Description of problem: After upgrade v3.9 to v3.10, original registry address return back to ip of docker registry which should be dns address before upgrade. //before upgrade # oc get is NAME DOCKER REPO TAGS UPDATED registry-console docker-registry.default.svc:5000/default/registry-console v3.9 //after upgrade # oc get is NAME DOCKER REPO TAGS UPDATED registry-console 172.30.38.105:5000/default/registry-console v3.9 # oc get dc docker-registry -o yaml |grep -A 1 "REGISTRY_OPENSHIFT_SERVER_ADDR" - name: REGISTRY_OPENSHIFT_SERVER_ADDR value: docker-registry.default.svc:5000 Version-Release number of the following components: openshift-ansible-3.10.0-0.27.0.git.0.abed3b7.el7.noarch How reproducible: always Steps to Reproduce: 1. Upgrade v3.9 to v3.10 2. 3. Actual results: Images was pull/push by ip address of docker-reregistry. Expected results: Images should be pull/push by dns address of docker-reregistry. Additional info: Please attach logs from ansible-playbook with the -vvv flag
We're likely losing REGISTRY_OPENSHIFT_SERVER_ADDR in /etc/origin/master/master.env during upgrade for some reason. Also older variables may have been deprecated. Need to make sure that we're no longer using deprecated variables and get the copy from /etc/sysconfig/atomic-openshift-master-api -> /etc/origin/master/master.env
Marking https://bugzilla.redhat.com/show_bug.cgi?id=1579227 as a duplicate. In 1579227 (3.9 -> 3.10 update with Service Catalog) the credentials from atomic-openshift-master-controllers are valid but the credentials found in master.env are different/invalid. The belief is the fix for 157608 will update master.env with correct values.
*** Bug 1579227 has been marked as a duplicate of this bug. ***
*** This bug has been marked as a duplicate of bug 1581876 ***