Use Ambari to install Hadoop clusters in CentOS 6.5

Use Ambari to install Hadoop clusters in CentOS 6.5

Ensure that the network is smooth during installation!

The previous article describes how to install Ambari in CentOS 6.5 (see). I will change the Host Name of installing Ambari to ambari.

1. Open the browser first. Firefox is generally used in Linux. Enter the URL ambari: 8080. Here, ambari is the host name of lz. Change it to your own host name.

Go to the logon page. The default username and password are admin. You can modify them in the system. A prompt is displayed during the installation of ambari.

2. Enter the system and name the cluster. Select the ambari version to be installed and the system platform to use. Here, lz uses 1.3.3 and CentOS6. There is a very important thing to explain,

If version 1.3.3 is selected, JDK 6 is required. If JDK 7 is selected when installing ambari, it cannot be passed in register! If you have selected JDK 7,

Return and re-run the installation process. Use JDK 6.

3. Select the cluster node to be installed. (Lz has been installed several times, but only once), Target Hosts selects the node in the/etc/hosts file and uses the name instead of the IP address.

Select SSH Private Key. The ssh folder is a hidden attribute. Therefore, you must hide the folder when searching. If you have manually installed the ambari agent on the target node,

You can select manual registration. Otherwise, you can directly go to the next step.

4.

Ambari automatically installs the Ambari agent on the target node and registers the agent. This may be slow and may cause errors. Just try it several times. If the error persists, check the cause of the error.

5.

Select the project to be installed. We recommend that you do not select all items at one time (lz is the first installation, and all items are selected here. The options in versions earlier than 2.0 are slightly different from those on the way, where YARN is only available in Hadoop2.0,

Even so, version 2.0 is not recommended. But HDFS, MapReduce, Nagios, and Ganglia are required. The last two items are used for monitoring.

6.

The project selected in the previous step is allocated nodes.

 

7. Select Server Load balancer and clients. If you have installed Multiple Server Load balancer instances, install the client on one of the Server Load balancer instances. However, DataNode and NodeManager must be selected (the NodeManager option is not available under 2.0, but TaskTraker is available. Select it ).

8.

Custom Service. Here there will be a red label to modify, mainly to define the relevant password. After the installation is complete, continue to the next installation.

 

9.

After the system is installed, the progress bar displays 33%. After all nodes are installed, the service is automatically started and success is enabled. But here is the most error-prone place. Click message to view the log. The error cause is displayed in the log.

Here are some suggestions that may be used.

(1)

After an error occurs, check whether the yum and rpm commands are available on each node. If the lz encounters several rpm database damages, perform the following steps to solve the problem:

Run the following commands in sequence:

Rm/var/lib/rpm/_ db * Delete the old files, 4 in total
Rpm -- rebuilddb
Yum clean all clear yum Cache

(2)

Some transactions may not be completed. Run the following command:

Yum-complete-transaction to complete the transaction

Package-cleanup -- dupes Delete possible duplicate packages

Package-cleanup -- problems Delete the problematic package

 

After success, you can see the successful interface. Add New hosts option under Actions of the Hosts option to Add a New node. The process is basically the same.

Install Hadoop Cluster Monitoring Tool Ambari

Use Ambari to quickly deploy the Hadoop Big Data Environment

This article permanently updates the link address: