[Unionfs] Re: rw nfs branch (Re: unionfs root on nfs)

[Hans-Peter Jansen]

Hi Junjiro, hi Martin, hi Shaya,

thank you all for the suggestions. Patching startproc wasn't a big deal 
in the end (albeit my C brain gritted..) and things are going quite 
well with unionfs-20060221-0341. Thanks god, SUSE isn't as paranoid as 
Debian ;-)

Am Sonntag, 5. März 2006 22:26 schrieb hooanon05 at yahoo.co.jp:
> Guten Morgen Herr Jansen.
>
> Hans-Peter Jansen:
> > I'm desperately seeking a (short time) fix for the /proc problem to
> > get my SuSE 9.3 diskless NFS setup going until the unionfs crew
> > come along with an "official" fix. How did you solved this in your
> > environment? Still based on this patch? The "statically linking"
> > order isn't very appealing, since I try very hard to keep the
> > distribution kernel untouched, if possible somehow.
>
> Current unionfs can handle hard link under nfs branch corrrectly, so
> we don't need the "nfs link patch" anymore, nor static link.

I struggled with this term of your previous mail:
JO>> nfs patch calls some kernel internal functions, so you need to link
JO>> unionfs statically instead of dynamic module.
but the intermediate solution of patching startproc and friends is 
working quite well. In fact, I got the unionfs based diskless client 
setup working with:
 - obligatory modification of mkinitrd (make it unionfs aware)
 - startproc patch (part of SUSE's sysvinit package)
 - minor modification of networking scripts 
   (don't shutdown unconfigured interfaces :-()

Every other modification is to cope with hardware setup and adjustments 
to local/personal preferences. Acting minimum invasive has a high 
priority for me, since I'm suffering far too 

> About /proc soft link problem, I guess Shaya Potter's mmap patch is
> the answer. But I didn't try it and it seems to be needed some
> fixing, for example lseek at readdir.

Yeah, Shaya's mmap support patch sounds very promising, and I'm going to 
try it as soon as my time window will allow more experiments. But 
pressure is high currently to deploy the stuff..

> BTW, I was glad to read your previous post and happy to know that my
> patch helped at least one unionfs user.

I'm a big fan of unionfs in general - following tightly the project 
advances and contributions since about a year, thus I'm a big fan of 
you, Shaya, Dave, Jeff and Charles. And not to forget, to whom we owe 
that project: looks like Erez is a really cool professor, who seems to 
be able to fill the gab between research and missing real world 
functionality. Thank you!

Let's hope, that unionfs gets ready to be included in the kernel in the 
next few releases. I don't know of anything not a quarter as valuable 
as this project for mainline.

Pete