Description of problem: Couldn't upgrade from v3.0.0.1 to v3.0.2.0, the upgrade failed at step "oadm policy reconcile-cluster-roles --confirm" Version-Release number of selected component (if applicable): atomic-openshift-utils-3.0.40-1.git.22.da21865.el7aos.noarch How reproducible: always Steps to Reproduce: 1.ansible-playbook -i /root/config/ose3020 /usr/share/ansible/openshift-ansible/playbooks/byo/openshift-cluster/upgrades/v3_0_minor/upgrade.yml 2. 3. ansible-playbook -i /root/config/ose3020 /usr/share/ansible/openshift-ansible/playbooks/byo/openshift-cluster/upgrades/v3_0_minor/upgrade.yml TASK: [oadm policy reconcile-cluster-roles --confirm] ************************* failed: [10.14.6.120] => {"changed": true, "cmd": ["oadm", "--config=/etc/openshift/master/admin.kubeconfig", "policy", "reconcile-cluster-roles", "--confirm"], "delta": "0:00:00.044817", "end": "2016-02-15 18:04:48.997194", "rc": 1, "start": "2016-02-15 18:04:48.952377", "warnings": []} stderr: Error: unknown flag: --confirm Run 'oadm help' for usage. FATAL: all hosts have already failed -- aborting PLAY RECAP ******************************************************************** to retry, use: --limit @/root/upgrade.retry 10.14.6.120 : ok=323 changed=33 unreachable=0 failed=1 10.14.6.126 : ok=113 changed=13 unreachable=0 failed=0 localhost : ok=25 changed=0 unreachable=0 failed=0 Expected results: Additional info:
Wow, great catch. I suspect at this point most customers are upgrading directly to 3.1 which is why this was never reported. Looking at git it seems this has been broken since Nov. I think we should fix this but it's definitely not a 3.2 blocker.