You are missing "m", like 8192m, without "m" the unit is byte, so only 8 kilobytes! Also note that 4G is usually enough for DN heap. ... View more

@Michael Bronson As for automatic clean up not getting triggered, well it may be due to (or at-least related to) this unresolved bug reported on Yarn. https://issues.apache.org/jira/browse/YARN-4540 ... View more

@Michael Bronson Following API call will tell us how many DataNodes are live. # curl -H "X-Requested-By: ambari" -u admin:admin -X GET http://newhwx1.example.com:8080/api/v1/clusters/NewCluster/services/HDFS/components/DATANODE?fields=ServiceComponentInfo/started_count,ServiceComponentInfo/total_count "started_count": Shows the live (running) DataNodes. "total_count" : Shows how many DataNodes are there in the cluster. In order to find out the Individual DataNode Status you will need to supply the hostnames where the DataNodes are installed in the following API call. # curl -H "X-Requested-By: ambari" -u admin:admin -X GET http://newhwx1.example.com:8080/api/v1/clusters/NewCluster/hosts/newhwx4.example.com/host_components/DATANODE?fields=HostRoles/host_name,HostRoles/state (OR other host) # curl -H "X-Requested-By: ambari" -u admin:admin -X GET http://newhwx1.example.com:8080/api/v1/clusters/NewCluster/hosts/newhwx1.example.com/host_components/DATANODE?fields=HostRoles/host_name,HostRoles/state ... View more