Hi Jérémy, Would you set "dfs.namenode.rpc-address" to "master:9000" in hdfs-site.xml? The NameNode RPC address is "localhost:8020" by default and that's why only the DataNode running on master is registered. DataNodes running on slave1/slave2 want to connect to "localhost:8020" and cannot find the NameNode because NameNode is not running on slave1 or slave2.
-Akira 2018年12月24日(月) 0:13 Jérémy C <[email protected]>: > > Hello everyone, > > > I installed hadoop 3.1.1 on 3 virtual machines with VMware on Ubuntu. When I > run hdfs namenode -format and start-all.sh then jps works correctly on my > master and two slaves nodes. > > However, with the command hdfs dfsadmin -report, I can see only one live data > node (I get the same result when I check on master:50070 or 8088). > > > I tried to disable firewall as follows: ufw disable but it didn't solve the > problem. The 3 machines can connect with each other (without passwd) with > ping and ssh. I also deleted the hadoop tmp folder with datanode and namenode > folders but it didn't work as well. There are also no issues shown in the log > files. > > > Do you have any solutions to get three live datanode instead of one? Thanks. > > > You will find attached my configurations files. > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
