hadoop - Adding a new Namenode to an existing HDFS cluster -
in hadoop hdfs federation latest step of adding new namenode existing hdfs cluster is:
==> refresh datanodes pickup newly added namenode running following command against datanodes in cluster:
[hdfs]$ $hadoop_prefix/bin/hdfs dfsadmin -refreshnamenodes <datanode_host_name>:<datanode_rpc_port>
witch best place execute flowing command: namenode or datanode ?
if have 1000 datanodes logical run 1ooo time ?
in namenode run command once.
$hadoop_prefix/sbin/slaves.sh hdfs dfsadmin -refreshnamenodes <datanode_host_name>:<datanode_rpc_port>
slaves.sh
script distribute command slave hosts mentioned in slaves
file (typically placed in $hadoop_conf_dir
)
Comments
Post a Comment