Origin: Builders not pushing images to integrated registry / Cannot resolve docker-registry.default.svc

[jtcressy]

On a fresh install using openshift-ansible (branch release-3.6), builds fail to push images to the docker-registry. I've already tried re-deploying the registry, restarting the masters, and output from oc adm diagnostics is attached below.

Version

3.6.1

Current Result

Builder pods cannot resolve docker-registry.default.svc thus can't push images.
Builder pods cannot push to docker-registry.default.svc because the registry is not serving https on port 5000

Expected Result

Builder pods push images to registry and registry is serving https at port 5000.

Additional Information

Output of oc adm diagnostics: sprunge paste
Build logs with BUILD_LOGLEVEL=5: accountmanager-7.log

EDIT:
Re-attached registry-certificates and registry-config to the registry pod, which solves the problem with the registry not serving https, but I still get the 'no such host' error.
Updated builder log with BUILD_LOGLEVEL=5: accountmanager-8.log
Log summary: F1120 18:58:15.463499 1 helpers.go:119] error: build error: Failed to push image: Get https://docker-registry.default.svc:5000/v1/_ping: dial tcp: lookup docker-registry.default.svc on 192.168.246.107:53: no such host

[jtcressy]

After digging through https://github.com/openshift/origin/issues/16097
and https://github.com/openshift/openshift-ansible/issues/5466

I solved it by running this on every host (master, infra and nodes)

echo "search cluster.local" >> /etc/resolv.conf

Now i'm just really hoping this sticks after a restart.

[marthenlt]

Encounter the same thing, but actually you don't have to add cluster.local manually in /etc/resolv.conf

All you need to do if you're running RHEL like myself is to restart your NetworkManager and it will auto populate the correct search value with proper nameserver IP.

systemctl restart NetworkManager