Description of problem: During installation a check is made for docker images, however if you have to use a proxy this test will fail on skopeo as proxy, are not set before the check. causing the skopeo to time out. openshift_disable_check docker_image_availability can be set to disable the check, however checking should support the use of proxy. How reproducible: All the time Steps to Reproduce: 1. Have servers behind proxy. 2. Setup proxy settings in inventory file. 3. Install Actual results: Availability check failed, failed to check images. docker-registry / etc / ... Expected results: Continue to install. Additional info: Checking the first server where the installer hangs, you can see it testing with skopeo --tls-verify=false docker://registry.access.redhat.com.... this check does not take the proxy settings from the docker daemon, and neither does it take other proxy settings. Two ways to overcome this.. 1. Set proxy settings in the profile on all servers. 2. Make install script set proxy settings before running the check. I dont have a way of uploading logs from this customer i am at.
This is a separate bug for the problem noted in https://bugzilla.redhat.com/show_bug.cgi?id=1480195#c6
https://github.com/openshift/openshift-ansible/pull/6716 addresses this for 3.9. Once that is approved I'll look into backports for 3.7 and 3.6 and/or adjust the release version here.
Fixes for this are going into master and 3.7. https://bugzilla.redhat.com/show_bug.cgi?id=1499358 tracks that. Backporting to 3.6 would be a little more tricky at this point. *** This bug has been marked as a duplicate of bug 1499358 ***