We have a 10 Node cluster . Master 8Gig Ram ,Slaves (9) 4Gig Ram each .
Yesterday we were copying data of size 110GB into the cluster using
"copyFromLocal" command from the Master server. (Master node is not being
used as a datanode). During this process, datanodes are frequently loosing
connection and this is generating "unreachable node" exceptions in the log.
This was happening quite frequently for many nodes one after the other.
Any parameters I must tune to remove this ? Any suggestions are highly
Thanks and Regards
4th Year undergraduate