On Wed, Feb 01, 2023 at 02:55:12PM +0200, Panagiotis Papadakos wrote: > Dear all, > > I am using XFS on an ICY-BOX 4-bay USB RAID enclosure which > unfortunately has been corrupted (probably due to some power-down). XFS filesystem shouldn't be corrupted due to a power failure at the first place, that's the main reason behind journaling filesystems. > > I have used xfs_repair -L, This is what I'd expect after a power failure, a dirty log which should be replayed *before* attempting to repair the filesystem, why did you discard the journal? Have you tried to mount/umount the filesystem before running xfs_repair -L? > which after a huge number of messages about > free inode references, bad hash tables, etc, fails with the following > error: > > fatal error -- couldn't map inode 13199169, err = 117 Always send the full output :) > Is there anything I can do or should I consider my data lost? The fact you discarded the journal in the first place, you've already lost some data, how much, depends on how much information was in the journal waiting to be replayed :( -- Carlos Maiolino
