GitHub Repository Index

Description

The Elasticsearch Mesos framework allows to deploy Elasticsearch on a large Mesos cluster which runs on any kind of hardware.

The framework helps to automate installation, running and scaling of the system. The user interface shows the the cluster state and performance overview.

Elasticsearch is part of the ELK stack

Features Documentation

STEP 1.Mesos cluster with Marathon and Chronos frameworks already running.

STEP 2a.ES framework scheduler is deployed to the master node with a configured number (3) of slaves to be launched.

STEP 2b.Scheduler receives resource offers from Mesos.

STEP 3.Once ES scheduler accepts the offers, it deploys executors together with ES nodes in containers to Mesos slaves.

STEP 4.ES nodes discovery. Zookeeper plugins in ES nodes registers with Zookeeper. This allows the nodes to find each other.

STEP 5.The ES nodes found each other and form a cluster.

    features

    Cloud

    Runs on ​_any_​ hardware, cloud or on-premises (* minimum RAM requirements for ES process).

    Resilence

    Restarts failed tasks automatically.

    Discovery

    Discovery via Elasticsearch Zookeeper plugin.

    Scaling

    The framework can be scaled up and down horizontally.

    Runtime

    Run the framework as Java processes or Docker containers.

    User Interface

    Web user interface on scheduler port 31100.

    api

    Access information of the UI via the API.

    Written in java

    The framework is implemented written in Java 8.

    Configurable

    Cluster settings such as dimensions, cluster name and Elasticsearch settings are configurable.

    Tested

    The major features of the framework are tested by minimesos system tests.

    Free

    The framework is open source and free to use.

      Download

      VERSION 0.6.0 released 9.11.2015gi

      Roadmap