Application developers who are pushing the limits of Apache Drill and Apache Spark will be pleased to hear that MapR has released MapR Kubernetes Ecosystem Operator components for both Apache Spark and Apache Drill.
These Kubernetes Operators provide the following:
- Direct access to robust, exabyte scale storage from your Drill and Spark applications via MapR's high performance data access APIs.
- Elastic scaling across multiple versions of Drill and or Spark on one Kubernetes cluster.
- Security as a first class citizen via clear namespace segregation, enforcement of RBAC controls and lowest privilege containers.
- Support for stateful applications for Spark and Drill.
- Native Kubernetes resource management, CI/CD support, LCM for Spark and Drill applications.
- Bootstrap scripts for simplified install and configuration.
- Integrated shared services: Hive Metastore (running in Kubernetes or on the MapR platform) and Spark History server.
MapR began its Kubernetes journey in 2017 releasing a Persistent Application Client Container bringing performant, secure, stateful storage to Kubernetes managed applications running directly over the MapR platform. MapR will continue to advance its Kubernetes offerings, adding additional operators and storage options in the future.
Read a product overview here.
Click here to try out the operators.
Install guide here.
Release notes are here.
Thanks for reading and stay tuned for more in 2020.
Components, versions and supported infrastructure in version 1.0:
- Prebuilt containers with CentOS 7.7 as base OS image
- Manual installer: python scripts using Python version: 2.7.5+, 3.7.0+
- Kubectl version: 1.4.10
- Mapr core 6.0.1, 6.1 (5.2.2 does not support metrics)
- Compute Spaces (tenant) will include shared services: MapR Hive Metastore 2.3 and Spark History Server (ver 2.4.4)
- Kubernetes distro version support: v1.14, 1.15, 1.16
- Cloud managed service: GKE (Not Supported in MKE v1.0: Azure, AWS)
- Other container platforms: OpenShift - v4.1 (limited to Kubernetes distro v1.13)
- Support mapr-spark 2.4.4
- Support mapr-drill 1.15 & 1.16 (latest patches included)
- Requires JDK v1.8
This blog post was published January 03, 2020.