Network connectivity flake during GCE installation #17187
Description
This error looks new since switching to bootstrapping - the network was temporarily unavailable after the API had been reported successfully up (and other actions ran?). At this point in the install there are no nodes active (waiting for bootstrapping) so it seems unlikely this is related to the openshift-sdn component on the master. Unfortunately we don't have master logs from the install portion yet on GCE.
TASK [openshift_hosted : Create default projects] ******************************
Saturday 04 November 2017 20:29:25 +0000 (0:00:07.580) 0:16:35.031 *****
changed: [ci-prtest-5a37c28-10903-ig-m-bx69] => (item={'key': u'default', 'value': {u'default_node_selector': u''}})
failed: [ci-prtest-5a37c28-10903-ig-m-bx69] (item={'key': u'logging', 'value': {u'default_node_selector': u''}}) => {"changed": false, "failed": true, "item": {"key": "logging", "value": {"default_node_selector": ""}}, "msg": {"cmd": "/bin/oc adm new-project logging --admin-role=admin --node-selector=", "results": {}, "returncode": 1, "stderr": "The connection to the server internal-api.prtest-5a37c28-10903.origin-ci-int-gce.dev.rhcloud.com:8443 was refused - did you specify the right host or port?\n", "stdout": ""}}
failed: [ci-prtest-5a37c28-10903-ig-m-bx69] (item={'key': u'openshift-infra', 'value': {u'default_node_selector': u''}}) => {"changed": false, "failed": true, "item": {"key": "openshift-infra", "value": {"default_node_selector": ""}}, "msg": {"cmd": "/bin/oc adm new-project openshift-infra --admin-role=admin --node-selector=", "results": {}, "returncode": 1, "stderr": "The connection to the server internal-api.prtest-5a37c28-10903.origin-ci-int-gce.dev.rhcloud.com:8443 was refused - did you specify the right host or port?\n", "stdout": ""}}