Description of problem: The config playbook (which should be idempotent and usable to keep your environment in a desired state) will fully upgrade container versions to whatever the latest in configured registries is. i.e. a 3.1 installation, once 3.2 were released, might run the config playbook before they're ready to actually do an upgrade. Version-Release number of selected component (if applicable): openshift-ansible-3.0.86-1 How reproducible: I think 100% if setup is followed. (which is a bit complicated today) Steps to Reproduce: 1. Install 3.1 containerized environment. 2. Simulate the release of 3.2 images by adding a custom registry that has 3.2.x images as the latest pushed. (QE or brew registries will do if you don't want to set one up yourself, provided the most recent push was 3.2) 3. Re-run playbooks/byo/openshift-cluster/config.yml Actual results: Before re-running config playbook: [cloud-user@ip-172-18-5-237 ~]$ docker ps CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES 55a8b643a984 openshift3/node:v3.1.1.6 "/usr/local/bin/origi" 7 minutes ago Up 7 minutes atomic-openshift-node 10ca26983028 openshift3/openvswitch:v3.1.1.6 "/usr/local/bin/ovs-r" 7 minutes ago Up 7 minutes openvswitch 4326ee0e5da9 openshift3/ose:v3.1.1.6 "/usr/bin/openshift s" 7 minutes ago Up 7 minutes atomic-openshift-master After: [cloud-user@ip-172-18-5-237 ~]$ docker ps CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES ec3e9aff1f2c openshift3/node:v3.2.0.20 "/usr/local/bin/origi" About an hour ago Up About an hour atomic-openshift-node 8d8255a9f361 openshift3/openvswitch:v3.1.1.6 "/usr/local/bin/ovs-r" About an hour ago Up About an hour openvswitch 6ea4ee4a3626 openshift3/ose:v3.2.0.20 "/usr/bin/openshift s" About an hour ago Up About an hour atomic-openshift-master Expected results: The images should NOT have been upgraded as we did not run an upgrade playbook. Additional info: Please ignore the fact that openvswitch was not upgraded, this is a separate bug, openvswitch does not get upgraded even if you run the upgrade playbook in some situations.
https://github.com/openshift/openshift-ansible/pull/1859/files
The fix works well, the image versions wasn't upgraded when run the playbook again.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2016:1208