Re: [RFC][PATCH 0/9] Network receive deadlock prevention for NBD

[Peter Zijlstra <a.p.zijlstra@chello.nl>]

On Wed, 2006-08-09 at 16:54 -0700, David Miller wrote:
> From: Peter Zijlstra <a.p.zijlstra@chello.nl>
> Date: Wed, 09 Aug 2006 15:32:33 +0200
> 
> > The idea is to drop all !NFS packets (or even more specific only
> > keep those NFS packets that belong to the critical mount), and
> > everybody doing critical IO over layered networks like IPSec or
> > other tunnel constructs asks for trouble - Just DON'T do that.
> 
> People are doing I/O over IP exactly for it's ubiquity and
> flexibility.  It seems a major limitation of the design if you cancel
> out major components of this flexibility.

We're not, that was a bit of my own frustration leaking out; I think 
this whole push to IP based storage is a bit silly. I'm just not going 
to help the admin who's server just hangs because his VPN key expired.

Running critical resources remotely like this is tricky, and every 
hop/layer you put in between increases the risk of something going bad.
The only setup I think even remotely sane is a dedicated network in the
very same room - not unlike FC but cheaper (which I think is the whole
push behind this, eth is cheap)