Description of problem: When OpenShift node runs on OpenStack, it can be notified about it by setting an authentication URL and creds. In that case it is trying to guess the instance id of the VM by searching for instance with instance name same as the hostname. There however can be more of such instances if more of same (heat) stacks of openshift are running under same tenant (while running in completely different networks). For example a staging and production environment. There is a better way to get the instance id -- use the uuid from the metadata : [cloud-user@ose3-master-h57q6 ~]$ curl 'http://169.254.169.254/openstack/latest/meta_data.json' | json_reformat % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 1711 100 1711 0 0 21236 0 --:--:-- --:--:-- --:--:-- 21387 { "random_seed": "...", "uuid": "44447f54-fed0-4d4a-9fed-b7cae0b610ce", "availability_zone": "nova", "keys": [ { "data": "...", "type": "ssh", "name": "jenkins" } ], "hostname": "ose3-master-h57q6", "launch_index": 0, "public_keys": { "jenkins": "ssh-rsa ..." }, "name": "ose3-master-h57q6" } Version-Release number of selected component (if applicable): kubernetes.x86_64 1.2.0-0.13.gitec7364b.el7 rhel-extras-candidate kubernetes-client.x86_64 1.2.0-0.13.gitec7364b.el7 rhel-extras-candidate kubernetes-master.x86_64 1.2.0-0.13.gitec7364b.el7 rhel-extras-candidate kubernetes-node.x86_64 1.2.0-0.13.gitec7364b.el7 rhel-extras-candidate How reproducible: always Steps to Reproduce: 1. Boot the stack, rename the instance to some name. 2. start the atomic-openshift-node 3. Actual results: # journalctl -xe zář 13 01:23:19 ose3-master-h57q6 atomic-openshift-node[50596]: I0913 01:23:19.645808 50596 openstack.go:289] Claiming to support Instances zář 13 01:23:19 ose3-master-h57q6 atomic-openshift-node[50596]: E0913 01:23:19.695103 50596 kubelet.go:1131] Unable to construct api.Node object for kubelet: failed to get external ID from cloud provider: Failed to find object Expected results: No complains about the failure to get external ID. Additional info:
This is a dup of 1321964. This is a known point of friction. Docs do state that the nodeName must equal the instance name for the Openstack integration to work: https://docs.openshift.org/latest/install_config/configuring_openstack.html#openstack-configuring-nodes-manually We are working on this limitation: https://bugzilla.redhat.com/show_bug.cgi?id=1367201 *** This bug has been marked as a duplicate of bug 1321964 ***