Join us in this 4-part series with the core committers of the Apache Hadoop projects (Pig, Zookeeper and YARN) and Hadoop experts to gain insight into current advances in Apache Hadoop, obtain use-cases and best practices on how to get started with Hadoop and live Q&A with the people at the center of the Hadoop movement.
Webinar 2: Deployment and Management of Hadoop Clusters with Ambari
Speaker: Matt Foley, Committer and PMC Member of the Apache Hadoop Project and Member of the Technical Staff @ Hortonworks
Deploying, configuring, and managing large Hadoop and HBase clusters can be quite complex. Just upgrading one Hadoop component on a 2000-node cluster can take a lot of time and expertise, and there have been few tools specialized for Hadoop cluster administrators. AMBARI is an Apache incubator project to deliver Monitoring and Management functionality for Hadoop clusters. This paper presents the AMBARI tools for cluster management, specifically: Cluster pre-configuration and validation; Hadoop software deployment, installation, and smoketest; Hadoop configuration and re-config; and a basic set of management ops including start/stop service, add/remove node, etc. In providing these capabilities, AMBARI seeks to integrate with (rather than replace) existing open-source packaging and deployment technology available in most data centers, such as Puppet and Chef, Yum, Apt, and Zypper.