ARC: How to run multiple containers pod properly?

[dpanas-sollers]

Why are you starting this discussion?

Question

What GitHub Actions topic or product is this about?

ARC (Actions Runner Controller)

Discussion Details

Hi. Im coming from jenkins world. Jenkins k8s plugin allows 2 important things:

1. Specify multiple containers in one pod.
   Main container is mandatory and default (it has agent service running) but doesnt even have to be used at all.

2. For each stage you can use any container specified in pod template.
   They do not have to be modified (added jenkins agent service with auth token) just to run commands on them. So i can use generic containers or those built by other projects. And yet jenkins is fully aware of them (their health) and commands can be run there.

So i tried to replicate my CI setup/pipeline - some maven container and oracle database container - where i run both in one pod. Currently o dont use maven runner yet.

This is my values.yaml

# based on https://github.com/actions/actions-runner-controller/blob/master/charts/gha-runner-scale-set/values.yaml
githubConfigUrl: "https://github.com/my-repo"
githubConfigSecret:
  github_token: "ghp_w-CENSORED-D117"
template:
  spec:
    containers:
      - name: runner
        image: ghcr.io/actions/actions-runner:latest
        command: ["/home/runner/run.sh"]
      - name: database
        image: mycustomregistry/mydatabase:latest
        env:
          - name: DB_SERVICE_NAME
            value: some-db-service
          - name: TZ
            value: Europe/Warsaw
        securityContext:
          runAsUser: 1001220000

And pipeline

name: test

on:
  push:
    branches:
      - master

jobs:
  hello:
    name: Run test
    runs-on: gh-runners-k3s-1

    steps:
      - name: Checkout repository
        uses: actions/checkout@v3

      - name: Print environment and system info
        run: |
          echo "Running on agent ubuntu2204 from domanpool"
          uname -a
          printenv | sort

And i already see that both containers are running. BUT, github is not aware of second container - it may fail or not start and if i dont use any command to connect to database, the pipeline will be green. But ok, before my main command i have to use healtchecks anyway to check if database is up and ready for connection.

So i assume point 1. is working (kind of)

But i don't know how to tacle point 2. in ARC. How to be able to run commands on those other pods? I mean i dont need that now, on this database container, but i use it in my other projects very often.

Overal what is correct approach in github actions world for such workflows?

[ChandrakantaMandal]

🧠 First, Why Use Multiple Containers in a Pod?
In Kubernetes, a Pod is the smallest deployable unit, and it can contain one or more tightly-coupled containers that:

Share the same network (localhost)

Share volumes

Are scheduled together on the same node

You typically use multiple containers in a Pod when they work together as a single unit — for example:

A main app container + a logging/metrics sidecar

A main app + a cache warm-up container

A proxy container alongside your backend

apiVersion: v1
kind: Pod
metadata:
name: multi-container-pod
spec:
containers:
- name: main-app
image: node:18
ports:
- containerPort: 3000
volumeMounts:
- name: shared-logs
mountPath: /var/log/app

- name: log-sidecar
  image: busybox
  command: ["sh", "-c", "tail -f /var/log/app/app.log"]
  volumeMounts:
    - name: shared-logs
      mountPath: /var/log/app

volumes:
- name: shared-logs
emptyDir: {}

[dpanas-sollers]

To answer your question:

1. To be able to run two ephemeric containers together and shut them down together when job ends. In my current situation CI spins up two containers - one is java8 container with ant, which runs liquibase changes on database container. Database container has to be ephemeric - "fresh" for each job. And should not be modified specifiacally for github actions.
2. I already had projects where CI used 3 containers. For example custom sonarqube plugin. It was compiled in one container, later was copied and some config was adjusted on sonarqube container. And then third container was running some testing scans. So pipeline run commands on all three containers.

Im open for ideas of how to run such pipeline different way (github actions way) as long as it ...

• wont be more troublesome
• wont take more time
• wont have some serious security implications
• allow concurrent builds.

[arielev-oasis]

@dpanas-sollers your second point is one of the most basic things i'm missing from github, doing this in jenkins k8s plugin was amazing, but sadly github arc is not k8s native.

@ChandrakantaMandal there are many reasons running multiple containers in a single pod, a use case i can tell about is for example, a CI that executed N binaries, lets say, a terraform code that part of the CI runs terraform fmt, tflint, checkov, terraform-docs and maybe others

you can, create a custom image, that contains all these binaries, maintain it, release it, upgrade it etc
or you don't need to do any of this, but just declare a k8s pod, that contains the official open source images, so i end up with a pod that contains 5 container, and when i need to execute checkov for example, i just run it from that specific container

[github-actions[bot]]

🕒 Discussion Activity Reminder 🕒

This Discussion has been labeled as dormant by an automated system for having no activity in the last 60 days. Please consider one the following actions:

1️⃣ Close as Out of Date: If the topic is no longer relevant, close the Discussion as out of date at the bottom of the page.

2️⃣ Provide More Information: Share additional details or context — or let the community know if you've found a solution on your own.

3️⃣ Mark a Reply as Answer: If your question has been answered by a reply, mark the most helpful reply as the solution.

Note: This dormant notification will only apply to Discussions with the Question label. To learn more, see our recent announcement.

Thank you for helping bring this Discussion to a resolution! 💬