[sf-lug] Linux mounts ... and overmounts and ...

[Rick Moen]

Quoting Michael Paoli (Michael.Paoli at cal.berkeley.edu):

> That's certainly an approach.
> 
> In the land of Linux, I like to not only look at the output of
> mount, but also consult /proc/mounts - as that's the kernel's idea
> of what's mounted where, whereas the output of mount is a bit
> more human friendly ... except when it isn't - like in some
> cases/version and when /etc/mtab is stale - e.g. root filesystem
> is mounted read-only.
> 
> Also, for most *nix filesystems, the inode number of the root directory
> of a filesystem, as seen from within that filesystem, is 2.
> So, e.g. ...
> $ (for mp in $(df -T -x tmpfs -x devtmpfs -x iso9660 -x fuse.zfs |
> awk '{if(NR!=1)print $7;}'); do cd "$mp" && ls -id .; done) | sort |
> uniq -c
>      18 2 .
> $
> 
> Now ... Linux bind mounts ... very handy ... but still need a better
> way to determine/show with existing bind mounts, exactly what's
> mounted on what.
> Better way might exist, but don't think I've found a good/great answer
> on that one yet.

Isn't it a whole lot simpler if wondering where the contents of the /usr
filesystem went to just be able to type 'ls' and get the answer returned
to you in capital letters?  Works fine, even when hitting an ENOCOFFEE
error.