Re: init scripts problems ?

I tested the last patch, and it works really well.

Now, when stopping an active client (i. e. already used by an FR1 
array),  it dies gracefully,  and  raid is correctly  notified (  nda  
is  set faulty, but remains in array )
that's solve another problem, that makes raid unstoppable in certain 
conditions, surely by the fact that the there was an enbd-client process 
blocked in the kernel.
Now starting  a client  automatically  restart reconstruction.

So, from my side, evirything seems ok, as all the fonctionnalities are 
operational, Just like with 2.4  kernels.
I let you analyze if the  kernel interaction need further 
investigation... It's clearly out of my skills !

There will be interesting of testing on a real SMP machine, I'm going to 
see if i can find an old one for testing... or maybe a new bi-Xeon dual 
core for christmas !

I'm going to experiment with 2.6.14 too, to see if the patch is robust...