Apache Ambari is a web application that manages Big Data clusters. It allows to control the variety of metrics such as disk, network and CPU usages on hosts, number of components alive, hosts states and many other specific indicators.
If you’re asking yourself “What does Ambari mean?”, Mahadev Konar (Hortonworks Co-Founder) explained the meaning in one of interviews. “Ambari” comes from Indian folklore and means the saddle on an elephant. This is a metaphor for Big Data world where elephant is a logo of Hadoop.
So, what is Apache Ambari and how does it work? It consists of Ambari Server and Ambari Agents. When Ambari Server is installed you can deploy the cluster with help of user-friendly wizard. It will automatically download and install Ambari Agents on every host and all needed elements of Hortonworks Data Platform. After all, agents gather and constantly send host information to the server using HTTP connection.
So, let’s start.
To log into Apache Ambari start the Sandbox VM and open 127.0.0.1:8080 in your browser. You will see the login page. Then use the default credentials to login into the system: username “admin”, password “admin”.
Now you are redirected to a dashboard page. Its main responsibility is to represent cluster metrics in real-time. This page has 3 tabs:
- Metrics – the set of useful metrics
- Heatmaps – the chart for detecting weak spots of the cluster
- Config History.
Services and Components
Ambari names all frameworks and sub-systems as Services with Components:
- HDFS is a service with components DataNode, NameNode and Secondary NameNode;
- HBase is a service with components HBase Master and Region Servers;
You can see all installed infrastructure on Services page where you can manage services and install new ones. When you click on each service it opens a Summary tab with main information. You can also change service configuration in Config tab.
Let’s click on Hive on the services list on the left. Then press Service Action button on the top right and select Run Service Check and confirm the pop up (it will trigger some tests to verify the correct functionality of Hive). This process will be displayed in the background operations window. You can press OK and the service check will run in background (you will see it in the counter on the top left).
Hosts page displays all host of your cluster and provides the possibility of bulk operations (stopping/starting components on selected hosts, etc.).
Ambari also has Alerts page (as well as alerts counter near BG-operations counter on the top left). It displays various system notifications.
Ambari was created by Hortonworks and licensed by Apache. There is an analog in Cloudera Data Platform called Cloudera Manager.
Now you know what is Apache Ambari.