08-13-2015 04:44 AM
We use for staging purpose Cloudera Enterprise Data Hub Edition Trial 5.4.3 with HBASE,HDFS,YARN and Zookeper.
There are 3 servers, one as master-namenode and two as datanodes.
We have the same environment on production site but with one master-namenode and six as datanodes.
On both environments we have the same error on the master-namenode and we lost the connection on GUI after while for a couple of seconds.
Detected pause in JVM or host machine (e.g. a stop the world GC, or JVM not scheduled): paused approximately 1781ms: GC pool 'ConcurrentMarkSweep' had collection(s): count=1 time=2116ms
I think the problem is that I have to tune Java but I don't know exactly what value I have to change.
Thanks in advance for your time.