You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@beam.apache.org by GitBox <gi...@apache.org> on 2020/12/15 01:44:09 UTC
[GitHub] [beam] robertwb commented on a change in pull request #13420: Rewrite container docs for custom containers
robertwb commented on a change in pull request #13420:
URL: https://github.com/apache/beam/pull/13420#discussion_r542980881
##########
File path: website/www/site/content/en/documentation/runtime/environments.md
##########
@@ -15,56 +15,168 @@ See the License for the specific language governing permissions and
limitations under the License.
-->
-# Container environments
+# Container Environments
-The Beam SDK runtime environment is isolated from other runtime systems because the SDK runtime environment is [containerized](https://s.apache.org/beam-fn-api-container-contract) with [Docker](https://www.docker.com/). This means that any execution engine can run the Beam SDK.
+The Beam SDK runtime environment is [containerized](https://www.docker.com/resources/what-container) with [Docker](https://www.docker.com/) to isolate it from other runtime systems. This means any execution engine can run the Beam SDK. To learn more about the container environment, read the Beam [SDK Harness container contract](https://s.apache.org/beam-fn-api-container-contract).
-This page describes how to customize, build, and push Beam SDK container images.
+Prebuilt SDK container images are released per supported language version during Beam releases and and pushed to [Docker Hub](https://hub.docker.com/search?q=apache%2Fbeam&type=image)
-Before you begin, install [Docker](https://www.docker.com/) on your workstation.
+## Custom Containers
-## Customizing container images
+Users may want to customize container images for many reasons, including:
-You can add extra dependencies to container images so that you don't have to supply the dependencies to execution engines.
+* pre-installing additional dependencies,
+* launching third-party software
+* further customizing the execution environment
-To customize a container image, either:
-* [Write a new](#writing-new-dockerfiles) [Dockerfile](https://docs.docker.com/engine/reference/builder/) on top of the original.
-* [Modify](#modifying-dockerfiles) the [original Dockerfile](https://github.com/apache/beam/blob/master/sdks/python/container/Dockerfile) and reimage the container.
+ This guide describes how to create and use customized containers for the Beam SDK.
-It's often easier to write a new Dockerfile. However, by modifying the original Dockerfile, you can customize anything (including the base OS).
+### Prerequisites
-### Writing new Dockerfiles on top of the original {#writing-new-dockerfiles}
+* You will need to have a version of the Beam SDK >= 2.21.0.
+* You will need to have [Docker installed](https://docs.docker.com/get-docker/).
+* You will need to have a container registry accessible by your execution engine or runner to host a custom container image. Options include [Docker Hub](https://hub.docker.com/) or a "self-hosted" repository, including cloud-specific container registries.
-1. Pull a [prebuilt SDK container image](https://hub.docker.com/search?q=apache%2Fbeam&type=image) for your [target](https://docs.docker.com/docker-hub/repos/#searching-for-repositories) language and version. The following example pulls the latest Python SDK:
+> **NOTE**: On Nov 20, 2020, Docker Hub put [rate limits](https://www.docker.com/increase-rate-limits) into effect for anonymous and free authenticated use, which may impact larger pipelines that pull containers several times.
+
+### Building and pushing custom containers
+
+Beam builds prebuilt images from [Dockerfiles](https://docs.docker.com/engine/reference/builder/). Users can build customized containers in one of two ways:
+
+1. **[Writing a new](#writing-new-dockerfiles) Dockerfile based on an existing prebuilt container**. This is sufficient for simple additions to the image, such as adding artifacts or environment variables.
+2. **[Modifying](#modifying-dockerfiles) an existing Dockerfile in [Beam source](https://github.com/apache/beam)**. This method requires building from Beam source but allows for greater customization of the container (including replacement of artifacts or base OS/language versions).
+
+#### Writing new Dockerfiles on top of the original {#writing-new-dockerfiles}
+
+Steps:
+
+1. Create a new Dockerfile that designates a base image using the [FROM instruction](https://docs.docker.com/engine/reference/builder/#from)
+
+2. Once you have a created a custom Dockerfile, [build](https://docs.docker.com/engine/reference/commandline/build/) and [push](https://docs.docker.com/engine/reference/commandline/push/) the image using Docker:
+
+As an example, this `Dockerfile`:
+
+```
+FROM apache/beam_python3.7_sdk:2.25.0
+
+ENV FOO=bar
+COPY /src/path/to/file /dest/path/to/file/
```
-docker pull apache/beam_python3.7_sdk
+
+uses the prebuilt Python 3.7 SDK container image [`beam_python3.7_sdk`](https://hub.docker.com/r/apache/beam_python3.7_sdk) tagged at (SDK version) `2.25.0`, and adds an additional environment variable and file to the image.
+
+```
+export BASE_IMAGE="apache/beam_python3.7_sdk:2.25.0"
+export IMAGE_NAME="myremoterepo/mybeamsdk"
+export TAG="latest"
+
+# Optional but recommended pull step to pull the base image into your local Docker daemon.
+docker pull "${BASE_IMAGE}"
+docker build -f Dockerfile -t "${IMAGE_NAME}:${TAG}" .
+docker push "${IMAGE_NAME}:${TAG}"
```
-2. [Write a new Dockerfile](https://docs.docker.com/develop/develop-images/dockerfile_best-practices/) that [designates](https://docs.docker.com/engine/reference/builder/#from) the original as its [parent](https://docs.docker.com/glossary/?term=parent%20image).
-3. [Build](#building-container-images) a child image.
-### Modifying the original Dockerfile {#modifying-dockerfiles}
+**NOTE**: After pushing a container image, you should verify the remote image ID and digest should match the local image ID and digest, output from `docker build` or `docker images`.
+
+#### Modifying the original Dockerfile {#modifying-dockerfiles} in Beam source
+
+This method will require building image artifacts from Beam source - see the [Contribution guide](contribute/#development-setup) for additional instructions on setting up your development environment.
+
+1. Clone the `beam` repository.
-1. Clone the `beam` repository:
```
git clone https://github.com/apache/beam.git
```
-2. Customize the [Dockerfile](https://github.com/apache/beam/blob/master/sdks/python/container/Dockerfile). If you're adding dependencies from [PyPI](https://pypi.org/), use [`base_image_requirements.txt`](https://github.com/apache/beam/blob/master/sdks/python/container/base_image_requirements.txt) instead.
-3. [Reimage](#building-container-images) the container.
-### Testing customized images
+2. Customize the `Dockerfile` for a given language. This file is typically in the `sdks/<language>/container` directory (e.g. the [Dockerfile for Python](https://github.com/apache/beam/blob/master/sdks/python/container/Dockerfile).. If you're adding dependencies from [PyPI](https://pypi.org/), use [`base_image_requirements.txt`](https://github.com/apache/beam/blob/master/sdks/python/container/base_image_requirements.txt) instead.
+
+3. Navigate to the root directory of the local copy of your Apache Beam.
+
+4. Run Gradle with the `docker` target.
+
+
+```
+# The default repository of each SDK
+./gradlew :sdks:java:container:java8:docker
+./gradlew :sdks:java:container:java11:docker
+./gradlew :sdks:go:container:docker
+./gradlew :sdks:python:container:py36:docker
+./gradlew :sdks:python:container:py37:docker
+./gradlew :sdks:python:container:py38:docker
+
+# Shortcut for building all Python SDKs
+./gradlew :sdks:python:container buildAll
+```
+
+To examine the containers that you built, run `docker images`:
+
+```
+$> docker images
+REPOSITORY TAG IMAGE ID CREATED SIZE
+apache/beam_java8_sdk latest ... 1 min ago ...
+apache/beam_java11_sdk latest ... 1 min ago ...
+apache/beam_python3.6_sdk latest ... 1 min ago ...
+apache/beam_python3.7_sdk latest ... 1 min ago ...
+apache/beam_python3.8_sdk latest ... 1 min ago ...
+apache/beam_go_sdk latest ... 1 min ago ...
+```
+
+If you did not provide a custom repo/tag as additional parameters (see below), you can retag the image and [push](https://docs.docker.com/engine/reference/commandline/push/) the image using Docker to a remote repository.
+
+```
+export IMAGE_NAME="myrepo/mybeamsdk"
+export TAG="latest"
+
+docker tag apache/beam_python3.6_sdk "${IMAGE_NAME}:${TAG}"
+docker push "${IMAGE_NAME}:${TAG}"
+```
+
+**NOTE**: After pushing a container image, verify the remote image ID and digest matches the local image ID and digest output from `docker_images`
+
+##### Additional Build Parameters
+
+The docker Gradle task defines a default image repository and [tag](https://docs.docker.com/engine/reference/commandline/tag/) is the SDK version defined at [gradle.properties](https://github.com/apache/beam/blob/master/gradle.properties). The default repository is the Docker Hub `apache` namespace, and the default tag is the [SDK version](https://github.com/apache/beam/blob/master/gradle.properties) defined at gradle.properties. With these settings, the
+`docker` command-line tool will implicitly try to push the container to the Docker Hub Apache repository.
+
+You can specify a different repository or tag for built images by providing parameters to the build task. For example:
+
+```
+./gradlew :sdks:python:container:py36:docker -Pdocker-repository-root=example-repo -Pdocker-tag=2019-10-04
+```
-To test a customized image locally, run a pipeline with PortableRunner and set the `--environment_config` flag to the image path:
+builds the Python 3.6 container and tags it as `example-repo/beam_python3.6_sdk:2019-10-04`.
+
+From 2.21.0, a `docker-pull-licenses` flag was introduced to add licenses/notices for third party dependencies to the docker images. For example:
+
+```
+./gradlew :sdks:java:container:java8:docker -Pdocker-pull-licenses
+```
+creates a Java 8 SDK image with appropriate licenses in `/opt/apache/beam/third_party_licenses/`.
+
+By default, no licenses/notices are added to the docker images.
+
+
+## Using Container Images in Pipelines
+
+The common method for providing a container image requires using the PortableRunner and setting the `--environment_config` flag to a given image path.
+Other runners, such as Dataflow, support specifying containers with different flags.
Review comment:
Regardless of the plans for using the Portability API for submitting jobs, we should standardize on a single flag for specifying the image across runners. It could be argued that environment_config is not the best flag for this; maybe we should take it to the list?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org