Comment 6 for bug 1740896

<<< Update: After removing the livenessProbe as suggested by ykarel on irc, the dashboard pod has stopped crashing
>>> This was just to check how kubectl apply behaves without liveliness.

<<< The systemd daemon for dashboard refuse to start perhaps because of its type i.e. oneshot. The "kube-dash.service" and "kube-system-namespace.service" are in dead and inactive state.

kube-dash systemd service is just to create service and deployment so it's dead state is expected.

<<< kubectl --namespace=kube-system get ep kubernetes-dashboard -- http://paste.openstack.org/show/640754/
>>> Here endpoint is missing. This is the issue that needs to be fixed.

<<< I also observed that at L106[1], there is no kube-dashboard deployment, this should be changed to "kubernetes-dashboard" in /usr/local/bin/kube-dash

[1] https://github.com/openstack/magnum/blob/stable/pike/magnum/drivers/common/templates/kubernetes/fragments/kube-dashboard-service.sh#L106

>> Yes, it's wrong

Was not able to reproduce it on devstack stable/pike. Trying to setup a osa setup to look into it.