So, I'm stumped on this one. I have functionally two volume groups. vgold and vgnew, and I migrated the lvms from the old one to the new one (via rsync of the files to an external drive first, not that it matters). After installing the new disk, putting the files in place and (clearly) rebooting:
# mount /dev/vgnew/foo /foo [ no errors returned ] # ls /foo [ empty directory still ] # mount | grep foo [ nothing ] # mount /dev/vgnew/foo /bar [ no errors returned ] # ls /bar [ all the files ]
/foo is the previous point point under the old volume group, but it doesn't matter I use to attempt to mount on /foo... Nothing will mount there at all.
strace shows that
mount() is getting called and returns 0.
/etc/mtab of course doesn't contain anything (being a symlink to kernel's /proc/self/mounts).
This is on a fedora 21 box, FYI. And I have about 10 mount points that are all acting the same way, at different depths of the file system.
------- UPDATE -------
A coworker found a archlinux page where someone had the same problem. Unfortunately, they have no solution there other than to rename things. But they did provide one clue that turned out to be true:
mount /foo && df
shows the mount very very quickly showing up and then going away.
------- UPDATE 2 -------
This stackexchange problem actually comes closest. It shows that my fstab entries (with noauto as the option) is the culprit:
/dev/vgfoo/foo /foo ext4 noauto 1 2
Aftering commenting that out and then doing a
systemctl daemon-reload, it actually mounts. But... I still want that line in the fstab so I mount it when I need it (it's on a disk that isn't always online).
systemd strikes again...