Wrong usage of lp_idmap_backend() value?

[David Collier-Brown -- Customer Engineering]

Simo Sorce wrote:
 > I would not, if nfs had decent locking (or is it decent now? any
 > opinion?) we could use the tdb backend as remote and have
 > centralised idmap at no cost.

   	In my opinion, even NFS 3/4 has sufficiently poor locking
	as to make it inadvisable to use **the locks that are
	provided with NFS**. It's not that they don't work, it's
	that it is hard to get right, and harder to know if you've
	got it right.

	This does not mean that one cannot use lockfree or 	
	latency-insensitive locking schemes.  Henry might
	be a good person to comment...

--dave
--
Those who do not understand Unix are condemned to reinvent it, poorly. 
  Henry Spencer,  <henry at spsystems.net>