Self-Hosted Gitea Runner: Docker Settings Ignored in Workflow

Actions
1

Hello,

I have a pipeline step, which starts a docker container itself which is in network_mode: bridge.
I want to do some requests to that container from my current gitea runner container. Further this runner container should have a mount of type “bind” from /tmp/ to /tmp/.

So I did:

name: CI/CD Pipeline (Deploy to Production)
on:
  push:
    tags:
      - "*"

jobs:
  build-keycloak-test-image:
    runs-on: ubuntu-latest
    container:
      image: docker:dind
      network: "" # already tried network_mode: "bridge" and network: "bridge"
      volumes:
        - /tmp:/tmp

After some google research I assumed that works, but both seems to be ignored.
My docker container with image “docker:dind” uses network: “GITEA-ACTIONS-TASK-79_WORKFLOW-CI-CD-Pipeline-Deploy-to-Production_JOB-build-keycloak-test-image-build-keycloak-test-image-network”

and the mounts are just the default ones:

 "Mounts": [
            {
                "Type": "volume",
                "Name": "act-toolcache",
                "Source": "/var/lib/docker/volumes/act-toolcache/_data",
                "Destination": "/opt/hostedtoolcache",
                "Driver": "local",
                "Mode": "z",
                "RW": true,
                "Propagation": ""
            },
            {
                "Type": "volume",
                "Name": "GITEA-ACTIONS-TASK-79_WORKFLOW-CI-CD-Pipeline-Deploy-to-Production_JOB-build-keycloak-test-image-env",
                "Source": "/var/lib/docker/volumes/GITEA-ACTIONS-TASK-79_WORKFLOW-CI-CD-Pipeline-Deploy-to-Production_JOB-build-keycloak-test-image-env/_data",
                "Destination": "/var/run/act",
                "Driver": "local",
                "Mode": "z",
                "RW": true,
                "Propagation": ""
            },
            {
                "Type": "volume",
                "Name": "GITEA-ACTIONS-TASK-79_WORKFLOW-CI-CD-Pipeline-Deploy-to-Production_JOB-build-keycloak-test-image",
                "Source": "/var/lib/docker/volumes/GITEA-ACTIONS-TASK-79_WORKFLOW-CI-CD-Pipeline-Deploy-to-Production_JOB-build-keycloak-test-image/_data",
                "Destination": "/workspace/dev/ember",
                "Driver": "local",
                "Mode": "z",
                "RW": true,
                "Propagation": ""
            },
            {
                "Type": "volume",
                "Name": "37c34b75cf487d36d0153dd6862829f0ca870085e7fdcc10b417d6f51ddb4858",
                "Source": "/var/lib/docker/volumes/37c34b75cf487d36d0153dd6862829f0ca870085e7fdcc10b417d6f51ddb4858/_data",
                "Destination": "/var/lib/docker",
                "Driver": "local",
                "Mode": "",
                "RW": true,
                "Propagation": ""
            },
            {
                "Type": "bind",
                "Source": "/var/run/docker.sock",
                "Destination": "/var/run/docker.sock",
                "Mode": "",
                "RW": true,
                "Propagation": "rprivate"
            }
        ],

Is this maybe a bug? Because the runner does not say something about invalid properties.

Thank you very much for your support :wink:

2

After some help, I found this documentation of the config.yml of the gitea runner:

I found the config property container.network.
After setting that to bridge it worked!
Further I could add ’ /tmp/’ to the container.valid_volumes and after that the mounted volume from the workflow file worked!
If somebody run in that issue, you are welcome!

3

It seems like mounting volumes to job containers does not work anymore. I’m trying this with the latest gitea and acc_runner v 0.3.0. Set container.valid_volumes to “m2-cache“, created the volume with docker volume create m2-cache (wrote something to it using a busybox, so it’s working) and the build always shows the same message: [m2-cache] is not a valid volume, will be ignored.

Followed instructions to the letter and still no clue what is going wrong.