Savanna 0.3 Targets Havana Release, Supports Elastic Data Processing
We’re excited to report that Savanna 0.3 has been officially released. The OpenStack Data Processing Project now provides elastic data processing (EDP), a feature that enables creating and launching MapReduce jobs with a single click. This release also offers extended support for dashboard, storage, network configuration, and Hadoop provisioning. For a full list of features and bug fixes, please visit the Savanna Launchpad and release notes.
With Savanna, you now can:
Create and launch MapReduce jobs using REST API or UI including Pig and Hive scripts in flexible, configurable manner.
Store jobs in both Savanna’s internal database and Swift (supported as data source for MapReduce job input and output).
When it comes to Hadoop provisioning, Savanna now fully supports data locality (rack and 4-level node group awareness for both HDFS and Swift). The latest Hadoop 1.2.1 clusters, including Oozie and Hive, may now be provisioned in parallel using a vanilla plugin, while Hortonworks Data Platform (HDP) plugin can be used to support HDP 1.3.2 parallel provisioning.
Among the core Savanna features also are extended network configuration and Neutron support (including nova-network).
Savanna’s release cycle closely follows OpenStack releases and the next major version will mirror the Icehouse (six-month) cycle. Four weeks ago, Savanna was accepted into OpenStack Incubation, a step toward becoming a part of the OpenStack integrated release.
It’s been a pleasure to see Savanna’s reviewer team grow over the last three months from 14 to 24 active members from five companies--Mirantis, Red Hat, Hortonworks, Rackspace, and IBM. We’d like to thank each and every one of them for their invaluable contributions.
If you’re going to the Hong Kong Summit, please attend our talk about Savanna on Thursday, November 7, from 2:40 pm to 3:20 pm. We will review the features implemented in 0.3 version and show a live demo of executing a Hadoop job in just one click on data stored in Swift using pre-configured Hadoop cluster template. Additionally, we have four slots at the Design Summit on Friday, November 8, from 1:30 pm to 4:50 pm. Please, feel free to propose a session until October 22.