Hi!

quoted

- We expect that the lots-of-dirty-anon-memory-over-swap-over-network
  scenario might still cause deadlocks.  

  I assert that this can be solved by putting swap on local disks.  Peter
  asserts that this isn't acceptable due to disk unreliability.  I point
  out that local disk reliability can be increased via MD, all goes quiet.

Putting swap on local disks really messes up the concept of stateless 
servers. I suppose you can do some sort of swap encryption, but
otherwise you need to scrub the swap partition on boot if you
re-purpose the hardware. You also then need to do hardware
configuration to make sure the local disks are all setup the 
same way across all server platforms so the common images can 
boot. 

We should really encrypt swap with random key generated at boot, for
all the machine. I believe it is possible (with some non-trivial
setup) today, but it would be nice to do it automagically.
								Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html