> Based on what you're saying, and being I'm using session (cookie) based load 
> balancing it would be true that data is rarely read or written (per user) on 
> a different server, that could be why data isn't replicating.

You've probably discovered this already but just in case, and for
others finding this ML thread: The placement of data is controlled by
your replication strategy and ring layout. Placement is not affected
by which node you happen to talk to when doing your writes or reads.
The node you talk to - the co-ordinating node - is responsible for
routing requests appropriately, and it won't e.g. store data from a
write if doesn't happen to also be part of the replica set by
co-incidence.

(For those that react: CL.ANY and hinted hand-off kind of violates
this claim as phrased, but not the spirit of it since with CL.ANY
you're making a conscious decision to complete writes prior to any
node in the replica set receiving the data, and you won't be reading
the data until it does end up in the right location.)

-- 
/ Peter Schuller

Reply via email to