0

I run a custom gitlab-runner to processes the ci-pipeline of my projects inside docker containers on my own machine. This machine is a centos 7 machine and hence uses podman instead of docker. I once managed to to set everything up and it worked flawlessly for quite some time. I now had to restart the machine and since then all pipelines fail with the error:

Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?

Since we use podman (vi the docker-cli) there is indeed no /var/run/docker.sock (but as I said it strangely worked in the past). So I followd all the steps described here. I verfied that the podman socket is /run/podman/podman.sock, but now I get:

Running with gitlab-runner 15.2.1 (32fc1585)
  on ci_pipeline_hostname key
Preparing the "docker" executor
ERROR: Failed to remove network for build
ERROR: Preparation failed: unable to parse docker host `/run/podman/podman.sock`

I'm obviously not the only one who runs a gitlab runner on a centos based OS. So there must be some way to get it (back to) working.

M0M0
  • 200
  • 9
  • I thing that you need to upgrade your gitlab-runner to >= 15.1. See https://gitlab.com/gitlab-org/gitlab-runner/-/issues/29108 – Claudio Weiler Aug 04 '22 at 13:02
  • this did not help, I get the exact same issue again. Only the version number has changed – M0M0 Aug 04 '22 at 14:11
  • when I run the gitlab-runner locally with gitlab-runner exec docker test everything works fine – M0M0 Aug 04 '22 at 14:28
  • running locally works fine. Actually removing the whole [runners.docker] section from settings.toml (so basically not using the workaround for podman) worked... I keep the question open since I do not understand why it works – M0M0 Aug 05 '22 at 08:32

0 Answers0