Question-76: How do you control that which hosts can connect to the NameNode and which cannot?
Answer: We can use the property dfs.hosts and dfs.hosts.exclude to determine which datanodes can connect to name node or which can not respectively.
Question-77: Which all properties you can set to allow and deny the nodes which are running NodeManager to connect to ResourceManager?
Answer: We can use following properties
- resourcemanager.nodes.include-path (to allow communication)
- resourcemanager.nodes.exclude-path (to not allow communication)\
Question-78: Is this required to restart the cluster once above properties are modified?
Answer: No, it is not required. You can simply invoke the refreshNodes command through dfsadmin for HDFS and rmadmin for YARN. To notify the NameNode and ResourceManager respectively.
Question-79: Which all types of hosts can be managed by Ambari?
Answer: Ambari can manage the the hosts which participates in Hadoop cluster. This hosts can only be managed for HDFS and YARN. Even for that you have to enable this feature by using the configuration property “manage.include.file” and once this is enabled Ambari would manages the files included and excluded and update their contents based on the state of the hosts and same will be visible in the Web UI of Ambari. So as soon as file changes, Ambari uses the command like refreshNodes to update the state of the DataNode, NameNode and ResourceManager.
Question-80: What all information Ambari displays?
Answer: As we know Hadoop is a cluster contains community level Hardware and number of nodes can be more than 1000 and to manage and collect the information from all the nodes, agents are installed on each node in the cluster. And information from all the node is collected by the Ambari Server and properly represented on the Ambari Web UI.
In general Ambari shows the summaries of each individual service like Spark, HDFS, HBase etc. as well as helps you in creating Alerts.