CDP (Cloudera Data Platform) Private Cloud 1.2 was recently released and builds on the success of CDP Private Cloud Base (see the 7.1.6 release blog). While Private Cloud Base is the ideal modernization of both CDH and HDP deployments for traditional workloads, Private Cloud adds cloud-native capabilities. In this blog, we’ll cover the complete range of new capabilities and updates for CDP Private Cloud as a whole (the platform) as well as for both the CDW (Cloudera Data Warehouse) and CML (Cloudera Machine Learning) services.
With the complete overview further in this blog, we wanted to highlight three key aspects of the release; one each from platform, CDW, and CML.
Platform – In-place Updates
To deliver its cloud-native capabilities, CDP Private Cloud leverages Red Hat’s OpenShift. Both CDP Private Cloud, as well as the OpenShift cluster it’s deployed on, may need occasional upgrading (this new 1.2 release is a case in point. Companies can now upgrade either component, CDP Private Cloud as well as OpenShift, in place and in any order thereby reducing system downtime.
CDW – Lower minimum hardware requirements
When initially released for Private Cloud, CDW’s resource requirements were fully driven by the minimum specifications for OpenShift. Yet for organizations that only want to get their toes wet and perhaps just evaluate the capability, the 16 cores, 128 GB RAM, and 600 GB of storage prevented them from doing just that. With Private Cloud 1.2, we introduce detailed low resource requirements that reduce the amount of CPU, RAM, and storage needed by up to 75%.
CML – Applied ML Prototypes
Applied ML Prototypes, AMPs for short, provide a revolutionary new way of developing and shipping enterprise ML use cases. AMPs provide complete ML projects that can be deployed with one click directly from Cloudera Machine Learning. With AMPs, data scientists can go from an idea to a fully working ML use case in a fraction of the time, with an end-to-end framework for building, deploying, and monitoring business-ready ML applications instantly. To find out more about AMPs, read this blog, which also links to the full documentation and a webinar on the subject.
Private Cloud 1.2 delivers more than the highlights above. Here’s a list of all its new capabilities and updates:
- Support for OpenShift 4.6:
- Customers can now deploy CDP on the “Extended Support Update” version of OpenShift, reducing the frequency of required OpenShift upgrades and, with that, benefit from lower operational impact.
- Resource Groups:
- The same roles can now be assigned to groups of individuals for simplified permission management.
- Custom External Alerts:
- CDP Private Cloud can now be connected to an enterprise alerts system of choice for centralized notification of any health-related issues.
- Custom TLS Certificates:
- Highly security-conscious organizations (e.g. financial services sector) can now provide their own TLS certificates for added security.
- Customizing Diagnostic Bundles:
- Diagnostic bundles can be zoomed in on relevant events through filters such as time period, environment, workload type, and more.
- Custom Principal Support for Hive/Impala:
- Hive, Impala, and other service names can now be changed to support the specific security requirements of customers in regulated industries. In previous versions, performing such a change would interfere with Kerberos that expected a hard-coded service name.
- CDW Impala support for ORC ACID reads:
- CDW Impala users can now run SQL queries on ORC ACID data, adding to the engine-supported data sources.
- Hive ACID Compaction Fixes:
- Multiple stability fixes were made for Hive ACID Compaction, delivering a more reliable service.
- Improved Air Gapped Deployments with support for Non-Transparent Proxy:
- Customers with Non-Transparent Proxy configured in OpenShift will have those settings propagated through to ML Workspace workloads (sessions, models, jobs), ensuring greater consistency.
- Improved NFS Support:
- Sharing project files is now supported more securely through NFS versions v3 and v4.x with additional external NFS security improvements.
- Custom Kerberos Principal Support (Tech Preview):
- Custom service principals for Hive, Ozone, and other services are now supported, meeting the specific security requirements of customers in regulated industries. MLGov also supports specifying a custom service principal during workspace creation.
- API v2 (Tech Preview):
- An early version of an updated public-facing API is available for Tech Preview, covering initial APIs regarding Projects and Jobs and adding additional automation and scripting capabilities for CML.
- Grafana Monitoring:
- CML monitoring is now included in the CDP centralized Grafana dashboard, delivering a one-stop shop for all CDP operational insight. Additional database metrics were added and alerts were improved.
Beyond PVC 1.2
With CDP Private Cloud 1.2 now GA, we’re already hard at work on future releases. Later this year, you can expect the long-anticipated Cloudera Data Engineering to be added to the list of services, delivering Spark on Kubernetes. Stay tuned for further announcements and communication this summer.
CDP Private Cloud Plus customers can access 1.2 directly from Cloudera Manager; to upgrade your subscription please contact your Cloudera account manager. For more information, please also read the CDP Private Cloud 1.2 documentation. If this release overview piqued your interest, you can trial CDP Private Cloud free for 60 days.