Re: sharing one r/w unix filesystem between different machines and users
On Fri, 26 Aug 2011 08:41:35 +0200
Christoph Groth <cwg@falma.de> wrote:
> "Robert Blair Mason Jr." <rbmj@verizon.net> writes:
>
> > Christoph Groth <cwg@falma.de> wrote:
> >
> > Will something like the following work? This works on any filesystem
> > using standard unix permissions (such as ext*, ufs, reiserfs*,
> > btrfs, etc.)
> >
> > Suppose your filesystem is mounted on /media/usb0.
> >
> > # chmod -R a+rwx /media/usb0
> >
> > This does, however, seem the Wrong Way To Do It, as it will not work
> > for files created after issuing the command.
>
> That's the point. If user A creates some directory, and user B (who
> has a different uid on his machine) wants to delete a file in this
> directory, user B will have first to become root and override the
> permissions. This is an unnecessary hassle and requires the user in
> question to be fairly technical. And it won't work if the user is
> unable to become root.
>
> > Unfortunately, there isn't an elegant way to do this. It seems like
> > it should be an option to mount (something like ignoreperms).
> > Looking at the internet, it appears that OSX has a mount option to
> > do this:
> >
> > http://developer.apple.com/library/mac/documentation/Darwin/Reference/ManPages/man8/mount.8.html
> > (search for 'noowners')
> >
> > This option does not appear to exist on Linux or BSD, however.
>
> Thanks, I've been looking for something like this (though an OSX
> solution doesn't help me). I really wonder why no one has ever fixed
> this issue for Linux -- probably everybody is happy simply using FAT
> and NTFS on shared removable media. And their importance is fading
> anyway.
>
I understand OSX doesn't help anyone here... perhaps someone could
request an implementation of this command in the mount source code?
It seems simple - just check if the filesystem has noowners set and if
so, instead of fetching permissions from disk, return -rwxrwxrwx. But
these things are never as simple as they seem...
--
rbmj
Reply to: