Created attachment 1286945 [details] Ansible log Description of problem: When trying to deploy metrics during installation time and using the advanced method, it fails with the error message: The conditional check 'persistent_volumes | length > 0 or persistent_volume_claims | length > 0' failed. The error was: {{ hostvars[groups.o o_first_master.0] | oo_persistent_volumes(groups) }}: |failed invalid storage kind 'emptydir' for component 'metrics' The error appears to have been in '/usr/share/ansible/openshift-ansible/roles/openshift_facts/tasks/main.yml': line 2, column 3, but may be elsewhere in the file depending on the exact syntax problem. The offending line appears to be: --- - name: Detecting Operating System ^ here Version-Release number of selected component (if applicable): openshift-ansible-callback-plugins-3.5.71-1.git.0.128c2db.el7.noarch openshift-ansible-playbooks-3.5.71-1.git.0.128c2db.el7.noarch openshift-ansible-3.5.71-1.git.0.128c2db.el7.noarch openshift-ansible-filter-plugins-3.5.71-1.git.0.128c2db.el7.noarch openshift-ansible-roles-3.5.71-1.git.0.128c2db.el7.noarch openshift-ansible-docs-3.5.71-1.git.0.128c2db.el7.noarch openshift-ansible-lookup-plugins-3.5.71-1.git.0.128c2db.el7.noarch How reproducible: Always Steps to Reproduce: 1. Set openshift_hosted_metrics_deploy=true and openshift_master_default_subdomain in the inventory file 2. Run the playbook /usr/share/ansible/openshift-ansible/playbooks/byo/config.yml Actual results: It fails with the error message: "failed invalid storage kind 'emptydir' for component 'metrics'" Expected results: To install openshift and deploy metrics with emptydir storage Additional information: If you install openshift with no metrics and after it you set: openshift_metrics_install_metrics=true openshift_hosted_metrics_storage_kind=emptydir And run the playbook /usr/share/ansible/openshift-ansible/playbooks/byo/openshift-cluster/openshift-metrics.yml Metrics are deployed with emptydir.
Host file: [OSEv3:children] masters nodes [OSEv3:vars] ansible_ssh_user=root deployment_type=openshift-enterprise openshift_master_identity_providers=[{'name': 'htpasswd_auth', 'login': 'true', 'challenge': 'true', 'kind': 'HTPasswdPasswordIdentityProvider', 'filename': '/etc/origin/master/htpasswd'}] openshift_hosted_metrics_deploy=true openshift_master_default_subdomain=apps.test.example.com openshift_hosted_metrics_storage_kind=emptydir [masters] master1.example.com [nodes] master1.example.com node1.example.com openshift_node_labels="{'region': 'primary', 'zone': 'east'}" node2.example.com openshift_node_labels="{'region': 'primary', 'zone': 'west'}"
Not sure if important, but parameter openshift_hosted_metrics_storage_kind=emptydir It is redundant and the behaviour is exactly the same without it.
If you don't set openshift_hosted_metrics_storage_kind at all you should get emptydir. Can you please verify and we'll move this over to a docs bug to clarify.
(In reply to Scott Dodson from comment #9) > If you don't set openshift_hosted_metrics_storage_kind at all you should get > emptydir. Can you please verify and we'll move this over to a docs bug to > clarify. As you can read at https://bugzilla.redhat.com/show_bug.cgi?id=1460594#c3 Not sure if important, but parameter openshift_hosted_metrics_storage_kind=emptydir It is redundant and the behaviour is exactly the same without it.
I can't reproduce this bug with latest versions: atomic-openshift-node-3.5.5.31-1.git.0.b6f55a2.el7.x86_64 atomic-openshift-clients-3.5.5.31-1.git.0.b6f55a2.el7.x86_64 atomic-openshift-master-3.5.5.31-1.git.0.b6f55a2.el7.x86_64 tuned-profiles-atomic-openshift-node-3.5.5.31-1.git.0.b6f55a2.el7.x86_64 atomic-openshift-utils-3.5.91-1.git.0.28b3ddb.el7.noarch atomic-openshift-docker-excluder-3.5.5.31-1.git.0.b6f55a2.el7.noarch atomic-openshift-excluder-3.5.5.31-1.git.0.b6f55a2.el7.noarch atomic-openshift-3.5.5.31-1.git.0.b6f55a2.el7.x86_64 atomic-openshift-sdn-ovs-3.5.5.31-1.git.0.b6f55a2.el7.x86_64 image: registry.access.redhat.com/openshift3/metrics-cassandra:3.5.0 image: registry.access.redhat.com/openshift3/metrics-hawkular-metrics:3.5.0 image: registry.access.redhat.com/openshift3/metrics-heapster:3.5.0 # cat /etc/ansible/hosts [OSEv3:children] masters nodes [OSEv3:vars] ansible_ssh_user=root deployment_type=openshift-enterprise openshift_master_identity_providers=[{'name': 'htpasswd_auth', 'login': 'true', 'challenge': 'true', 'kind': 'HTPasswdPasswordIdentityProvider', 'filename': '/etc/origin/master/htpasswd'}] openshift_hosted_metrics_deploy=true openshift_master_default_subdomain=apps.test.example.com [masters] master35.example.com [nodes] master35.example.com nodeinfra.example.com openshift_node_labels="{'region': 'infra'}" nodeprimary.example.com openshift_node_labels="{'region': 'primary'}"
Created attachment 1309418 [details] ansible logs from customer
Created attachment 1309419 [details] ansible logs from my reproducer
Hi Ruben, Looks like Javier was awaiting feedback on this from engineering?
Latest comment in the case from the customer was that this problem doesn't exist in 3.6. Moving back to Installer but marking CLOSED WONTFIX.