Skip to content

robinhuiser/concourse-arm64

 
 

Repository files navigation

Concourse CI for linux/arm64

This repository helps you build both the web and worker arm64 components for Concourse CI - prebuilt Docker images can be found on Docker Hub rdclda/concourse.

Concourse on arm64 screenshot

Bundled resources

Concourse git github-release registry-image semver time mock s3 slack-alert
v7.1.0 v1.12.0 v1.5.2 v1.2.0 v1.3.0 v1.5.0 v0.11.1 v1.1.1 v0.15.0
v7.2.0 v1.12.1 v1.5.2 v1.2.1 v1.3.0 v1.6.0 v0.11.1 v1.1.1 v0.15.0
v7.3.2 v1.14.0 v1.6.1 v1.3.0 v1.3.1 v1.6.0 v0.11.2 v1.1.1 v0.15.0
v7.4.0 v1.14.0 v1.6.2 v1.4.0 v1.3.4 v1.6.1 v0.12.2 v1.1.2 v0.15.0
v7.4.1 v1.14.0 v1.6.4 v1.4.0 v1.3.4 v1.6.1 v0.12.2 v1.1.2 v0.15.0
v7.4.2 v1.14.0 v1.6.4 v1.4.0 v1.3.4 v1.6.1 v0.12.2 v1.1.2 v0.15.0
v7.4.3 v1.14.0 v1.6.4 v1.4.0 v1.3.4 v1.6.1 v0.12.2 v1.1.2 v0.15.0
v7.4.4 v1.14.0 v1.6.4 v1.4.0 v1.3.4 v1.6.1 v0.12.2 v1.1.2 v0.15.0
v7.5.0 v1.14.4 v1.6.4 v1.4.1 v1.3.4 v1.6.2 v0.12.3 v1.1.3 v0.15.0
v7.6.0 v1.14.5 v1.6.4 v1.4.1 v1.3.4 v1.6.2 v0.12.3 v1.1.3 v0.15.0
v7.7.0 v1.14.5 v1.6.4 v1.5.0 v1.3.4 v1.6.3 v0.12.3 v1.2.0 v0.15.0
v7.7.1 v1.14.5 v1.6.4 v1.5.0 v1.3.4 v1.6.3 v0.12.3 v1.2.0 v0.15.0
v7.8.0 v1.14.5 v1.7.1 v1.5.1 v1.4.0 v1.6.3 v0.12.4 v1.2.1 v0.15.0
v7.8.1 v1.14.5 v1.7.2 v1.5.1 v1.4.0 v1.6.3 v0.12.4 v1.2.1 v0.15.0
v7.8.2 v1.14.5 v1.7.2 v1.5.1 v1.4.0 v1.6.3 v0.12.4 v1.2.1 v0.15.0
v7.8.3 v1.14.5 v1.7.2 v1.5.1 v1.4.0 v1.6.3 v0.12.4 v1.2.1 v0.15.0
v7.9.0 v1.14.6 v1.8.0 v1.7.0 v1.5.0 v1.6.3 v0.12.4 v1.2.1 v0.15.0
v7.9.1 v1.14.7 v1.8.0 v1.7.1 v1.5.1 v1.6.3 v0.12.4 v1.2.1 v0.15.0

Bundled CLIs

Each Docker image includes the CLIs for Linux/Mac/Windows for the Intel platform - they can be downloaded from the Concourse web console.

Available Concourse tasks

The following tasks have been ported to the arm64 platform:

Task Image Version Description
dcind rdclda/concourse-dcind-task 1.0.0 A task for running Docker Compose in Docker
oci-build rdclda/concourse-oci-build-task 0.9.0 A task for building OCI images

Deploy

Copy the example docker-compose.yaml to your Raspberry Pi and update the external IP address setting CONCOURSE_EXTERNAL_URL.

# On your Macbook M1/M2 machine
$ podman-compose up -d

# Login using fly - update your IP here too ;-)
$ export FLY_TARGET=my-m2
$ fly --target=$FLY_TARGET login \
    --concourse-url=http://concourse.localtest.me:8080 \
    --username=test \
    --password=test                                                      

You can now access the Concourse web console using http://concourse.localtest.me:8080/.

Tests

These tests are provided to verify the correct working of the bundled resource types.

Prepare some surrounding service requirements first:

# create a public s3 bucket
$ export S3_BUCKET_NAME=rdclda-concourse-s3-test
$ aws s3api create-bucket --acl public-read \
   --bucket $S3_BUCKET_NAME --region us-east-1

# push test file to s3 bucket
$ echo "Looks like the s3 resource is working." | \
   aws s3 cp - s3://$S3_BUCKET_NAME/testfile.txt \
   --acl public-read

# define your Slack webhook URL
$ export SLACK_WEBHOOK_URL=https://hooks.slack.com/services/T00000000/B00000000/XXXXXXXXXXXXXXXXXXXXXXXX

Now, we can deploy and run the tests:

# deploy & kick off the bundled resource tests
$ for resource in registry-image time git s3 mock slack-alert; do
    fly -t $FLY_TARGET set-pipeline -n -p test-${resource}-resource -c tests/$resource-resource.yaml \
      --var s3-bucket-name=$S3_BUCKET_NAME \
      --var slack-webhook-url=$SLACK_WEBHOOK_URL
    fly -t $FLY_TARGET unpause-pipeline -p test-${resource}-resource
    fly -t $FLY_TARGET trigger-job --job test-${resource}-resource/test-job
done

# deploy & kick off the external task test
$ for task in dcind oci-build; do
    fly -t $FLY_TARGET set-pipeline -n -p test-${task}-task -c external-tasks/${task}/example/pipe.yaml
    fly -t $FLY_TARGET unpause-pipeline -p test-${task}-task
    fly -t $FLY_TARGET trigger-job --job test-${task}-task/test-job
done

Use the web console to verify the status of the tests.

BIY

Follow the steps below if you want to build the images yourself.

Prerequisites

  • Raspberry Pi 4 with 8Gb of memory (if you want to build elm)
  • Docker daemon + Docker CLI (buildx enabled)
  • 4Gb of (Docker assigned) memory
  • Bash shell

Build elm

Elm is a build dependency for the Concourse web component, but is not available for arm64 - therefore elm v0.19.1 has been pre-compiled on arm64 and packaged under ./dist within this repository.

The two main reasons to not make the elm native binary compilation part of the Concourse CI build are:

  • Docker buildx fails (crashes) when trying to compile this on amd64 platform
  • Takes too long

In case you want to build elm yourself, follow the steps below:

# Based upon Ubuntu 20.04
# Raspberry Pi 4 with 8Gb memory and SSD storage attached
# Expect build to take up to 3+ hours
apt-get update && apt-get install ghc cabal-install -y
apt-get install git curl -y

git config --global user.email "info@rdc.pt" && \
git config --global user.name "Robin Daniel Consultants, Lda."

mkdir -p /tmp/build && cd /tmp/build
git clone https://github.com/dmy/elm-raspberry-pi.git ./elm-raspberry-pi
cd ./elm-raspberry-pi && git checkout tags/20200611

cd /tmp/build
git clone https://github.com/elm/compiler.git ./elm/compiler
cd ./elm/compiler && git checkout tags/0.19.1

git am -s /tmp/build/elm-raspberry-pi/patches/elm-0.19.1.patch
cabal new-update
cabal new-configure --ghc-option=-split-sections
cabal new-build

After the last step, the build will output the elm binary path.

Build Concourse

You will find under the ./build-specs directory the available configurations for building Concourse CI for arm64.

# Kick off the build - specify the concourse version you want to build
# Using Intel / Docker Compose:
./build.sh 7.9.1

# Using ARM64 (M1, M2) and podman:
./build_podman_arm64.sh 7.9.1

The generated Docker image will be pushed to the specified repository defined in the .env file.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Shell 62.3%
  • Dockerfile 37.7%