MongoDB初学者疑问：/data/db作用、副本集数据存储及主节点关联

Hey there! I totally get where you're coming from—when I first started with MongoDB, the whole data directory setup for standalone instances vs replica sets had me scratching my head too. Let's break down your questions one by one:

1. /data/db 是不是始终对应主节点，还是仅用于启动mongod？ #

Short answer: /data/db is just MongoDB's default data storage path for mongod processes, and it has no inherent link to being a primary node.

Here's the full breakdown:

• When you start a mongod instance without explicitly specifying a data directory via the --dbpath flag, MongoDB will automatically try to use /data/db (on Unix-like systems).
• In a replica set, every node (primary, secondary, arbiter) can have its own distinct data directory. You could set your primary's path to /data/db-primary, a secondary to /data/db-secondary, and an arbiter to /data/db-arbiter—none of these have to be /data/db unless you want them to be.
• Even if you use /data/db for one node, it doesn't make that node the primary. The primary role is determined by replica set election, not the data directory path.

2. 搭建副本集后数据存储在哪里？ #

Once you have a replica set up:

• Each data-bearing node (primary and secondaries) stores its own full copy of the dataset in its specified data directory. For example:
  • If your primary uses --dbpath /data/primary-db, all writes to the primary are stored here first.
  • Your secondary nodes (say, using --dbpath /data/secondary-db-1 and --dbpath /data/secondary-db-2) will replicate the primary's operations (via the oplog) and maintain their own copies of the data in their respective directories.
• Arbiter nodes don't store the actual dataset—they only hold replica set configuration metadata. So their data directory (if you don't skip it) will have minimal files, no actual collection data.

3. /data/db 的作用和是否必要？ #

• 作用: It's the default fallback data directory for mongod when you don't tell it where to store data. It's a convenience for quick testing or standalone setups where you don't want to mess with custom paths.
• 是否必要?: No, not at all. You can completely avoid using /data/db by always specifying a custom path with --dbpath when starting mongod. For example:

  mongod --dbpath /my/custom/mongo-data --replSet myReplicaSet
  

  The only time you need to create /data/db is if you're starting mongod without the --dbpath flag—if the directory doesn't exist, MongoDB will throw an error and fail to start.

Pro tip: In production environments, it's almost always better to use custom, descriptive data paths (like separating primary/secondary directories) to make maintenance, backups, and troubleshooting easier.

内容的提问来源于stack exchange，提问作者Amal Kostali Targhi