[Users] clustering

[Andreas Granig]

Bogdan-Andrei Iancu wrote:
> yes, that's right - I will take a look to see how difficult is to change 
> t_replicate() to be able to sent to more destinations in parallel.

I'd really encourage everybody evaluating a clustered solution to try 
cacheless usrloc in combination with mysql-cluster. Scales very well, 
without the side-effect of inevitable usrloc inconsistency...

Just my 2 cents,
Andy