Introduce sharding rules to MongoDB collections

build/docker/sharding/README.md

@@ -0,0 +1,90 @@
+## Components
+A MongoDB sharded cluster consists of the following components.
+1. shard: Each shard contains a subset of the sharded data.
+2. mongos: The mongos acts as a query router, providing an interface between client applications and the sharded cluster.
+3. config servers: Config servers store metadata and configuration settings for the cluster.
+
+For a production deployment, consider the following to ensure data redundancy and system availability.
+* Config Server (3 member replica set): `config1`,`config2`,`config3`
+* 3 Shards (each a 3 member replica set):
+	* `shard1-1`,`shard1-2`, `shard1-3`
+	* `shard2-1`,`shard2-2`, `shard2-3`
+	* `shard3-1`,`shard3-2`, `shard3-3`
+* 2 Mongos: `mongos1`, `mongos2`
+
+## How to deploy
+Go to the `prod` or `dev` directory and execute the following command.
+It takes 1~2 minutes to completely deploy a cluster.
+```bash
+docker-compose up -d
+```
+
+## How to examine
+You can examine the configuration and status using the following commands.
+### Config server
+```bash
+docker-compose exec config1 mongosh --port 27017
+rs.status()
+```
+
+### Shards
+```bash
+docker-compose exec shard1-1 mongosh --port 27017
+rs.status()
+```
+
+### Mongos
+```bash
+docker-compose exec mongos1 mongosh --port 27017
+sh.status()
+```
+
+## How to connect
+You can use the sharded cluster by adding `mongo-connection-uri` option in Yorkie.
+Note that the Yorkie should be on the same network as the cluster.
+For the production,
+```
+--mongo-connection-uri "mongodb://localhost:27020,localhost:27021"
+```
+For the development,
+```
+--mongo-connection-uri "mongodb://localhost:27020"
+```
+
+
+## Details
+As mongos determines which shard a document is located and routes queries, it's necessary to configure mongos to identify shards and sharding rules.
+Therefore, the following commands should be applied to the primary mongos.
+* `sh.addShard()` method adds each shard to the cluster.
+* `sh.shardCollections()` method shards each collection with the specified shard key and strategy.
+
+Two strategies are available for sharding collections.
+1. Hashed sharding: It uses a hashed index of a single field as the shard key to partition data across your sharded cluster.
+2. Range-based sharding: It can use multiple fields as the shard key and divides data into contiguous ranges determined by the shard key values.
+
+```javascript
+sh.addShard("shard-rs-1/shard1-1:27017,shard1-2:27017,shard1-3:27017")
+sh.addShard("shard-rs-2/shard2-1:27017,shard2-2:27017,shard2-3:27017")
+sh.addShard("shard-rs-3/shard3-1:27017,shard3-2:27017,shard3-3:27017")
+
+sh.enableSharding("yorkie-meta")
+sh.shardCollection("yorkie-meta.projects", { _id: "hashed" })
+sh.shardCollection("yorkie-meta.users", { username: "hashed" })
+sh.shardCollection("yorkie-meta.clients", { project_id: "hashed" })
+sh.shardCollection("yorkie-meta.documents", { project_id: "hashed" })
+sh.shardCollection("yorkie-meta.changes", { doc_id: "hashed", server_seq: 1 })
+sh.shardCollection("yorkie-meta.snapshots", { doc_id: "hashed" })
+sh.shardCollection("yorkie-meta.syncedseqs", { doc_id: "hashed" })
+```
+
+Considering the common query patterns used in Yorkie, it's possible to employ Range-based sharding using the following command.
+```javascript
+sh.shardCollection("<database>.<collection>", { <shard key field> : 1, ... } )
+```
+
+## Considerations
+
+
+## References
+* https://www.mongodb.com/docs/v6.0/tutorial/deploy-shard-cluster/
+* https://www.mongodb.com/docs/manual/core/sharded-cluster-components/

build/docker/sharding/dev/docker-compose.yml

@@ -0,0 +1,126 @@
+version: '3'
+services:
+
+  # Config Server
+  config1:
+    image: mongo:7.0.1
+    container_name: mongo-config1
+    command:
+      - /bin/sh
+      - -c
+      - | # run the init script for the primary replica after mongod has been started.
+        sh -c "sleep 5 && mongosh < /scripts/init-config1.js" &
+        mongod --port 27017 --configsvr --replSet config-rs --bind_ip_all
+    volumes:
+      - ./scripts:/scripts
+    ports:
+      - 27100:27017
+    restart: always
+    networks:
+      - sharding
+
+  # Shards
+  # Shards 1
+  shard1-1:
+    image: mongo:7.0.1
+    container_name: mongo-shard1-1
+    command:
+      - /bin/sh
+      - -c
+      - | # run the init script for the primary replica after mongod has been started.
+        sh -c "sleep 5 && mongosh < /scripts/init-shard1-1.js" &
+        mongod --port 27017 --shardsvr --replSet shard-rs-1 --bind_ip_all
+    volumes:
+      - ./scripts:/scripts
+    ports:
+      - 27110:27017
+    restart: always
+    networks:
+      - sharding
+
+  shard1-2:
+    image: mongo:7.0.1
+    container_name: mongo-shard1-2
+    command: mongod --port 27017 --shardsvr --replSet shard-rs-1 --bind_ip_all
+    volumes:
+      - ./scripts:/scripts
+    ports:
+      - 27111:27017
+    restart: always
+    networks:
+      - sharding
+
+  shard1-3:
+    image: mongo:7.0.1
+    container_name: mongo-shard1-3
+    command: mongod --port 27017 --shardsvr --replSet shard-rs-1 --bind_ip_all
+    volumes:
+      - ./scripts:/scripts
+    ports:
+      - 27112:27017
+    restart: always
+    networks:
+      - sharding
+
+  # Shards 2
+  shard2-1:
+    image: mongo:7.0.1
+    container_name: mongo-shard2-1
+    command:
+      - /bin/sh
+      - -c
+      - | # run the init script for the primary replica after mongod has been started.
+        sh -c "sleep 5 && mongosh < /scripts/init-shard2-1.js" &
+        mongod --port 27017 --shardsvr --replSet shard-rs-2 --bind_ip_all
+    volumes:
+      - ./scripts:/scripts
+    ports:
+      - 27113:27017
+    restart: always
+    networks:
+      - sharding
+
+  shard2-2:
+    image: mongo:7.0.1
+    container_name: mongo-shard2-2
+    command: mongod --port 27017 --shardsvr --replSet shard-rs-2 --bind_ip_all
+    volumes:
+      - ./scripts:/scripts
+    ports:
+      - 27114:27017
+    restart: always
+    networks:
+      - sharding
+
+  shard2-3:
+    image: mongo:7.0.1
+    container_name: mongo-shard2-3
+    command: mongod --port 27017 --shardsvr --replSet shard-rs-2 --bind_ip_all
+    volumes:
+      - ./scripts:/scripts
+    ports:
+      - 27115:27017
+    restart: always
+    networks:
+      - sharding
+
+  # Mongos
+  mongos1:
+    image: mongo:7.0.1
+    container_name: mongos1
+    command:
+      - /bin/sh
+      - -c
+      - | # run the init script for the primary replica after the config servers and the shards has been configured
+        sh -c "sleep 40 && mongosh < /scripts/init-mongos1.js" &
+        mongos --port 27017 --configdb config-rs/config1:27017 --bind_ip_all
+    ports:
+      - 27017:27017
+    restart: always
+    volumes:
+      - ./scripts:/scripts
+    networks:
+      - sharding
+
+networks:
+  sharding:

build/docker/sharding/dev/scripts/init-config1.js

@@ -0,0 +1,9 @@
+rs.initiate(
+    {
+      _id: "config-rs",
+      configsvr: true,
+      members: [
+        { _id : 0, host : "config1:27017" },
+      ]
+    }
+  )

build/docker/sharding/dev/scripts/init-mongos1.js

@@ -0,0 +1,11 @@
+sh.addShard("shard-rs-1/shard1-1:27017,shard1-2:27017,shard1-3:27017")
+sh.addShard("shard-rs-2/shard2-1:27017,shard2-2:27017,shard2-3:27017")
+
+sh.enableSharding("yorkie-meta")
+sh.shardCollection("yorkie-meta.projects", { _id: "hashed" })
+sh.shardCollection("yorkie-meta.users", { username: "hashed" })
+sh.shardCollection("yorkie-meta.clients", { project_id: "hashed" })
+sh.shardCollection("yorkie-meta.documents", { project_id: "hashed" })
+sh.shardCollection("yorkie-meta.changes", { doc_id: "hashed", server_seq: 1 })
+sh.shardCollection("yorkie-meta.snapshots", { doc_id: "hashed" })
+sh.shardCollection("yorkie-meta.syncedseqs", { doc_id: "hashed" })

build/docker/sharding/dev/scripts/init-shard1-1.js

@@ -0,0 +1,10 @@
+rs.initiate(
+    {
+      _id : "shard-rs-1",
+      members: [
+        { _id : 0, host : "shard1-1:27017" },
+        { _id : 1, host : "shard1-2:27017" },
+        { _id : 2, host : "shard1-3:27017" }
+      ]
+    }
+  )

build/docker/sharding/dev/scripts/init-shard2-1.js

@@ -0,0 +1,10 @@
+rs.initiate(
+    {
+      _id : "shard-rs-2",
+      members: [
+        { _id : 0, host : "shard2-1:27017" },
+        { _id : 1, host : "shard2-2:27017" },
+        { _id : 2, host : "shard2-3:27017" }
+      ]
+    }
+  )