I have a tough time seeing why one would need a failover sink where both
write to the same HDFS cluster.  Can one describe to me why this is a good
idea?

Reply via email to