[Nutch] Hadoop dynamically deletes Datanode nodes and Tasktracker nodes

In the previous post there is a description of dynamically adding a node, this time explains how to dynamically delete a node.

In the previous post there is a tutorial on how to limit the connection of a node, to dynamically delete a node, you can configure on this basis.

1. Configure the Dfs.hosts.exclude on the host1

Add Host4 to the file specified by Dfs.hosts.exclude:

Then execute the following command:

hadoop   dfsadmin   -refreshNodes

Then use the following command to view:

hadoop dfsadmin -report

As follows:

2. Web interface

As you can see from the Web management interface, there are currently only 3 nodes, and Host4 has no more:

3. Remove Host4 from the slaves file

4. Remove Host4 from dfs.hosts

5. Configure the Mapred.hosts.exclude in Host3

Add Host4 to the file specified by Mapred.hosts.exclude:

Use the following command to refresh:

-refreshNodes

As follows:

6. Remove Host4 from the slaves file

7. Remove Host4 from mapred.hosts

8. View Host4

On the host4 we can still see Tasktracker running:

Use the following command to stop actively:

workspace/hadoop-1.2.1stop tasktracker

As follows:

[Nutch] Hadoop dynamically deletes Datanode nodes and Tasktracker nodes