On 12/07/17 13:42, Andrei Borzenkov wrote:
On Wed, Jul 12, 2017 at 3:26 PM, Istvan Gabor <suseuser04@gmail.hu> wrote:
What is the size of individual partitions that comprise md9? It is assembled from /dev/sdb9 and /dev/sdc9.
# sfdisk -l /dev/sdb Units: sectors of 1 * 512 = 512 bytes Sector size (logical/physical): 512 bytes / 512 bytes I/O size (minimum/optimal): 512 bytes / 512 bytes Disklabel type: dos Device Boot Start End Sectors Size Id Type /dev/sdb9 298005813 360916289 62910477 30G fd Linux raid
62910477 == 31455238,5KiB
...
fsck:
# fsck /dev/md9 fsck from util-linux 2.28 e2fsck 1.42.11 (09-Jul-2014) The filesystem size (according to the superblock) is 7863809 blocks The physical size of the device is 7863791 blocks Either the superblock or the partition table is likely to be corrupt! Abort<y>? yes
(I am puzzled what is the block size here)
Most likely 4K (default block size for ext*). Use "tune2fs -l" or "dumpe2fs" to verify.
7863809 * 4KiB == 31455236KiB which is 2.5KiB smaller than physical size 7863791 * 4KiB == 31455164KiB
So it looks like filesystem was created on physical device. You are lucky if you did not observe data corruption/crashes before. As to why you could mount it on earlier version - may be it did not verify it properly.
Anyway, at this point you need to reduce size of your filesystem to match actual Linux MD size.
Okay. First things first. Go read https://raid.wiki.kernel.org/index.php/Asking_for_help, and get and run lsdrv. That will be needed if anything goes wrong. I note your array is a v1.0 mirror. That means you should be able to mount the disk partition - sdb9 or sdc9 - directly. Try doing that read-only so you know you can recover your data (or, we know 12.2 can mount it ...) So an option is to delete the array on one drive, recreate it, copy the data across, and add the other drive back in ... I'm trying to get my head round the implications of the partition eating up the entire space of md9 - you're lucky not to have filled the drive and overwritten the superblock and stuff ... But I suspect that going back into 12.2, and shrinking the filesystem (MAKE SURE you shrink the md device, not the sd device), will fix the problem for leap (the v1.0 superblock is at the end of the md device, so probably the newer kernel is looking after the filesystem for the superblock, and getting confused ... Incidentally, all the defaults are making it harder and harder to make the N in mdN and sdXN the same. Probably for good reason :-) If you did accidentally create the filesystem on the sd I'm surprised 12.2 didn't pick it up but leap does, but I'm not too surprised that it worked in 12.2. Cheers, Wol -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org