Running a task fails in /opt/resource/check with 'failed to ping registry'

[douardda]

I am trying to follow the tutorial using the docker-compose.yml to run concourse.

When I try to execute the hello_world task, it fails with:

resource script '/opt/resource/check []' failed: exit status 1

stderr: failed to ping registry: 2 error(s) occurred:

• ping https: Get https://registry-1.docker.io/v2/: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)
• ping http: Get http://registry-1.docker.io/v2/: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)

I can curl the docker registry from within the running worker (using a docker-compose exec concourse-worker bash).

My host is running Debian stretch with

• kernel 4.18.6-1~bpo9+1
• docker 18.09.1
• docker-compose 1.23.2 (pip installed in a python 3.5 venv).
• concourse image 474635a78a6c , concourse 4.2.2

Any idea?

Thanks, David

[douardda]

Ok so the problem looks similar to https://github.com/concourse/concourse/issues/2388

I finally found the problem: my network does not allow dns traffic other than the internal dns server, so I have modified the docker-compose file to force using this dns server adding the CONCOURSE_GARDEN_DNS_SERVER environment variable.

This DNS quirck should be documented in the readme file or so, I guess.

[kingcoconut]

I was having this issue when running a task on the amidos/dcind image in privileged mode. I was unable to pull images from my ECR registry. I am running concourse on k8s, deployed via the helm/charts. Setting CONCOURSE_GARDEN_DNS_SERVER to 8.8.8.8 fixed my issue