How to Build a High‑Availability MongoDB Sharded Replica Set Cluster on Ubuntu

MongoDB is a widely used NoSQL database, now ranking among the top six. This guide explains how to build a high‑availability MongoDB cluster with sharding and replica sets on Ubuntu 18.04.

Key Concepts

1. mongos – the entry point for client requests, routing queries to the appropriate shard. 2. Config server – stores metadata for routing and sharding; runs as a replica set for reliability. 3. Shard – a partition of data stored on separate machines to scale storage and load. 4. Replica set – provides redundancy for each shard, preventing data loss. 5. Arbiter – a lightweight MongoDB instance that participates in elections without storing data.

Environment Preparation

Three Ubuntu 18.04.1 LTS servers (10.20.71.135, 10.20.73.182, 10.20.70.61) with the package mongodb-org-server_4.0.6_amd64.deb are used. Directories for conf, mongos, configsvr, and each shard ( shard1, shard2, shard3) are created under /usr/local/mongodb.

Install MongoDB Source

sudo apt-key adv --keyserver hkp://keyserver.ubuntu.com:80 --recv 9DA31620334BD75D9DCB49F368818C72E52529D4

echo "deb [ arch=amd64 ] https://repo.mongodb.org/apt/ubuntu bionic/mongodb-org/4.0 multiverse" | sudo tee /etc/apt/sources.list.d/mongodb-org-4.0.list

sudo apt-get update
# sudo apt-get install -y mongodb-org

echo "mongodb-org hold" | sudo dpkg --set-selections

echo "mongodb-org-server hold" | sudo dpkg --set-selections

echo "mongodb-org-shell hold" | sudo dpkg --set-selections

echo "mongodb-org-mongos hold" | sudo dpkg --set-selections

echo "mongodb-org-tools hold" | sudo dpkg --set-selections

Configure Config Server

Create configsrv.conf:

storage:
  dbPath: /data/mongodb/config/data
  journal:
    enabled: true
systemLog:
  destination: file
  logAppend: true
  path: /data/mongodb/config/log/configsrv.log
net:
  port: 21000
  bindIp: 0.0.0.0
processManagement:
  timeZoneInfo: /usr/share/zoneinfo
security:
  keyFile: /data/mongodb/conf/key
  authorization: enabled
replication:
  replSetName: configs
sharding:
  clusterRole: "configsvr"

Service file mongod-configsrv.service (systemd):

[Unit]
Description=MongoDB Config Server
After=network.target
Documentation=https://docs.mongodb.org/manual

[Service]
User=mongodb
Group=mongodb
EnvironmentFile=-/etc/default/mongod
ExecStart=/usr/bin/mongod --config /data/mongodb/conf/configsrv.conf
PIDFile=/data/mongodb/config/log/configsrv.pid
LimitFSIZE=infinity
LimitCPU=infinity
LimitAS=infinity
LimitNOFILE=64000
LimitNPROC=64000
LimitMEMLOCK=infinity
TasksMax=infinity
TasksAccounting=false

[Install]
WantedBy=multi-user.target

Start and enable:

systemctl daemon-reload
systemctl start mongod-configsrv.service
systemctl enable mongod-configsrv.service

Initialize the config replica set:

mongo --port 21000
config = {_id: "configs", members: [{_id:0, host:"10.20.71.135:21000"},{_id:1, host:"10.20.73.182:21000"},{_id:2, host:"10.20.70.61:21000"}]}
rs.initiate(config)

Configure Shard Replica Sets

Each shard (shard1, shard2, shard3) has its own shardX.conf (replace X with 1‑3) with appropriate dbPath, port (27017, 27018, 27019), and the same security settings. Example for shard1:

storage:
  dbPath: /data/mongodb/shard1/data
  journal:
    enabled: true
systemLog:
  destination: file
  logAppend: true
  path: /data/mongodb/shard1/log/shard1.log
net:
  port: 27017
  bindIp: 0.0.0.0
processManagement:
  timeZoneInfo: /usr/share/zoneinfo
security:
  keyFile: /data/mongodb/conf/key
  authorization: enabled
replication:
  replSetName: shard1
sharding:
  clusterRole: "shardsvr"

Corresponding systemd service mongod-shard1.service (similarly for shard2 and shard3) uses --config /data/mongodb/conf/shard1.conf. Start and enable each service, then initialize each replica set:

# For shard1
mongo --port 27017
use admin
config = {_id:"shard1", members:[{_id:0, host:"10.20.71.135:27017"},{_id:1, host:"10.20.73.182:27017"},{_id:2, host:"10.20.70.61:27017", arbiterOnly:true}]}
rs.initiate(config)

Repeat with ports 27018 and 27019 for shard2 and shard3, adjusting host IPs accordingly.

Configure mongos Router

Create mongos.conf:

systemLog:
  destination: file
  logAppend: true
  path: /data/mongodb/mongos/log/mongos.log
net:
  port: 20000
  bindIp: 0.0.0.0
processManagement:
  timeZoneInfo: /usr/share/zoneinfo
security:
  keyFile: /data/mongodb/conf/key
  clusterAuthMode: keyFile
replication:
  localPingThresholdMs: 15
sharding:
  configDB: "configs/10.20.71.135:21000,10.20.73.182:21000,10.20.70.61:21000"

Systemd service mongos.service runs /usr/bin/mongos --config /data/mongodb/conf/mongos.conf. Start and enable:

systemctl daemon-reload
systemctl enable mongos.service
systemctl start mongos.service

Enable Sharding

Connect to mongos and add each shard replica set:

mongo --port 20000
use admin
db.auth("weadmin","123456")
sh.addShard("shard1/10.20.71.135:27017,10.20.73.182:27017,10.20.70.61:27017")
sh.addShard("shard2/10.20.71.135:27018,10.20.73.182:27018,10.20.70.61:27018")
sh.addShard("shard3/10.20.71.135:27019,10.20.73.182:27019,10.20.70.61:27019")
sh.status()

Enable sharding for database test and collection table1 on field id:

db.runCommand({enablesharding:"test"})
db.runCommand({shardcollection:"test.table1", key:{id:1}})

Testing

Create admin and test users, insert 5,000 documents into test.table1, then verify distribution:

# Create admin user
mongo --port 20000 -u weadmin -p 123456 --authenticationDatabase admin
use admin
db.createUser({user:"weadmin", pwd:"123456", roles:[{role:"userAdminAnyDatabase", db:"admin"},{role:"clusterManager", db:"admin"}], passwordDigestor:"server"})

# Create test user and insert data
mongo --port 20000 -u test -p 123456 --authenticationDatabase test
db.createUser({user:"test", pwd:"123456", roles:[{role:"dbOwner", db:"test"}], passwordDigestor:"server"})
for (var i=1;i<=5000;i++) { db.table1.save({id:i, name:"shuke"}); }

# Verify sharding
mongo --port 20000 -u weadmin -p 123456 --authenticationDatabase admin
use admin
db.auth("weadmin","123456")
sh.enableSharding("test")
sh.shardCollection("test.table1", {id:1})
use test
db.table1.stats()

Conclusion

Manually deploying a MongoDB sharded replica set cluster is complex; using a managed cloud service or an automation tool such as ansible-mongodb-cluster can simplify the process.