CNI broken in containerized install, pods fail to start

[sjenning]

$ oc version
oc v3.6.0-rc.0+245158d-9
kubernetes v1.6.1+5115d708d7
features: Basic-Auth GSSAPI Kerberos SPNEGO

openshift v3.6.0-rc.0+245158d-9
kubernetes v1.6.1+5115d708d7

E0714 15:23:23.974280   16573 cni.go:257] Error adding network: failed to find plugin "loopback" in path [/opt/loopback/bin /opt/cni/bin]
...
W0714 15:23:24.140803   16573 docker_sandbox.go:263] NetworkPlugin cni failed on the status hook for pod "docker-registry-1-deploy_default": Unexpected command output nsenter: cannot open : No such file or directory
...
E0714 16:00:34.359759   16573 cni.go:275] Error deleting network: failed to find plugin "openshift-sdn" in path [/opt/cni/bin /opt/openshift-sdn/bin]

exec'ing into the node container, /opt is indeed empty.

[pweil-]

@dcbw setting to p1 for triage, feel free to lower if this isn't release blocking.

[sosiouxme]

Is there a workaround for this? My v3.6.0-rc.0 install is dead in the water... does latest do any better?

[eparis]

fixed by: #15542