Web UIs of HDFS components - E-MapReduce - Alibaba Cloud Documentation Center

This topic describes how to access the web user interfaces (UIs) of Hadoop Distributed File System (HDFS) components. This topic also describes the endpoint of a NameNode, the homepage of the NameNode UI, and how to start the NameNode UI.

Access the web UIs of HDFS components

You can access the web UIs of HDFS components by using an SSH tunnel or the E-MapReduce (EMR) console. For more information, see Create an SSH tunnel to access web UIs of open source components and Access the web UIs of open source components.

NameNode endpoint


Version	Endpoint	Description
Hadoop 3.x	http://${namenode_hostname}:9870	Replace the value of the ${namenode_hostname} parameter with the hostname of the corresponding NameNode.
Hadoop 2.x	http://${namenode_hostname}:50070	Replace the value of the ${namenode_hostname} parameter with the hostname of the corresponding NameNode.

NameNode UI homepage

Overview

In the first line of the Overview page, the value after Overview indicates the hostname of the NameNode. A value of active or standby in parentheses indicates that the NameNode is the active or standby NameNode in a high availability (HA) cluster. The following table describes the other parameters.


Parameter	Description
Namespace	The address of the namespace to which the NameNode belongs.
NameNode ID	The service ID of the NameNode.
Started	The start time of the NameNode.
Version	The version number of HDFS.
Compiled	The related information that is compiled.
Cluster ID	The ID of the HDFS cluster.
Block Pool ID	The ID of the HDFS block pool.

Summary

Description of some key information in the preceding figure:

Security: indicates whether Kerberos is enabled for the cluster.
Safemode: indicates whether the cluster is in safe mode and read-only.
The numbers of files, directories, and data blocks that are calculated by the active NameNode and standby NameNode may be different. This is a normal occurrence. The numbers that are calculated by the active NameNode prevail.

NameNode Journal Status

In an HA cluster in which two NameNodes are deployed, you must pay attention to the following issues:

Active NameNode in write mode
You must pay attention to the value of Written txid of each JournalNode when you view the Quorum Journal Manager (QJM) state of the active NameNode. If the ID is always 0 or a high value for an extended period of time, the JournalNode needs to be reset.
Standby NameNode in read mode
The QJM state of a standby NameNode is open for read. In most cases, you do not need to pay attention to the state information.

NameNode Storage

The checkpoint operations on FsImage files of NameNodes are performed asynchronously. Therefore, you must pay attention to the state of the image type. For example, if the image type of the standby NameNode is in the Failed state for an extended period of time, the EditLog files are overstocked. If the NameNode is restarted, it takes an extended period of time to read the EditLog files.

How to start the NameNode UI

To start a NameNode, perform the following steps: Load the FsImage files, load the EditLog files, and wait for reported blocks. In most cases, if the operations to load the FsImage files and wait for reported blocks take an extended period of time, you can check whether to adjust the garbage collection (GC) parameters. If the operation to load the EditLog files takes an extended period of time, you can check whether a large number of EditLog files are overstocked on the corresponding JournalNode. For information about the directory that is used by a JournalNode, see Local disk directories used by HDFS components.