Description of problem: When trying to add a new node to a pre-installed env using atomic-openshift-installer, the installer changed the node configuration of installed nodes. [root@openshift-v3 node]# diff /etc/origin/node/node-config.yaml /etc/origin/node/node-config.yaml.2015-11-09\@17:19:04~ 8c8 < format: openshift3/ose-${component}:${version} --- > format: registry.access.redhat.com/openshift3/ose-${component}:${version} 12c12 < networkPluginName: redhat/openshift-ovs-subnet --- > networkPluginName: redhat/openshift-ovs-multitenant The new node-config.yaml is using some default value the for the parameters. Version-Release number of selected component (if applicable): atomic-openshift-utils-3.0.7-1.git.131.21caa21.el7aos.noarch How reproducible: Always Steps to Reproduce: 1. Prepare an ose env with oreg_url=registry.access.redhat.com/openshift3/ose-${component}:${version} and os_sdn_network_plugin_name=redhat/openshift-ovs-multitenant 2. Add a new node to this env using 'atomic-openshift-installer install' 3. After installation of the node, check the configuration of preexisting nodes. Actual results: registry_url and sdn_network_plugin_name in node-config.yaml were changed. Expected results: Should not make changes to the currently running environment. Additional info:
https://github.com/openshift/openshift-ansible/pull/851 which should address this.
Verify this bug with atomic-openshift-utils-3.0.7-1.git.156.70ed54c.el7aos.noarch. After adding new nodes to pre-installed env, the configuration on pre-installed env is not modified. And the new nodes have the default configuration. Move this to VERIFIED.
Sorry for the previous comment, with further testing, found the installer will also rewrite the node configuration on the all-in-one node. When a pre-existing node also act as master, this host would be added into the node list in the inventory, and rewritten along with the new nodes.