MetalLB - how to troubleshoot - no availble IPs

General Discussions
1 
### Cluster information:

Kubernetes version: **1.18.2**
Cloud being used: **Bare-Metal**
Installation method: **Ansible - https://github.com/zimmertr/Bootstrap-Kubernetes-with-QEMU**
Host OS: **Debian 10**
CNI and version: **Calico - v3.8.8-1**
CRI and version: Docker - **19.03.8, build afacb8b7f0**

Hi,

I’m trying to get MetalLB to work with my setup.
pfSense: OpenBGBD enabled and status says it sending messages to the workers.

I have two VLANs one for Kubernetes and one for MetalLB deployments:

  • 10.64.64.0/24 - Kuberentes net
  • 10.65.65.0/24 - MetalLB net

I am running OpenBGBD on a virtual router IP adress: 10.1.10.2.
I can reach 10.1.10.2 port 179/tcp from my workers so no FW issue from what I can gather.

My MetalLB config:

apiVersion: v1
kind: ConfigMap
metadata:
  namespace: metallb-system
  name: config
data:
  config: |
    peers:
    - my-asn: 64513
      peer-asn: 64512
      peer-address: 10.1.10.2
    address-pools:
    - name: my-ip-space
      protocol: bgp
      addresses:
      - 10.65.65.0/24

The logs say:
{"caller":"service.go:90","error":"no available IPs","msg":"IP allocation failed","op":"allocateIP","service":

Now the workers only have one network card and that is connected to the 10.64.64.0/24 net. Is that correct or should I be adding a 2nd network adapter and connect it to the 10.65.65.0/24 net as well?

2

Hey @CluelessTechnologist. Just found this post while Googling my username. I am the person that wrote the project you used to deploy Kubernetes and probably MetalLB. Feel free to send me an email or open a GitHub issue on the project if you still need help. :slight_smile:

tj@tjzimmerman.com

Also, for what it is worth, I have deprecated this project in favor of TKS.