OK, I will admit up front that some of the naming in my system set up is probably strange to long-time Linux/UNIX admins, but it all is allowable according to the rules.
Recently when removing a package I no longer use it threw a couple of errors with the remount-filesystems script failing. Looking at the script, I can see why.
The script scans /etc/fstab and issues a mount for the filesystems using the first parameter in each parsed line. However, there are a number of constructs that are perfectly legal to use in /etc/fstab that do not work when issued by the mount command. In my case, here are 2 examples:
Perhaps the script can be changed to use the second parameter of each /etc/fstab line instead, as I think that will always be acceptable to a mount command. Unless anyone knows differently.
Now I think more, that will only work for the first of my “oddities”, the wild card mounting 6 drives in a mergerfs pool.
The second example, the space in the directory name, which requires the “/040” construct to work, could occur in either, or both, the mounted directory and the mount point.
And another thing to consider. You can’t re-mount “fuse” type file systems:
Mar 8 18:17:49 NethServer esmith::event[3435]: fuse: mountpoint is not empty
Mar 8 18:17:49 NethServer esmith::event[3435]: fuse: if you are sure this is safe, use the 'nonempty' mount option
Mar 8 18:17:49 NethServer esmith::event[3435]: Action: /etc/e-smith/events/fstab-update/S20remount-filesystems FAILED: 1 [3.489736]