mrkeyoor.comAsk

moby/mobytested 09 Oct 2026commit a5b58c9

moby review

Use Moby for an engine contribution or a container system whose engine behavior you intend to maintain.

  • Install 160 s
  • Build 180 s
21/28passed, run exited non-zero
  • 0 packages
Dev Toolsevaluationupdated 31 Oct 2026
+33stars / 7d

Choose Moby when you need to build or modify a container engine. Moby builds container engines. Docker uses it as an upstream, and other container systems can reuse its components. For an application that communicates with Docker Engine, start with the supported client and API modules.

Verdict

Use Moby for an engine contribution or a container system whose engine behavior you intend to maintain. Arrange a host with the documented tooling and permissions before investigating the recorded networking failures. The stored run completed its install and build steps but left testing unfinished; it does not validate a working deployment. For an Engine API application, start with the supported client and API modules.

We ran it

Lab card: what happened when we ran moby
Lab card from our sandbox run
Screenshot of moby (mobyproject.org)
mobyproject.org
Install✓ · 160s0 packages
Build✓ · 180s
Tests✗ timed out · 900s21 passed · 7 failed of 28 (go test)
Repo3383 files~398,979 lines of source · 31 MB · 21 CI workflows · Dockerfile

Answers from our run

Does moby build from source?

Dependencies installed in 160 seconds (0 packages), and the build succeeded in 180 seconds. We cloned commit a5b58c9 into a clean Debian container with 3 CPUs and no project-specific setup.

Do moby's tests pass?

No. The project's own test command finished in 900 seconds and reported 21 passed, 7 failed (go test).

Who should not use moby?

Application teams seeking a finished container product with commercial support.

What are the alternatives to moby?

Podman, containerd, BuildKit. Use Moby for an engine contribution or a container system whose engine behavior you intend to maintain.

Setup2/5Material setup constraints: install and build returned zero, but testing remained unfinished with missing iptables and denied namespace operations in the raw tail.
Docs4/5Public module boundaries and the development sequence are explicit, with a qualification for dated contributor-guide examples.
Community5/5Assesses documented openness to participation and reuse only: external contributors are welcomed and other upstream users receive equal treatment. Response speed, adoption and support quality were not measured.
Maturity4/5Assesses the documented upstream role and explicit module and release boundaries, qualified by best-effort support and the root module's lack of API stability guarantees.

Discussed on

  1. hnA new upstream project to break up Docker into independent components571 points
  2. hnDocker 19.03: Rootless Mode (Experimental)97 points
  3. hnThe Moby Project: accelerated software containerization36 points
  4. hnDocker is still silently punching holes in your firewall after 5 years23 points
  5. hnDocker is transitioning to the Moby Project7 points

Who it’s for

Engineers contributing changes to container-engine behavior.
Infrastructure integrators assembling container systems from replaceable components.
Go developers choosing between supported Engine API modules and engine source.

Who it’s NOT for

Application teams seeking a finished container product with commercial support.
Go applications requiring stable imports from daemon internals.
Contributors unable to provide the host access required by the documented privileged development container.
Readers seeking production workload benchmarks or deployment reliability evidence: the recorded run tested neither.

Setup reality

The stored October 9, 2026 run at commit a5b58c9 records a 160-second install, a 180-second build step and an unfinished test step stopped at 900 seconds. Install and build returned exit code 0; tests returned exit code 124 with the timeout flag set. No new execution was performed for this correction.

The sandbox configuration records 3 CPUs, 8 GB of RAM and the golang:1.24-bookworm image. The install tail records a Go 1.26.8 download, so the image tag does not establish the compiler version used throughout the run.

The stored parser summary records 21 passing and 7 failing package results, with a total of 28. These are not individual-test counts or verified exhaustive unique coverage. The raw tail reports missing iptables, denied network namespace operations and a package timeout; it does not explain every recorded failure.

Choose the source when you need to change the engine

Moby builds container engines. Its README addresses engineers and integrators who want to modify or assemble container systems. Docker uses Moby as an upstream, and the project encourages other systems to reuse its components.

For application deployment, begin with the container product you intend to operate. Moby's README describes release support as best effort and directs customers seeking commercial support to Docker Desktop or Mirantis Container Runtime. Adopting the source tree does not supply a support contract.

What the stored run establishes

Our run on October 9, 2026, at commit a5b58c9 recorded a 160-second install, a 180-second build step and a test timeout at 900 seconds. These are stored observations, not results from a new execution.

The sandbox had 3 CPUs, 8 GB of RAM and the golang:1.24-bookworm image. The install tail says go: downloading go1.26.8 (linux/amd64). The base image's name therefore does not establish that Go 1.24 performed the later work. The dependency counter records 0 packages; that counter cannot mean that nothing was downloaded.

Install and build each returned exit code 0. The build tail contains vet-exit:0. The test step returned exit code 124 and was marked timed out. These command outcomes do not establish successful daemon startup, container execution, image building, CLI deployment or workload performance.

The MrKeyoor testing methodology explains how the publisher records command outcomes and the limits of sandbox testing. That page supplies provenance context, not evidence that this October 9 run used every detail of the currently supplied methodology.

Read package summaries separately from failure details

The stored parser summary records 21 passing and 7 failing package results, with a total of 28. Those values do not establish exhaustive unique coverage or count individual test cases. No complete-log reconciliation was performed.

The raw tail identifies narrower problems. Several named tests in the iptables package report iptables not found. It also reports failed attempts to enter or restore a thread network namespace because the operation was not permitted. A six-minute package timeout occurred while TestRule was running. That package timeout is separate from the overall test step being stopped at 900 seconds.

The netutils package contains a named failure accompanied by denied namespace operations. Other lines show successful package results and packages with no test files. Packages with no test files are not failed or skipped tests. The excerpt does not explain all 7 failing package results, so missing iptables and denied permissions cannot be assigned as the cause of every failure.

Before investigating these networking paths, provide the documented development environment. The stored run identifies missing tooling and denied operations in this sandbox; it does not establish how those paths behave on a suitably configured host.

Check host access before using the development shell

The development-container guide shows make shell launching a container with --privileged. Its workflow then builds binaries, installs them inside the development container, starts the daemon and runs a hello-world container. These are documented steps, not outcomes demonstrated by the stored run.

The prerequisites list Git, Make and Docker. The development environment supplies Go, and contributing through GitHub requires a GitHub account. On Linux, the instructions assume Docker access through the docker group or sudo.

Confirm that your host permits the required access. The guides retain Docker Toolbox references and old version examples, so check the instructions relevant to your host. Review the cleanup commands before running them on a machine with images or containers you need.

Use supported modules for an Engine API application

The root Go module has no API stability guarantees. The README identifies github.com/moby/moby/v2 as the codebase for building engine binaries and says it is not intended for import as a Go library.

The supported public modules are github.com/moby/moby/client, the Docker Engine API client, and github.com/moby/moby/api, the shared API types. Start with these for an application that communicates with an engine. Their tags are versioned independently from engine release tags.

The README also says the old github.com/docker/docker module is deprecated. Engine tags with the docker- prefix are for building engine binaries and must not be consumed through go get. Check module boundaries and migration notes before replacing imports; compilation alone does not establish API stability.

Choose alternatives by responsibility

Podman supports running containers and pods without root or other elevated privileges. Its README documents a Docker-compatible CLI and no manager daemon. Consider it when you need to operate containers through that model. Rootless setup can require administrator configuration; the Moby run does not establish Podman compatibility with the same sandbox.

Containerd is designed to be embedded into a larger system. Its README describes a daemon managing container lifecycle tasks, including image storage, execution and supervision. Evaluate it when your platform needs that runtime layer, then compare the responsibilities your system requires.

BuildKit converts source code into build artifacts. Its README says Docker builds use Buildx and BuildKit by default since Docker Engine 23.0. BuildKit also complements the Docker engine stack. Consider standalone BuildKit when you need direct builder control. The BuildKit review addresses that separate decision.

What the ratings mean

These scores are editorial judgments about contributor suitability. They do not measure adoption, security or workload performance. The scale is 1 for substantial obstacles or weak supporting evidence, 2 for material constraints, 3 for a mixed assessment, 4 for strong support with a meaningful qualification, and 5 for the strongest fit within the stated dimension.

Setup ease is 2/5: the recorded test step remained unfinished, with missing tooling and denied operations in the tail. Documentation is 4/5: module boundaries and the contributor sequence are explicit, although the guides retain dated examples.

Community is 5/5 specifically for documented openness to participation and reuse. The README welcomes external maintainers and contributors and says other upstream users will receive equal treatment. This narrow judgment does not assess response times, contributor volume or support quality.

Maturity is 4/5 for Moby's documented role as Docker's upstream and its explicit module and release boundaries, qualified by best-effort support and unstable root-module APIs. These scores do not certify a production configuration.

Alternatives

ProjectWhat it isPick it when
Podman gh↗A container and pod manager with a Docker-compatible CLI, rootless operation and no manager daemon.Choose it for managing containers through its operating model. Verify host configuration requirements separately; the Moby run is not a Podman compatibility test.
containerd gh↗A container lifecycle runtime designed for embedding in larger systems.Evaluate it when your platform needs an embedded runtime layer. Compare required responsibilities before replacing an engine built from Moby.
BuildKit gh↗A build toolkit that also complements Docker's engine stack.Consider standalone BuildKit when you need direct builder control. Its artifact-building role is separate from managing container lifecycles.

Sources

  1. Moby README and module guidance
  2. Moby development container
  3. Moby development prerequisites
  4. Docker Engine 29.9.0 release
  5. Health checks retaining stale healthy status
  6. Podman README
  7. containerd README
  8. BuildKit README

More dev tools reviews