On Wed, Nov 6, 2013 at 3:55 PM, Sergey Gerasimov
<[EMAIL PROTECTED]> wrote:
> But I still don’t understand why hadoop engine tries to connect to
> DataNodes from client(!) machine during posting jar from client machine to
> the cluster.
Only metadata traffic goes to the NN, once metadata operations resolve
the block locations after that direct streaming to the DNs begins.