Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
MapReduce >> mail # user >> Node failure


What's the expected behaviour on the client side when the node fails with
replication of 3 and minimum replication of 2? Is the expected behaviour
that client takes the bad node out of the list before sending blocks to
that node? Could someone help me understand how this works during a node
failure?