On Friday 08 July 2016 19:18:57 Chris Murphy wrote:
Jul 08 19:11:52 linux-04te systemd[1]: Starting File System Check on /dev/disk/by-uuid/df4040f2-8331-4e8b-838e-3855ebf13d5b... Jul 08 19:11:52 linux-04te systemd-fsck[830]: /sbin/fsck.xfs: XFS file system. Jul 08 19:11:52 linux-04te systemd[1]: Started File System Check on /dev/disk/by-uuid/df4040f2-8331-4e8b-838e-3855ebf13d5b. Jul 08 19:11:52 linux-04te systemd[1]: Mounting /home... Jul 08 19:11:52 linux-04te kernel: SGI XFS with ACLs, security attributes, realtime, no debug enabled Jul 08 19:11:52 linux-04te kernel: XFS (sdb4): Mounting V5 Filesystem Jul 08 19:11:53 linux-04te kernel: XFS (sdb4): Starting recovery (logdev: internal) Jul 08 19:12:03 linux-04te systemd[1]: home.mount mount process exited, code=killed status=9 Jul 08 19:12:03 linux-04te systemd[1]: Failed to mount /home. Jul 08 19:12:03 linux-04te systemd[1]: Dependency failed for Local File Systems. Jul 08 19:12:03 linux-04te systemd[1]: Dependency failed for Postfix Mail Transport Agent. Jul 08 19:12:03 linux-04te systemd[1]: Triggering OnFailure= dependencies of local-fs.target. Jul 08 19:12:03 linux-04te systemd[1]: Unit home.mount entered failed state.
The XFS /home has failed to mount, journal replay failed. There is a chance you can run xfs_repair from the emergency shell if xfsprogs is in the initrd. Otherwise you'll need to boot from alternate media. If you don't have a backup of /home, you should probably do this booted from a livecd so you can use xfs_repair -n so you can see what the complaints are, and the chance of fixing it up. You're almost certainly better off using a newer xfsprogs, e.g. from a Tumbleweed LiveCD for the actual repair, but it really depends on what the problem is. As it's a v5 file system, I personally would use the most recent xfsprogs I could get my hands on.
Jul 08 19:12:03 linux-04te kernel: XFS (sdb4): _xfs_buf_find: Block out of range: block 0x7fffffff8, EOFS 0x34f32800
Definitely do not use xfs_repair -l unless an XFS expert (not me) or developer says it's a good idea to do it, it's a big hammer.
An always valid advice is: You need a backup. If it is about a non-root
partition it might be simpler and safer to just recreate the filesystem and
the content from backup. Even if you manage to repair the filesystem some
files might be corrupted and maybe you won't notice this until your backup
also gets invalid when you overwrite the backup with your data. If you don't
have a backup, this is your lesson when you learn how they are important ;-)
For trying the recovery I can recommend to follow the approach what
professionals in the data recovery domain and even law enforcemente would do:
Copy the complete filesystem to another medium and only do recovery
experiments on that image copy. For example you can call from a live medium
```
dd_rescue