MongoDBClusterDeploy

Introduction

A deploy tool written in Python, can deploy mongoDB cluster simply. Easy to use and easy to manage your cluster.

Environment

This toolset may only deploy MongoDB cluster on linux machines, it based on python3, you may install dependencies of python:

sudo apt-get install python3-setuptools python3-pip

and install needed packages:

python3 -m pip install django paramiko psutil

And you should allow the port we use (you can indicate it by yourself and default is 10086) can pass the firewall.

Deploy of manage tools

For managing process, you may deploy process manage utility first, and modify the code in ProcessManager/AutoDeploy.py to deploy manage tool on your cluster.

if __name__ == "__main__":
    IP_list = ["172.16.9.81", "172.16.9.86", "172.16.9.87"]
    for IP in IP_list:
        deploy_manage(IP, your_port, "root", "your_root_passwd") # I have save passwd on my cluster so you may

The explaination of function deploy_manage：

deploy_manage(Address of machine, manage port, ssh user,ssh passwd)

The deploy tool will auto-test the service after deploy it, if your machine start service slow than you image, the test will failed, but you can verify it manually: http://Address of machine:manage port/monitor/cpu

Copy MongoDB executable file

You may download your binaries on https://www.mongodb.com/, amd open MongoDBClusterDeploy/setup_mongodb.py modify:

local_filepath = "mongodb-linux-x86_64-ubuntu1604-3.4.6/bin" # where your executable files located in
remote_location = "/usr/local/mongodb" # fix to you binaries where will located in your remote machine

And openMongoDBClusterDeploy/config/machine.py modify the variable machine_list The explain of each unit：

"m1": { # the name of the machine, should be unique
    "ip": "172.16.9.81", # Address
    "port": 22, # SSH port
    "user": "root", # SSH user
    "passwd": "txj@618", # SSH password
    "trojan":{
        "port":10086, # manager port
        "passwd":"b115Admin" # manager password, you may modify it in MongoDBClusterDeploy/ProcessManager/ProcessTrojan/ManageAPI/views.py key_list variable
    }
},

Then run MongoDBClusterDeploy/setup_mongodb.py to upload your files.

Config and start your cluster

Open MongoDBClusterDeploy/config/mongo.py some variable you may modify:

mongo_path = "/usr/local/mongodb/" # where your mongodb binaried located in
dbfile_path = "/media/data/mongodb" # where your data to storage
replset_oplog_size = 1024  # OPLOG size, megabyte
shard_service_pre_machine = [1, 1]  # The number of shards in eache repl set, each machine

To ensure CLEANING_MODE = True in MongoDBClusterDeploy/remote_runner.py if it is your first time to config your cluster。 Then run MongoDBClusterDeploy/remote_runner.py to config your cluster.

The start and shutdown of MongoDB cluster

Use db.ShutdownServer() to shutdown your cluster. If your're starting up cluster, modify CLEANING_MODE = False in MongoDBClusterDeploy/remote_runner.py and ensure the configures as same as you deploying. Then run MongoDBClusterDeploy/remote_runner.py to config your cluster.

Monitoring of MongoDB

TODO, run django project in ClusterMonitor can monitor mongoDB stat.