Whitespaces in 'addresses' configuration property

[gprossliner]

The bug itself, as detailed as you can:

When using this configuration ('config'), all works as expected:

address-pools:

- name: ingress
  protocol: layer2
  addresses:
    - 10.0.0.80-10.0.0.80

- name: default
  protocol: layer2
  addresses:
    - 10.0.0.81-10.0.0.99

But when there are spaces, metallb is unable to provide an IP, so the EXTERNAL-IP keeps :

address-pools:

- name: ingress
  protocol: layer2
  addresses:
    - 10.0.0.80 - 10.0.0.80

- name: default
  protocol: layer2
  addresses:
    - 10.0.0.81 - 10.0.0.99

The logs of the speaker shows the error:

{
  "caller": "k8s.go:361",
  "configmap": "kube-metallb/config",
  "error": "parsing address pool #2: invalid CIDR \"10.0.0.81 - 10.0.0.99\" in pool \"default\"",
  "event": "configStale",
  "msg": "config (re)load failed, config marked stale",
  "ts": "2019-11-22T12:19:02.449737153Z"
}

Version of MetalLB

$ kubectl -n kube-metallb get pod -ojson | jq -r ".items[].spec.containers[].image"
metallb/controller:v0.8.2
metallb/speaker:v0.8.2
metallb/speaker:v0.8.2

Version of Kubernetes

$ kubectl version -ojson | jq -r ".serverVersion.gitVersion"`
v1.16.0

Name and version of network addon (e.g. Calico, Weave...)

$ kubectl -n kube-flannel get pod -ojson | jq -r ".items[].spec.containers[].image"
quay.io/coreos/flannel:v0.11.0-amd64
quay.io/coreos/flannel:v0.11.0-amd64

Whether you've configured kube-proxy for iptables or ipvs mode

$ kubectl -n kube-system logs kube-proxy-9jxfn
...
server_others.go:149] Using iptables Proxier.
...

[gprossliner]

Thank you!