Docker Networking Fundamentals

Overview

Create a user-defined bridge network, connect containers by name, publish ports, and know when host/overlay/macvlan apply.

Default bridge isolates containers; user-defined bridges add DNS. Host shares the host stack. Overlay spans Swarm/multi-host. Port mapping publishes container ports to the host.

This is a core tutorial in Module 8 · Networking of the REBASH Academy Docker for Cloud & DevOps Engineers series — written for Cloud, DevOps, Platform, and SRE engineers.

Prerequisites

• Volumes and Persistent Storage
• Networking basics from the Networking track help

Learning Objectives

By the end of this tutorial, you will be able to:

• Create and inspect a bridge network
• Reach containers via DNS names
• Publish ports with -p
• Contrast bridge vs host
• Outline overlay / macvlan use cases

Architecture

This topic’s control points and relationships are shown below.

Theory

What

Docker networks connect containers to each other and to the outside world. The default bridge network suits single-host apps; host shares the host network namespace; none isolates; overlay serves multi-host Swarm; macvlan makes containers appear as LAN hosts. Published ports map container ports to the host.

Why

Most “container cannot connect” tickets are network misunderstandings: wrong network, unpublished ports, or DNS name mismatches. Compose and Swarm rely on user-defined networks so service names resolve via embedded DNS.

How it works

User-defined bridge networks give containers IP addresses and DNS entries based on container or Compose service names. docker run --network attaches at start; you can also connect later. Publishing -p 8080:80 forwards host port 8080 to container port 80. Troubleshoot with docker network inspect, docker exec plus wget/nc, and host tools such as ss or lsof for published ports. Firewall rules on the host can still block traffic after Docker’s iptables/nftables integration.

Driver	Typical use
bridge	Single-host apps (default for Compose)
host	Max performance / special networking
none	Locked down (no NIC)
overlay	Multi-host Swarm
macvlan	Appear as LAN hosts

Key concepts

• Embedded DNS — service discovery on user-defined networks
• Publish vs expose — publish creates host forwarding
• Hairpin / localhost — Desktop vs Linux differences
• Network policies — Docker alone is not Kubernetes NetworkPolicy

Common pitfalls

• Using the legacy default bridge without DNS service names
• Binding only to 127.0.0.1 then wondering why other hosts cannot connect
• Assuming containers share localhost with the host (they do not, except host mode)
• Overlapping subnet CIDRs with corporate VPNs

Hands-on Lab

Objective

Create a custom bridge network, attach two containers, resolve each other by name, publish a port to the host, and capture docker network inspect evidence.

Prerequisites

• Docker Engine or Docker Desktop
• curl on the host

Lab environment

Workspace: ~/rebash-docker/module-08

Custom network rebash-mod08-net; host port 18086 for the web container.

Terminal

mkdir -p ~/rebash-docker/module-08 && cd ~/rebash-docker/module-08

Real-world scenario

Two microservices on the same Docker host must talk over a private network with DNS names, while only the web tier exposes a port to engineers on localhost. You create the network, start nginx and an Alpine client, curl from client to web by service name, and inspect the network attachment.

Step-by-step tasks

Task 1 – Create custom bridge network

Terminal

cd ~/rebash-docker/module-08
docker network create rebash-mod08-net | tee network-create.txt
docker network ls --filter name=rebash-mod08-net | tee network-ls.txt
grep -q 'rebash-mod08-net' network-ls.txt

Expected output

network-ls.txt lists rebash-mod08-net as bridge driver.

Task 2 – Start web and client on the network

Terminal

cd ~/rebash-docker/module-08
docker run -d --name rebash-mod08-web --network rebash-mod08-net -p 18086:80 nginx:1.27-alpine
docker run -d --name rebash-mod08-client --network rebash-mod08-net alpine:3.20 sleep 600
docker ps --filter network=rebash-mod08-net --format '{{ "{{" }}.Names{{ "}}" }}' | tee network-containers.txt
grep rebash-mod08-web network-containers.txt
grep rebash-mod08-client network-containers.txt

Expected output

Both container names appear attached to the network.

Task 3 – DNS by name, host curl, and network inspect

Terminal

cd ~/rebash-docker/module-08
docker exec rebash-mod08-client wget -qO- http://rebash-mod08-web/ | head -n 5 | tee client-to-web.txt
grep -qi 'nginx' client-to-web.txt
curl -sI http://127.0.0.1:18086 | head -n 3 | tee host-curl.txt
grep -qi 'HTTP/' host-curl.txt
docker network inspect rebash-mod08-net --format '{{ "{{" }}range .Containers{{ "}}" }}{{ "{{" }}.Name{{ "}}" }} {{ "{{" }}end{{ "}}" }}' | tee network-inspect-names.txt
grep rebash-mod08-web network-inspect-names.txt

Expected output

Client resolves rebash-mod08-web and returns HTML; host curl gets HTTP headers; inspect lists both container names.

Validation steps

• Custom network rebash-mod08-net exists
• Client container reached web container via DNS name on the network
• Host port 18086 responds and network-inspect-names.txt lists attachments

Common errors and fixes

Error	Cause	Fix
wget: bad address	Containers on default bridge	Attach both to rebash-mod08-net
port is already allocated	18086 in use	Stop conflicting container or change port
network with name exists	Previous lab	docker network rm rebash-mod08-net after cleanup

Challenge exercise

Add a second client that fails to resolve a name off-network — prove isolation by pinging the web IP from inside the client (should work) versus wrong hostname:

Terminal

cd ~/rebash-docker/module-08
docker exec rebash-mod08-client ping -c 1 rebash-mod08-web | tee ping-dns.txt
grep -q '1 packets transmitted' ping-dns.txt

Expected output

ping-dns.txt shows one successful ping to the web container by name.

Learning outcomes

• Created an user-defined bridge with embedded DNS
• Published only the web tier while keeping client internal
• Used docker network inspect to audit container attachments

Cleanup

Terminal

cd ~/rebash-docker/module-08
docker rm -f rebash-mod08-web rebash-mod08-client 2>/dev/null || true
docker network rm rebash-mod08-net 2>/dev/null || true
docker rmi nginx:1.27-alpine alpine:3.20 2>/dev/null || true

Validation

• Lab commands run under ~/rebash-docker/module-08/
• client-to-web.txt and network-inspect-names.txt prove DNS and inspect goals
• You can explain each Theory section in your own words
• You can describe one production failure mode for this topic

Code Walkthrough

Production practice for Docker Networking Fundamentals always combines:

1. Inspect before you change (status, plan, logs, dry-run)
2. Prefer reversible, documented changes (Git, IaC, drop-ins, version pins)
3. Capture evidence (command output, pipeline logs) for handovers
4. Prefer current tools and APIs over legacy shortcuts
5. Least privilege — escalate credentials only when required

Keep runbooks short enough to follow under pressure. Automate checks; keep humans for judgement.

Security Considerations

• Treat credentials and tokens for docker as privileged — never commit them
• Prefer short-lived auth (OIDC, roles, SSO) over long-lived keys
• Validate blast radius before apply/deploy/delete operations
• Restrict who can approve production changes
• Collect audit logs; limit who can read sensitive traces

Common Mistakes

Using the legacy default bridge without DNS service names

Validate assumptions against the Theory section and official docs before changing production.

Binding only to 127.0.0.1 then wondering why other hosts cannot connect

Lab shortcuts (open security groups, admin roles, skip approvals) must not ship unchanged.

Changing production without a rollback path

Always know how to revert (previous artefact, prior release, state rollback, DNS failback).

Best Practices

• Encode Docker Networking Fundamentals changes as code and review them in pull requests
• Pin versions (images, modules, actions, provider plugins)
• Separate environments with clear promotion gates
• Alert on symptoms with runbooks attached
• Destroy lab resources; tag everything with owner and expiry where possible

Troubleshooting

Symptom	Likely cause	Fix
Auth / permission denied	Wrong identity, policy, or scope	Check caller identity, roles, and least-privilege policies
Timeout / no route	Network, DNS, security group, or endpoint	Trace path, DNS, and allow-lists before retrying
Drift / unexpected plan	Manual change or wrong state/workspace	Reconcile desired vs actual; avoid click-ops on managed resources
Pipeline/job red	Flaky step, cache, or missing secret	Read failing step logs; bisect recent workflow/config changes
Cost spike	Idle load balancer, NAT, oversized compute	Inventory billable resources; stop/delete labs promptly

Summary

Docker Networking Fundamentals is essential for Cloud and DevOps engineers working with docker. Practise the lab until the inspection and change path is muscle memory, then continue the track.

Interview Questions

1. Bridge versus host versus none networks?
2. How do containers resolve each other by name?
3. Published ports versus container ports?
4. When do custom networks beat the default bridge?
5. How do you inspect connectivity failures?

Sample answer — question 2

Use docker network inspect and confirm both containers share the network.

Sample answer — question 4

Avoid host networking unless required; it weakens isolation.

Related Tutorials

• Course overview
• Docker Compose Fundamentals

References

• Docker networking overview