Cloudera Manager 4.0: Customer Feedback and Adoption
For those new to it, Cloudera Manager is the first and market-leading management platform for CDH (Cloudera’s Distribution Including Apache Hadoop). Enterprise customers are coming to expect an end-to-end tool that manages the entire lifecycle of their Hadoop operations. In fact, in a recent Cloudera customer survey, an overwhelming 95% emphasized the need for this approach.
Cloudera Manager sets the standard for enterprise deployment by delivering granular visibility into and control over every part of CDH – empowering operators to improve cluster performance, enhance quality of service, increase compliance and reduce administrative costs. We have also a FREE edition to get started, so try it out today! (BTW, for more information on this subject, you can attend a free Webinar on Wednesday, Sept. 19, on the topic “How CBS Interactive Uses Cloudera Manager to Effectively Manage Their Hadoop Cluster”.)
Cloudera Manager Service Dashboard
The adoption of CM4 has been fantastic! A good percentage of our existing customers have deployed CM4 and are managing their production/development/test clusters using this release. In fact, one customer is using CM4 to collectively manage greater than 1,000 Hadoop nodes. This pace of adoption and use by major Enterprise customers is a great testament of the resiliency and enterprise readiness of the product offering.
Among the many conversations with customers, we’ve found that several features are particularly appreciated. I wanted to highlight some of them:
- Managing Multiple CDH Versions. CM4 supports both CDH3 and CDH4, which conveniently allows customers to adopt the latest version of Cloudera Manager (and derive all the benefits from new features) without necessarily upgrading to the latest release of the platform (CDH). Of course, we always encourage customers to upgrade to the latest and greatest version of CDH but also fully understand the operational challenges and time frames involved in platform upgrades. In the (near) future, Cloudera will solve these upgrade challenges for customers.
- Cloudera Manager API. For most enterprises, Hadoop Operations doesn’t live in isolation – they need to tie into existing IT operations and workflows. CM4 provides a rich set of APIs to make this integration possible. Through a RESTful API, customers now have access to all the content in Cloudera Manager that they can integrate with the existing IT management and monitoring tools. To help you get started, we have written a simple integration to Nagios using the CM API. (Feedback welcome!) Several of our customers have also successfully integrated the CM4 API with tools like Zenoss, BMC, Nagio, and so on. If there are plugins that you have written to integrate CM with such tools and would like to share with the broader community, please let us know.
- Client Configurations on Gateway Nodes. Most Hadoop clusters have certain nodes dedicated as “edge/bastion/gateways”, primarily for end users of the clusters to be able to log in and submit their jobs. For security reasons, end users are typically allowed access to these edge nodes only. CM4 makes it easy to have the client configuration to be synced to these nodes on any configuration changes to the services.
The entire list of features in CM4 is available here.
Since the GA release in CM4 in June 2012, there have been a few update releases (CM4.0.2, CM4.03, and CM4.04). We are now working toward delivering CM4.1 in the next few months. This release will include support for the forthcoming CDH4.1 and also include a few new features like Zookeeper monitoring, maintenance mode (which several CM users/Ops folks have asked for), host-level decommissioning, and several usability enhancements. The next major release is Cloudera Manager 4.5, currently scheduled for early 2013.
As always, I want to thank our customers for their constant support and feedback. Please keep it coming!