worked around it for now by telling namenode to stop doing reverse dns checks in aws.
dfs.namenode.datanode.registration.ip-hostname-check=false On Sat, Jul 26, 2014 at 12:47 PM, hadoop hive <[email protected]> wrote: > Did you allowed RPC and TCP communication in you security group, which you > have added to you hosts. > > Please also check your exclude file and third point is to increase your dn > heapsize and start it. > > Thanks > > On Jul 27, 2014 1:01 AM, "Ed Sweeney" <[email protected]> wrote: >> >> All, >> >> New AWS cluster with Cloudera 4.3 RPMs. >> >> dfs.hosts contains 3 host names, they all resolve from each of the 3 >> hosts. >> >> the datanode on the same machine as the namenode starts fine (once I >> added it's longname hostname to dfs.hosts file). >> >> the 2 remote datanodes both get the error below. >> >> org.apache.hadoop.hdfs.server.protocol.DisallowedDatanodeException: >> Datanode denied communication with namenode because hostname cannot be >> resolved (ip=10.0.7.61, hostname=10.0.7.61): >> DatanodeRegistration(0.0.0.0, >> datanodeUuid=de84029d-107b-4c80-b503-c990a3621a40, >> >> It is AWS VPC so no reverse dns and I don't want to add anything to >> the /etc/hosts files - shouldn't have to since the long and short >> names all resolve properly. >> >> Seeing hostname field in the error message has the ip field, I tried >> using dfs.client.use.datanode.hostname = true but no change. >> >> Any help appreciated! >> >> -Ed
