An Ambari Stack for MongoDB Cluster  

Ambari stack for easily installing and managing MongoDB on HDP cluster with any kind of cluster architecture.

 Author: Anderson Santos

_{Built with ❤︎ by Anderson Santos and contributors}

Table of contents

• Features
  • Some configurations
  • Features in development stage
  • Future Features and development
• MongoDB Replicaset Cluster Architecture
• MongoDB Sharding Cluster Architecture
• Assumptions
• Getting Started
  • Setup
  • Default ports
  • What must be done carefully
  • Managing by rest
• Removing the service
• Running the tests
  • Last Test Coverage Report
• Authors
• License
• Acknowledgments

Features

• Auto install mongodb in the following ways:
  • As a sharding cluster
  • As a replicaset
  • As a standalone deploy
• Auto install the following components;
  • Mongod server instances
  • Mongo config server instances
  • Mongos (Query router) server instances
  • A client wrapper to be used to connect to the cluster
• Supports executing more than one instance of the same server type per host
  • This is useful when the number of hosts are limited
  • In order to do that, you must specify more than one port in the port list configuration for that instance type
• Supports a very flexible port configuration string
  • Eg.: 27019,27020,27025-27027
• Shard architecture configuration:
  • Configure which instances will be in which shard
  • Configure which instances will act as an arbiter
• Automatically configure the shards adding the nodes to it
  • Automatically detect the primary host to do it
• Automatically add the shard to the mongos (Query Router) shard list
• Support cluster scaling:
  • Adding new shards to the cluster
  • Adding new instances to the shard
  • Adding new instances to the shard that will execute in the same host
• It has a service advisor that will warn if the cluster architecture is not recommendable
  • It also validates the configuration
  • The validation list is here
• A client script that automatically detects the mongo server that is online and must be used
  • It automatically detects the hostname and port of the available servers
  • It considers the architecture before deciding which server
  • It checks if the server is on before trying to connect
  • It can be called using mogok
  • You can also use any mongo paramer such as: mogok --user admin --eval 'rs.status()'
• Testing
  • An test wrapper that execute integrated tests at various linux distros using docker
  • An docker instance to setup an testing environment and execute the integrated tests for CentOS 7 distro
• Well documented and tested

Some configurations

Features in development stage

• Add logOpSize config option
• An alert when the cluster is up but the shards are not 100% ok
  • If there are missing nodes on it
• An alert when the shards are not in the mongos (Query Router) shard list
• An command to re-configure the cluster:
  • Adding the missing nodes to the shard
  • Adding the missing shards to the mongos list

Future Features and development

We are needing help for these features. We have already developed some draft scripts and solutions for it.

• Testing
  • The tests for the service_advisor script
• Few modifications to support other linux distros (including testing)
• A metrics monitor to send several metrics from the mongo instances to ambari
  • And the ambari metrics screen :)
• Useful Tasks:
  • A command to delete nodes and shards
  • A command to change the shard order
  • A command to backup the databases
  • A command to restore the databases

MongoDB Replicaset Cluster Architecture

MongoDB Sharding Cluster Architecture

Assumptions

• Ambari is installed and running.
• No previous installations of Mongo DB exist. If there any, you can either remove it or rename it.

Getting Started

Setup

Follow given step to install and manage Mongo DB using Ambari.

1. Connect to the VM via SSH (password hadoop for sandbox image) and start Ambari server

ssh root@ambari.machine

2. To deploy the Mongo DB, run below

VERSION=`hdp-select status hadoop-client | sed 's/hadoop-client - \([0-9]\.[0-9]\).*/\1/'`
sudo git clone https://github.com/supergarotinho/ambari-mongodb /var/lib/ambari-server/resources/stacks/HDP/$VERSION/services/MongoDB

3. Restart ambari server

#sandbox
service ambari restart

#non sandbox
sudo service ambari-server restart

4. Then you can click on 'Add Service' from the 'Actions' dropdown menu in the bottom left of the Ambari dashboard:

5. On bottom left -> Actions -> Add service -> check MongoDB -> Next -> Next -> Next -> Deploy

• The images of the installation sequence are here

6. On successful deployment you will see the MongoDB as part of Ambari stack and will be able to start/stop the service from here:

Important Notes:

• It is recommended to start all nodes of the same component type together
• If you are not going to restart all components at the same time, use the following recommended order:
  1. Start the mongo config servers first
  2. The start the mongos (Query Route) servers
  3. Start the mongod instances

Default ports

• mongos port 27017
• mongo config port 27019
• mongo replica port 27018

What must be done carefully

You must consider to backup the databases and understand how the database and log names are chosen by the script

• Change the sharding order
  • The shards have numbered names, removing a shard must be done carefully as the shard names are used to locate the node database and logs
• Removing a shard
  • The shards have numbered names, removing a shard must be done carefully as the shard names are used to locate the node database and logs
  • If are doing this, it is advisable to delete or move the database files and logs of the removed shard
• Removing a node
  • It is advisable to remove or move the database files and logs of the removed node

Managing by rest

• One benefit to wrapping the component in Ambari service is that you can now monitor/manage this service remotely via REST API

export SERVICE=MongoDB
export PASSWORD=admin
export AMBARI_HOST="your_ambari_hostname"
export CLUSTER="your_ambari_cluster_name"

#get service status
curl -u admin:$PASSWORD -i -H 'X-Requested-By: ambari' -X GET http://$AMBARI_HOST:8080/api/v1/clusters/$CLUSTER/services/$SERVICE

#start service
curl -u admin:$PASSWORD -i -H 'X-Requested-By: ambari' -X PUT -d '{"RequestInfo": {"context" :"Start $SERVICE via REST"}, "Body": {"ServiceInfo": {"state": "STARTED"}}}' http://$AMBARI_HOST:8080/api/v1/clusters/$CLUSTER/services/$SERVICE

#stop service
curl -u admin:$PASSWORD -i -H 'X-Requested-By: ambari' -X PUT -d '{"RequestInfo": {"context" :"Stop $SERVICE via REST"}, "Body": {"ServiceInfo": {"state": "INSTALLED"}}}' http://$AMBARI_HOST:8080/api/v1/clusters/$CLUSTER/services/$SERVICE

Removing the service

• To remove the MongoDB:

  • Stop the service via Ambari

  • Delete the service

    curl -u admin:admin -i -H 'X-Requested-By: ambari' -X DELETE http://replace_with_your_ambari_hostname.com:8080/api/v1/clusters/ambari_cluster_name/services/MongoDB

  • Remove artifacts

    VERSION=`hdp-select status hadoop-client | sed 's/hadoop-client - \([0-9]\.[0-9]\).*/\1/'`
    rm -rf /var/lib/ambari-server/resources/stacks/HDP/$VERSION/services/MongoDB

  • Restart Ambari

    #sandbox
    service ambari restart
    
    #non sandbox
    sudo service ambari-server restart

Running the tests

package/test/runAllTests.sh

Last Test Coverage Report

Authors

• Anderson Santos - Initial work - supergarotinho

See also the list of contributors who participated in this project.

License

This project is licensed under the BSD-3 License - see the LICENSE.md file for details

Acknowledgments

• This is a 99% rewritten version of the "geniuszhe" version: https://github.com/cas-bigdatalab/ambari-mongodb-cluster