User namespace remapping in Docker - Step-by-Step Execution

Process Flow - User namespace remapping

Docker daemon starts

↓

Check userns-remap config

↓

If enabled

→Map container user IDs to host user IDs

↓

Create subuid and subgid ranges

↓

Run container with remapped IDs

↓

Container user ID != Host user ID

↓

Improved host security

↓

End

Docker daemon checks if user namespace remapping is enabled, then maps container user IDs to different host user IDs to improve security.

Execution Sample

Docker

dockerd --userns-remap=default

docker run -it alpine id

Starts Docker daemon with user namespace remapping enabled, then runs a container to show user ID mapping.

Process Table

Step	Action	User ID in Container	Mapped Host User ID	Result
1	Docker daemon starts with --userns-remap=default	-	-	User namespace remapping enabled
2	Docker creates subuid and subgid ranges for remapping	-	-	Ranges assigned to dockremap user
3	Run container with 'id' command	0 (root inside container)	100000 (mapped host user)	Container root mapped to non-root host user
4	Container process runs as remapped user	1000 (normal user inside container)	101000 (mapped host user)	Normal user remapped accordingly
5	Container exits	-	-	User namespace remapping ends with container lifecycle

💡 Container stops; user namespace remapping applies only during container runtime

Status Tracker

Variable	Start	After Step 3	After Step 4	Final
Container User ID	-	0	1000	-
Mapped Host User ID	-	100000	101000	-
Docker Daemon State	No remap	Remap enabled	Remap enabled	Remap enabled

Key Moments - 3 Insights

Why does the container root user (ID 0) map to a high host user ID like 100000?

Because user namespace remapping shifts container user IDs to a safe range on the host to avoid running as real root, as shown in execution_table step 3.

Does the container user ID inside the container change when remapping is enabled?

No, inside the container the user IDs remain the same (e.g., 0 for root), but they map to different host IDs, as seen in execution_table steps 3 and 4.

What happens if user namespace remapping is not enabled?

Container user IDs map directly to the same host user IDs, meaning root in container is root on host, which is less secure. This is implied by the initial daemon state in variable_tracker.

Visual Quiz - 3 Questions

Look at the execution table at step 3: what is the mapped host user ID for container root user?

A 101000

B 0

C 100000

D 500

💡 Hint

Check the 'Mapped Host User ID' column at step 3 in the execution_table.

At which step does the container process run as a remapped normal user (ID 1000 inside container)?

A Step 3

B Step 4

C Step 2

D Step 5

💡 Hint

Look for '1000 (normal user inside container)' in the 'User ID in Container' column in execution_table.

If user namespace remapping was disabled, what would happen to the container root user ID mapping?

A It would remain as 0 on the host

B It would map to a high host user ID like 100000

C It would map to 101000

D It would map to 500

💡 Hint

Refer to key_moments about what happens without remapping and variable_tracker Docker Daemon State.

Concept Snapshot

User namespace remapping shifts container user IDs to different host IDs.
Enables running containers as root inside but non-root on host.
Configured via dockerd --userns-remap.
Improves host security by isolating container users.
Container user IDs stay same inside container.
Host sees remapped user IDs in a safe range.

Full Transcript

User namespace remapping in Docker lets the container run users like root inside, but maps those user IDs to different, non-root user IDs on the host. When Docker daemon starts with --userns-remap, it creates subuid and subgid ranges for remapping. For example, container root user ID 0 maps to host user ID 100000. This prevents containers from having real root privileges on the host, improving security. Inside the container, user IDs remain unchanged, but on the host they are shifted. This remapping applies only during container runtime and ends when the container stops.