When using OVN, LoadBalancer members cannot access themselves through the LoadBalancer VIP. Thus, if kuryr-controller pod (which uses host-networking) is running on a master node (where the OpenShift API is running), then the kuryr-controller cannot access the OpenShift API
Verified on openshift-ansible-3.11.141 on top of OSP 13 2019-06-25.1 puddle. Kuryr-controller pod runs on an infra node: [openshift@master-0 ~]$ oc get pods -n kuryr -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE kuryr-cni-ds-6h8q2 2/2 Running 1 2h 192.168.99.16 master-0.openshift.example.com <none> kuryr-cni-ds-gnr79 2/2 Running 1 2h 192.168.99.7 infra-node-2.openshift.example.com <none> kuryr-cni-ds-j9pcr 2/2 Running 0 2h 192.168.99.14 app-node-1.openshift.example.com <none> kuryr-cni-ds-n4klv 2/2 Running 0 2h 192.168.99.6 app-node-0.openshift.example.com <none> kuryr-cni-ds-sks7m 2/2 Running 1 2h 192.168.99.9 master-1.openshift.example.com <none> kuryr-cni-ds-tcvk8 2/2 Running 0 2h 192.168.99.30 infra-node-0.openshift.example.com <none> kuryr-cni-ds-x6f66 2/2 Running 2 2h 192.168.99.23 master-2.openshift.example.com <none> kuryr-cni-ds-zg5j2 2/2 Running 0 2h 192.168.99.12 infra-node-1.openshift.example.com <none> kuryr-controller-7cf75d55c9-fnq9q 1/1 Running 3 2h 192.168.99.7 infra-node-2.openshift.example.com <none>
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2019:2816