Discover HDP 2.1: Using Apache Ambari to Manage Hadoop Clusters

Summary of the Apache Ambari webinar

Last Thursday we hosted the last of our seven Discover HDP 2.1 webinars, Using Apache Ambari to Manage Hadoop Clusters. Over 140 people attended and joined in the conversation.

The speakers gave an overview of Apache Ambari, discussed new features, and showed an end-to-end demo.

Thanks to our presenters Justin Sears (Hortonworks’ Product Marketing Manager), Jeff Sposetti (Hortonworks’ Senior Director of Product Management), and Mahadev Konar (Hortonworks’ Co-founder, Committer, and PMC Member for Apache Hadoop, Apache Ambari, and Apache Zookeeper) who presented the webinar. The speakers covered:

  • Themes Driving Apache Ambari’s continued development
    • Operate Hadoop at scale
    • Integrate with the Enterprise
    • Extend for the Ecosystem
  • New Apache Ambari features for cluster management
    • Maintenance Mode
    • Rolling Restarts
    • Ambari Blueprints
    • Ambari Stacks

If you missed the webinar, here is the complete recording:

And here is the presentation deck.

Webinar Q & A

Question Answer
Does Ambari include Nagios as one of its operational tools? Yes, using Ambari, you can optionally install Nagios as the monitoring and alerting system into the Hadoop cluster.
Where can I find sample client code that uses the Ambari Rest API? The Apache Ambari wiki page has a wealth of information about Ambari, including sample code and examples how to use the API.
Can an administrator selectively pick and choose to install the clients on the selected nodes? For example, I want to install Sqoop and Hive on an edge node but not other clients.

Yes. Ambari allows the Hadoop Operator to assign master and slave components to hosts, as well as clients. You can customize the component layout based on your cluster requirements, which can include installing an “edge node” with client libraries and configurations.

After cluster install, you can add more hosts to the cluster with the same flexibility to create an “edge node” as well.

Are metrics available via the API? Everything you saw today in the UI is requested via an API, and the documentation on the wiki explains the metrics and their relative API calls to fetch them.
Did you show both existing as well as new Apache Ambari features?

In the talk as well as the demo, we highlighted a combination of existing and new features. The demo is the shipping version of Ambari, so what we demonstrated today is available as part of Ambari 1.6.0.

The “Features + Roadmap” section of the Ambari wiki page enumerates feature lists for each release.

Is the UI in HTML5? The core front-end technology is JavaScript leveraging Ember, Bootstrap, and Handlebars for front-end framework and presentation layer. More information on the Ambari “Technology Stack” is available on the Ambari wiki page.
How can you add a new node to an existing cluster? And how does it get its configuration and service definitions?

You may add a new host quite easily. The “Add Host Wizard” will walk you through the installation process.

Ambari offers you two options to bootstrap an Ambari Agent on a host:

  1. Use SSH between the new host and Ambari Server host
  2. Pre-installed Ambari Agent on the new host.

In either case, the registration and installation process ensures that all the configuration and service definitions manifest onto the new node.

Is there a plan to have MapReduce job info displayed in addition to existing Tez DAG? Yes, we are planning to incorporate MapReduce job display information.

What’s Next?

Categorized by :
Ambari HDP 2.1 Operations & Management

Leave a Reply

Your email address will not be published. Required fields are marked *

If you have specific technical questions, please post them in the Forums

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Hortonworks Data Platform
The Hortonworks Data Platform is a 100% open source distribution of Apache Hadoop that is truly enterprise grade having been built, tested and hardened with enterprise rigor.
Get started with Sandbox
Hortonworks Sandbox is a self-contained virtual machine with Apache Hadoop pre-configured alongside a set of hands-on, step-by-step Hadoop tutorials.
Modern Data Architecture
Tackle the challenges of big data. Hadoop integrates with existing EDW, RDBMS and MPP systems to deliver lower cost, higher capacity infrastructure.