Re: [PATCH 0/5] hybrid union filesystem prototype

[Neil Brown <neilb@suse.de>]

On Thu, 2 Sep 2010 15:15:37 +0200 (CEST)
Jan Engelhardt <jengelh@medozas.de> wrote:

> 
> On Monday 2010-08-30 12:18, Miklos Szeredi wrote:
> >
> >> My comment about set-theory unions being commutative set me thinking.  I
> >> really don't think "union" is the right name for this thing.  There is
> >> nothing about it which really fits that proper definition of a union.
> >
> >We could call it overlayfs.  People learn new names quickly :)
> 
> There is a much larger issue that you should be very well aware about —
> 
> "The name wanted to be a clever acronym for "Filesystem in
> USErspace", but it turned out to be an unfortunate choice. The author
> has since vowed never to name a project after a common term, not even
> anything found more than a handful of times on Google."
> 
> overlayfs already exists. Right next to fuse on sourceforge...

lol


No, I mean it.  "Linux Over-Lays".  :-)

> 
> 
> Oh and I what I like to see is support for multiple readonly branches :)

I think we very nearly have that, assuming I understand your requirement
correctly.
The lower filesystem can itself be an overlay, providing it is mounted
read-only.

So if /mnt/ro1 /mnt/ro2 /mnt/ro3 are all read-only branches then

 mount -o ro,lowerdir=/mnt/ro1,upperdir=/mnt/ro2 meaninglessstring /mnt/ov1
 mount -o ro,lowerdir=/mnt/ov1,upperdir=/mnt/ro3 meaninglessstring /mnt/ov2
 mount -o lowerdir=/mnt/ov2,upperdir=/mnt/rw ignoreme /mnt/overlay

and /mnt/overlay will be the combination of 3 read-only filesystems and one
writable one.
(this doesn't work with the code as-is, but it is really just a few bug-fixes
away).

NeilBrown