ZFS mount points

[Forrest Aldrich]

What precautions could I take to prevent users from writing to a portion 
of a ZFS mount point that might fill up the underlying filesystem that 
wouldn't impede writes to that ZFS mount point.  For example:

/foo/bar/volume

If some program accidentally started writing to /foo/bar, as far as I 
understand, it would end up filling / and not the zpool. 

I could re-do this so that /zpool1 was /foo to mitigate this - though 
I've already established production mounts pointing to /foo/bar/volume 
(which is actually zpool1/volume).  Would changing zpool1 to /dce create 
havoc, since I'm NFS exporting /foo/bar/volume.


Thanks.