Install MongoDB and mongosh or Provision a Free Learning Environment

Learning outcomes

The first two lessons established the product and responsibility model. Now the goal is reproducibility: a second learner should be able to build the same Chapter 01 environment, identify exactly what was installed, connect with mongosh, and remove the lab without leaving a mysterious service or dataset behind.

1. Pick the environment from the evidence you need

A native installation is best when you need to learn Windows services, Linux package paths, macOS service managers, real configuration-file locations, OS permissions, or host filesystems. A Docker container is best when you need a disposable, pinned server process with isolated data and a predictable port. Atlas Free is useful when you need to practice a managed connection workflow without operating a server process, but it cannot expose host-level process/data-directory behavior and currently runs a different server series.

2. Native installation: follow current official OS instructions, then verify locally

Do not copy a years-old package command from a blog. MongoDB publishes separate current Community installation tutorials for Windows, macOS, Ubuntu/Debian, Red Hat-family distributions, SUSE, and other supported platforms. Follow the official page for your OS, then verify the installed artifacts rather than trusting an installer success dialog.

mongod --versionmongosh --version

On Windows, the server binary commonly lives under a versioned C:\Program Files\MongoDB\Server\8.3\bin directory when installed with the official MSI, but service name, configuration path, and data/log paths depend on installer choices. On Linux/macOS, package/service paths likewise depend on the official packaging method. This lesson therefore teaches inspection commands instead of pretending one hard-coded path is universal.

3. Docker installation: pin the server and keep the host exposure local

The official MongoDB Community Docker documentation demonstrates the MongoDB-maintained image and supports version-specific tags. For a reproducible course lab, use the exact 8.3.8 Ubuntu 22.04 slim tag verified during generation instead of latest.

docker pull mongodb/mongodb-community-server:8.3.8-ubuntu2204-slimdocker image inspect mongodb/mongodb-community-server:8.3.8-ubuntu2204-slim --format "{{.RepoTags}} {{.Id}}"docker run --name atlasmart-mongo-l3 \  -p 127.0.0.1:27017:27017 \  -d mongodb/mongodb-community-server:8.3.8-ubuntu2204-slimdocker ps --filter name=atlasmart-mongo-l3docker logs --tail 30 atlasmart-mongo-l3

The image tag pins the server release line, but a production supply-chain policy can go further by recording the pulled image digest. The lesson does not hard-code a digest because platform manifests can differ; record the digest returned by your environment in the lab notebook.

4. Install or verify mongosh separately

mongosh has its own release cadence. At this review point, 2.10.0 is the current release. A server container may bundle a different shell version, so do not infer the shell version from the server version. Install mongosh from the official shell installation/download path for your OS, or use the shell bundled in the container only as a fallback and print its version.

# Host shellmongosh --version# Optional fallback: inspect the shell bundled in the running server imagedocker exec atlasmart-mongo-l3 mongosh --version

5. Smoke test the connection and classify failures

A successful TCP connection is not yet proof of the server version or topology. Connect and run hello plus build information. A failure should be classified rather than “fixed” by changing random settings.

mongosh "mongodb://127.0.0.1:27017/?directConnection=true" --quiet --eval 'printjson(db.adminCommand({hello:1}));printjson({serverVersion: db.serverBuildInfo().version});' 

6. Optional Atlas Free path

If Docker/native installation is unavailable, Atlas Free can provide a free managed learning database. Current documentation says Free clusters never expire, allow one Free cluster per project, and run MongoDB 8.0. Create a database user, add only the required client IP to the project access list, obtain the service-generated connection string, and connect with mongosh. Treat the URI as a secret-bearing configuration value; do not commit it.

# Shape only — obtain the real SRV URI from your own Atlas project.# Never paste a real password into source control.mongosh "mongodb+srv://<cluster-host>/" --username <db-user>

7. Deliberately wrong approach: moving tags and production URIs in shell history

Two habits destroy reproducibility quickly: using mongodb/mongodb-community-server:latest and pasting a production URI with an embedded password into commands or documentation. The moving tag can point to a different server tomorrow. A credential-bearing URI can leak through shell history, process inspection, screenshots, logs, or Git.

The repair is to pin the image tag, record the resulting digest, use a disposable local database for labs, and provide credentials through interactive prompts or protected environment/secret files when authentication is required. Chapter 01 never requires a production endpoint.

8. Cleanup and production judgment

docker rm -f atlasmart-mongo-l3# Optional: remove the pinned image only if you do not need it for later lessons.# docker image rm mongodb/mongodb-community-server:8.3.8-ubuntu2204-slim

For a long-lived developer workstation, native packages can be more ergonomic. For CI and disposable experiments, pinned containers usually improve repeatability. For managed-service learning, Atlas is valuable but introduces cloud identity/network/tier semantics that should not be smuggled into local-server lessons. Keep each environment's evidence separate.

The next lesson zooms into the runtime objects you now have: processes, ports, data paths, connection strings, databases, and collections.

Verification checklist

• You can print exact mongod and mongosh versions.
• Your Docker track uses an exact 8.3.8 tag, not latest.
• Your unauthenticated local server is published only to 127.0.0.1.
• You can distinguish connection-refused, authentication, URI/DNS, and version/edition failures.
• If you used Atlas Free, you labeled its server version separately and did not commit credentials.

Authoritative references

• MongoDB release notes — Official source for the current stable server series and patch notes.
• MongoDB 8.3 release notes — Official 8.3 behavior, patch, compatibility, and security/reliability change log.
• mongosh release notes — Official shell release history; 2.10.0 was released 13 August 2026.
• Install MongoDB — Official OS-specific Community/Enterprise installation index.
• Install Community with Docker — Official MongoDB-maintained container instructions and version-specific tags.
• Install mongosh — Official shell installation guidance.
• Deploy a Free Atlas cluster — Official managed free-learning path and security quickstart.