Red Hat Bugzilla – Bug 1492786
Installer fails at Create OpenShift router step
Last modified: 2017-11-28 17:11:25 EST
Created attachment 1327518 [details] ansible.log Description of problem: Installation of OCP fails when giving a wildcard certificate to the installer. Version-Release number of the following components: openshift-ansible-3.6.173.0.21-2.git.0.44a4038.el7.noarch ansible-2.3.2.0-2.el7.noarch How reproducible: Every time Steps to Reproduce: 1. Deploy OCP with openshift_hosted_router_certificate={"certfile": "/root/wildcard.apps.lab.kupo.se.crt", "keyfile": "/root/wildcard.apps.lab.kupo.se.key", 'cafile': '/root/ca-chain.pem' } 2 Actual results: TASK [openshift_hosted : Create OpenShift router] ************************************************************************************************************ Monday 18 September 2017 09:28:38 +0000 (0:00:01.519) 1:07:19.473 ****** failed: [master1.lab.kupo.se] (item={u'name': u'router', u'certificate': {'keyfile': u'/root/wildcard.apps.lab.kupo.se.key', 'certfile': u'/root/wildcard.apps.lab.kupo.se.crt', 'cafile': u'/root/ca-chain.pem'}, u'replicas': u'3', u'serviceaccount': u'router', u'namespace': u'default', u'stats_port': 1936, u'edits': [{u'action': u'put', u'value': 1, u'key': u'spec.strategy.rollingParams.intervalSeconds'}, {u'action': u'put', u'value': 1, u'key': u'spec.strategy.rollingParams.updatePeriodSeconds'}, {u'action': u'put', u'value': 21600, u'key': u'spec.strategy.activeDeadlineSeconds'}], u'images': u'registry.lab.kupo.se:5000/openshift3/ose-${component}:${version}', u'selector': u'infra=true', u'ports': [u'80:80', u'443:443']}) => { "failed": true, "item": { "certificate": { "cafile": "/root/ca-chain.pem", "certfile": "/root/wildcard.apps.lab.kupo.se.crt", "keyfile": "/root/wildcard.apps.lab.kupo.se.key" }, "edits": [ { "action": "put", "key": "spec.strategy.rollingParams.intervalSeconds", "value": 1 }, { "action": "put", "key": "spec.strategy.rollingParams.updatePeriodSeconds", "value": 1 }, { "action": "put", "key": "spec.strategy.activeDeadlineSeconds", "value": 21600 } ], "images": "registry.lab.kupo.se:5000/openshift3/ose-${component}:${version}", "name": "router", "namespace": "default", "ports": [ "80:80", "443:443" ], "replicas": "3", "selector": "infra=true", "serviceaccount": "router", "stats_port": 1936 }, "module_stderr": "Shared connection to master1.lab.kupo.se closed.\r\n", "module_stdout": "Traceback (most recent call last):\r\n File \"/tmp/ansible_FoO2lC/ansible_module_oc_adm_router.py\", line 3185, in <module>\r\n main()\r\n File \"/tmp/ansible_FoO2lC/ansible_module_oc_adm_router.py\", line 3176, in main\r\n results = Router.run_ansible(module.params, module.check_mode)\r\n File \"/tmp/ansible_FoO2lC/ansible_module_oc_adm_router.py\", line 3093, in run_ansible\r\n api_rval = ocrouter.create()\r\n File \"/tmp/ansible_FoO2lC/ansible_module_oc_adm_router.py\", line 2895, in create\r\n self.needs_update()\r\n File \"/tmp/ansible_FoO2lC/ansible_module_oc_adm_router.py\", line 2940, in needs_update\r\n not Utils.check_def_equal(self.prepared_router['ServiceAccount']['obj'].yaml_dict,\r\n File \"/tmp/ansible_FoO2lC/ansible_module_oc_adm_router.py\", line 2665, in prepared_router\r\n results = self._prepare_router()\r\n File \"/tmp/ansible_FoO2lC/ansible_module_oc_adm_router.py\", line 2825, in _prepare_router\r\n rfd.write(open(self.config.config_options['cert_file']['value']).read())\r\nIOError: [Errno 2] No such file or directory: '/etc/origin/master/wildcard.apps.lab.kupo.se.crt'\r\n", "rc": 0 } MSG: MODULE FAILURE Expected results: The installer should copy the file to the needed location. Additional info: The same inventory file worked with v3.6.173.0.5
Johan, Thanks for filing your issue. I have a few questions so that we can get on the same page. 1. Do you have any further logs from the installation regarding the openshift_hosted role and the router.yml installation? The previous step is what copies the certificates from the controller host onto the server that is receiving openshift. These are the lines I'm referring to: https://github.com/openshift/openshift-ansible/blob/master/roles/openshift_hosted/tasks/router/router.yml#L21-L28 Could you find the logs for this step? Was this step successful? 2. The error from the router explains that it was looking for the certificate but it was not copied to the destination upon where the oc_adm_router module was expecting it to be (/etc/origin/master/wildcard.apps.lab.kupo.se.crt) Could you verify the contents of /etc/origin/master/ to see if the certificates are where they were copied from step #1? This will help us know if they copy step was successful. 3. Are the following inventory variables set? openshift_hosted_router_create_certificate openshift_hosted_router_certificate Thanks.
(In reply to Kenny Woodson from comment #1) > Johan, > > Thanks for filing your issue. I have a few questions so that we can get on > the same page. > > 1. Do you have any further logs from the installation regarding the > openshift_hosted role and the router.yml installation? > > The previous step is what copies the certificates from the controller host > onto the server that is receiving openshift. These are the lines I'm > referring to: > https://github.com/openshift/openshift-ansible/blob/master/roles/ > openshift_hosted/tasks/router/router.yml#L21-L28 > > Could you find the logs for this step? Was this step successful? > Is this the one you're looking for? Seems like it skips the step for some reason. 2017-09-18 15:32:14,215 p=6001 u=root | TASK [openshift_hosted : Get the certificate contents for router] ******************************************************************************************** 2017-09-18 15:32:14,215 p=6001 u=root | task path: /usr/share/ansible/openshift-ansible/roles/openshift_hosted/tasks/router/router.yml:42 2017-09-18 15:32:14,216 p=6001 u=root | Monday 18 September 2017 15:32:14 +0000 (0:00:00.069) 0:53:16.629 ****** 2017-09-18 15:32:14,276 p=6001 u=root | skipping: [master1.lab.kupo.se] => (item=/root/wildcard.apps.lab.kupo.se.key) => { "changed": false, "item": "/root/wildcard.apps.lab.kupo.se.key", "skip_reason": "Conditional result was False", "skipped": true } 2017-09-18 15:32:14,294 p=6001 u=root | skipping: [master1.lab.kupo.se] => (item=/root/ca-chain.pem) => { "changed": false, "item": "/root/ca-chain.pem", "skip_reason": "Conditional result was False", "skipped": true } 2017-09-18 15:32:14,295 p=6001 u=root | skipping: [master1.lab.kupo.se] => (item=/root/wildcard.apps.lab.kupo.se.crt) => { "changed": false, "item": "/root/wildcard.apps.lab.kupo.se.crt", "skip_reason": "Conditional result was False", "skipped": true } > 2. The error from the router explains that it was looking for the > certificate but it was not copied to the destination upon where the > oc_adm_router module was expecting it to be > (/etc/origin/master/wildcard.apps.lab.kupo.se.crt) Could you verify the > contents of /etc/origin/master/ to see if the certificates are where they > were copied from step #1? This will help us know if they copy step was > successful. > They are not there, as they were skipped in the previous step. > 3. Are the following inventory variables set? > openshift_hosted_router_create_certificate > openshift_hosted_router_certificate > openshift_hosted_router_certificate is set, openshift_hosted_router_create_certificate is not set.
Proposed fix: https://github.com/openshift/openshift-ansible/pull/5465
Reproduce this bug with a rpm install using openshift-ansible-3.7.0-0.125.0.git.0.91043b6.el7.noarch, verified this bug with openshift-ansible-3.7.0-0.143.3.git.0.6c65767.el7.noarch, and PASS.
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/RHSA-2017:3188