mongosh, connection strings and tools

conceptThe concept

mongosh is a Node.js-based shell. Everything you type is JavaScript, plus a few helpers (use, show, it). Connect with a URI:

mongosh "mongodb://app:secret@db1:27017,db2:27017,db3:27017/shop?replicaSet=rs0&authSource=admin&tls=true"
mongosh "mongodb+srv://cluster0.example.net/shop"   # hosts and options come from DNS SRV/TXT records
mongosh --host localhost --port 27017 -u admin --authenticationDatabase admin

Listing every replica-set member in the URI (or using SRV) lets the driver discover the topology and follow failovers. directConnection=true pins one host and disables discovery.

find() returns a cursor. The shell prints the first 20 documents and it fetches the next batch. In scripts you consume cursors with toArray(), forEach() or for...of.

whyWhy it matters

Most production troubleshooting happens in mongosh: checking rs.status(), reading db.currentOp(), running explain(), or fixing data with a careful update. Knowing the helpers and cursor behaviour makes you fast and safe.

Use caseReal-world scenario

On-call at 2 a.m.: mongosh into the primary, run db.currentOp({ secs_running: { $gt: 5 } }) to find a runaway query, check its planSummary, kill it with db.killOp(), then add the missing index during the day.

Try itExamples

Run and edit these in the interactive lesson.

db.getName()
db.version()
show collections

db.orders.find({}, { orderNo: 1, total: 1 }).limit(25)

const statuses = db.orders.distinct("status")
statuses.forEach(s => print(s.padEnd(10), db.orders.countDocuments({ status: s })))
db.orders.find({ total: { $gt: 3000 } }).toArray().map(o => o.orderNo)

db.hello()
db.serverStatus().opcounters

internalsUnder the hood

The shell wraps the official Node.js driver, so it follows the same server selection, retry and connection pool rules as your application. Each statement you run is a separate command; a cursor keeps a server-side cursor id open (10 minutes idle timeout by default) until it is exhausted or closed.

Avoid / preferDo this, not that

mongosh "mongodb://admin:S3cret@db1:27017/shop"

mongosh "mongodb://db1:27017,db2:27017,db3:27017/shop?replicaSet=rs0&tls=true" \
  --username admin      # prompts for the password

db.events.find().toArray()   // 40 million documents

db.events.find({ type: "error" }, { ts: 1, type: 1 }).limit(20)

PitfallsCommon mistakes

• Connecting to one replica-set member with directConnection=true in an application, which breaks automatic failover.
• Calling toArray() on a huge result in a script and running out of memory. Iterate the cursor instead.
• Running write scripts against production without first running the same filter with find() and countDocuments().

LimitsWhen not to use it

mongosh is for interactive administration and small scripts. Business logic that runs on a schedule belongs in an application with a driver, proper error handling, retries and monitoring. For bulk export and import, mongoexport, mongodump and mongorestore are faster and resumable.

SecuritySecurity notes

mongosh keeps a history file. Its redactHistory setting (default remove) drops lines that look sensitive, such as createUser calls and connection strings with credentials, but do not rely on it: avoid typing secrets, and use passwordPrompt() in commands such as db.createUser({ user: "app", pwd: passwordPrompt(), roles: [...] }).

ChallengeTry it yourself

({ version: db.version(), db: db.getName(), collections: db.getCollectionNames().length })

InterviewQuestions you should be able to answer

QuizCheck your understanding

4 questions with instant feedback in the interactive lesson; a wrong answer goes into your review deck.