On 05/02/2025 at 13:03, Raffaele Morelli wrote:
Last week we found it was in read only mode, I've stopped and tried to reassemble it with no success. dmesg recorded this error [7013959.352607] buffer_io_error: 7 callbacks suppressed [7013959.352612] Buffer I/O error on dev md126, logical block 927915504, async page read [7013959.352945] EXT4-fs (md126): unable to read superblock
No error messages from the underlying drives ?
We've found one of the drive with various damaged sectors so we removed both and created two images first ( using ddrescue -d -M -r 10 ).
Is either image complete or do both have missing blocks ?
We've set up two loopback devices (using losetup --partscan --find --show) and would like to recover as much as possible. Should I try to reassemble the raid with something like mdadm --assemble --verbose /dev/md0 --level=1 --raid-devices=2 /dev/loop18 /dev/loop19
If the RAID members were partitions you must use the partitions /dev/loopXpY, not the whole loop devices.
If either ddrescue image is complete, you can assemble the array in degraded mode from a single complete image. If both images are incomplete and the array has a valid bad block list, you can try to assemble the array from both images.
In either case, assemble the array read-only.
