > and then you'll just have to take a stab at it with hexdump. > > So kind of like, > > hexdump --skip $((17*1024*1024*1024)) --length 4096 /dev/md42 > hexdump --skip $((17*1024*1024*1024)) --length 4096 /dev/md43 > > If that produces the same random-looking data then your reshape > might have progressed 17G-ish and you could try to use that as > a starting point for a linear device mapping that uses the > first 17G of the 6 disk raid and everything else from the 5disk. > > The further the grow processed the larger a zone of overlap there > should be since more data ends up on the additional drive so > the original representation of the same data isn't yet written > into on the old drives. > > Does that make sense? > > 5 disks: a b c d e : f g h i j : k l m n o : p q r s t : ... > 6 disks: a b c d e f : g h i j k l : m n o p q r : s t : ... > > If it stopped at "t", both have "r s t"... and that "r s t" > would be what you have to find. You have to be wary of false > matches though (such as zeroes or other common patterns of data > you might find anywhere). I just thought of something: since /dev/sdf is a completely new disk, shouldn't I be able to locate the starting point by looking at how much of the new disk is filled. It comes filled with zeros from factory, right? Best regards // Patrik -- To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
