> Tried that too... but no luck.
> The only thing I've modified was
> "write core.mode 040755" and after some time of "printing inodes" I've
> runned xfs_repair in modify mode but nothing good happened. Almost all
> inodes have "next_unlinked" set to null and the time in 1 Jan 1970. The
> All folder has the time in 2008 but even after xfs_repair moved the
> folder in lost+found it's still empty and named 186.
> The conclusion is: all data lost? If so what is the best way to
> reformat the volume to not have the same problems again? The "no space
> left on the device" and "lose all data" problems?
> I have an UPS on that server, so the bad shutdown corruption will not
> I know all the files and folders are still there because no write
> opperations occured but I don't have a way to read them because the
> filesystem can't find the "connections"...
If you have no backup and the data is important enough for you, you can
try to use photorec from the testdisk package, foremost or magicrescue.
photorec seems to have the most formats supported out of the box,
foremost and magicrescue can be extended via recipes. Don't expect to be
able to recover file and directory names tough.
With photorec I was able to recover photos from a smartmedia card that
showed as empty in Linux while the camera showed the photos on it.