Docker 如何选择图像的操作系统/架构

Question

I started to learn how to work with Docker and I had this question. When I make "docker pull <image-name>" I usually pull the image from docker hub.

But images can have different versions that differ in OS and architecture. For example a python image has versions with linux and windows.

I'm wondering what image I'll get if I just do a "docker pull python" and how I can choose the OS and architecture myself.

I actually prefer to use linux images because it's lightweight and I don't want accidentally pull a windows image.

Answer 1

It's based on the host platform where docker is running. You can see this in docker info :

$ docker info --format '{{ .OSType }}/{{ .Architecture }}'
linux/x86_64

The code for this is in containerd and the OS and Arch are looked up slightly differently in go:

$ go env
GO111MODULE=""
GOARCH="amd64"
...
GOOS="linux"
...

Answer 2

I'm wondering what image I'll get if I just do a "docker pull python"

From " Leverage multi-CPU architecture support "

Most of the Docker Official Images on Docker Hub provide a variety of architectures.
For example, the busybox image supports amd64 , arm32v5 , arm32v6 , arm32v7 , arm64v8 , i386 , ppc64le , and s390x .
When running this image on an x86_64 / amd64 machine, the amd64 variant is pulled and run.

So it depends on the OS used when you do your docker pull .

---

and how I can choose the OS and architecture myself.

The same page adds:

You can also run images targeted for a different architecture on Docker Desktop.

You can run the images using the SHA tag, and verify the architecture.
For example, when you run the following on a macOS:

 docker run --rm docker.io/username/demo:latest@sha256:2b77acdfea5dc5baa489ffab2a0b4a387666d1d526490e31845eb64e3e73ed20 uname -m

aarch64

 docker run --rm docker.io/username/demo:latest@sha256:723c22f366ae44e419d12706453a544ae92711ae52f510e226f6467d8228d191 uname -m

armv71

In the above example, uname -m returns aarch64 and armv7l as expected, even when running the commands on a native macOS or Windows developer machine