title: What’s new in Hue 2.2 author: admin type: post date: 2013-02-15T05:00:00+00:00 url: /whats-new-in-hue-2-2/ tumblr_gethue_permalink:
This post is about the new release of Hue, an open source web-based interface that makes Apache Hadoop easier to use, that’s included in CDH4.2.
Hue lets you interact with Hadoop services from within your browser without having to go to a command-line interface. It features a file browser for HDFS, an Apache Oozie Application for creating workflows of data processing jobs, a job designer/browser for MapReduce, Apache Hive and Cloudera Impala query editors, a Shell, and a collection of Hadoop APIs.
The goal of this release was to add a set of new features and improve the user experience. Read on for a list of the major changes (from 304 commits).
With the Oozie Application you can chain jobs and schedule them repeatedly without having to write XML anymore. Workflow and Coordinator management got extra focus and now matches all the Oozie functionalities:
Drag & Drop Workflow Editor
Workflow Dashboard
Coordinator Wizard
Rerun a Workflow (left) or a Coordinator (right)
A number of user experience improvements make it simpler to query your data with SQL:
Query Editor
Wide result page with column lookup
Impala can now be queried from a new interface. More features will be supported when Impala is GA.
Cloudera Impala query
FileBrowser lets you navigate and manage HDFS files in a UI. Its front end was totally redesigned and new filesystem operations were added. You do not need to use the hadoop fs command anymore!
Bulk editing
JobBrowser lists MapReduce jobs with their statistics and statuses. It was prettified and now supports MR2 jobs running over YARN:
MR1/MR2 job list and direct log access
Groups and Hue permissions can be assigned to the users through the UserAdmin application. Access to Hue applications can be customized for each user. (For example, Bob can only see and use the Oozie and Impala applications.) The application has been restyled and simplified, and is no longer accessible by default to non-superuser:
Group permission editing
Desktop is the core library of Hue and every application is built on top of it. In this release, the user experience has been improved with more informative errors (now with stack traces and line numbers) and new status messages (such as when critical services like Oozie are down).
On the technical side, users can now upload files to a federated cluster, some XSS vulnerabilities were fixed, and database greenlet support was introduced for more performance. Hue now fully supports transactional databases like MySQL MyISAM and PostgreSQL.
Hue is also internationalized and available in English, French, German, Spanish, Portuguese (and Brazilian Portuguese), Korean, Japanese and simplified Chinese.
With this 2.2 release, a big part of the Hadoop user experience gap was filled in.
The next 2.3 release will target users who wish to better leverage the multiple query solutions in CDH (Beeswax/Hive, Impala, and Pig). A new document model (HUE-950) would make each query (e.g. Hive query) searchable and shareable with your colleagues or importable into an Oozie workflow without any duplication. A trash and source control versioning system (HUE-951) is also discussed as well as Oozie bundle (HUE-869) integrations.
Many past feature and bugs were discussed on the hue-user list, so feel free to chime in! A Hue meetup group was created and it would be a pleasure to meet you in person and see how analyzing your data with Hadoop could be made easier.