Cluster Management with MapR 6.0
MapR Control System (MCS) is a simple and unified interface for managing ALL data for ANY infrastructure with 24/7 monitoring. Using MCS, administrators get greater control of the end-to-end lifecycle of cluster management with intuitive workflows that simplify platform configurations for business continuity and built-in security. MCS in MapR 6.0 expands on the Spyglass initiative by co-relating events with metrics and logs and providing actionable recommendations.
Monitoring Cluster Health
Volumes, Tables, and Streams: All Together
Leveraging Core Enterprise Capabilities
New and Cool Features in MCS with MapR 6.0
User-centric design: With laser focus on user experience, MCS has been designed to streamline common user workflows. All user actions are built with pre-set defaults and categorized with common templates to make these actions intuitive and repeatable with less learning.
New Dashboard with alarm co-relation: Cluster Dashboard is often the most used part of the application, and in MCS, it offers a bird's-eye view of critical cluster information that includes utilization and resource metrics, YARN statistics, node health by service and by topology, and active alarms. The active alarms are now part of the metric chart timeline that helps administrators identify the data points around critical cluster events. The details on each alarm list the entities affected, a description of the alarm, and also recommended next steps to resolve.
Template-based workflows for volume, table, and stream management: The user interface benefits from a template-based workflow design that makes it intuitive to do all common data management actions. Whether the user wants to create a volume, table, or a stream, the experience is consistent and seamless and drastically reduces the learning curve for day-to-day actions.
View capacity trends and resource usage: Backed by a scalable monitoring engine, MCS UI shows graphs for time-series metrics for the most important cluster parameters. Administrators can not only plan for overall cluster capacity but also see the top volumes used, which users/groups are using the most, and how the usage is trending per volume.
MCS also helps you track resource utilization (CPU, memory) per node and identify heavily used nodes which can affect topology mappings. All other host metrics can be visualized for trend analysis. DB and Stream operations per node also help identify hot-spots. For YARN, apart from knowing the overall resource manager metrics for container allocation, administrators can quickly see top YARN queues and find the busiest and least used queues. For services like Drill, it shows granular metrics for fine-grained analysis of query execution and resource impact.
Direct access to Service Logs: MCS helps users to troubleshoot issues by pinpointing them to the necessary logs with a single click. Depending upon whether you want to look at specific services logs or services logs on a single node, it has pre-built filters with these parameters that help reduce MTTR drastically.
MapR Command Line Tools (maprcli)
MapR Client is a tool to submit YARN applications and to run MapR CLI, hadoop fs, and hadoop mfs commands. To run MapR CLI commands on a MapR client, you need to establish an ssh session to a node in the cluster.
The MapR command line tools are ideal for:
- Power users, shell scripts
- Monitoring: status, statistics, alarms
- Data Provisioning: Volumes, Tables, Streams
MapR API (HTTP/REST)
The MapR API provides the ability to run most commands by making REST requests programmatically or in a browser.