Comment # 25 on bug 1197490 from Robert Munteanu

I ran a full test again, with a fresh VM. The problem seems fixed for me. I
used
registry.opensuse.org/home/favogt/boo1197490/container/kubic/weave-kube:2.8.1-21.16
for my tests, selected by checking the build logs from
https://build.opensuse.org/package/show/home:favogt:boo1197490/kubic-weave-kube-image.

VM details:

$ cat /etc/os-release
NAME="openSUSE MicroOS"
# VERSION="20220330"
ID="opensuse-microos"
ID_LIKE="suse opensuse opensuse-tumbleweed"
VERSION_ID="20220330"
PRETTY_NAME="openSUSE MicroOS"
ANSI_COLOR="0;32"
CPE_NAME="cpe:/o:opensuse:microos:20220330"
BUG_REPORT_URL="https://bugs.opensuse.org"
HOME_URL="https://www.opensuse.org/"
DOCUMENTATION_URL="https://en.opensuse.org/Portal:MicroOS"
LOGO="distributor-logo-MicroOS

I ran the usual kubeadm setup commands

$ kubeadm init
$ mkdir -p $HOME/.kube
$ sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
$ sudo chown $(id -u):$(id -g) $HOME/.kube/config

And then overrode the weave patch as follows

$ cp /usr/share/k8s-yaml/weave/weave.yaml weave.yaml
$ cat >> kustomization.yaml
apiVersion: kustomize.config.k8s.io/v1beta1
kind: Kustomization

images:
- name: registry.opensuse.org/kubic/weave-kube
  newName:
registry.opensuse.org/home/favogt/boo1197490/container/kubic/weave-kube
  newTag: 2.8.1-21.16

resources:
- weave.yaml

# optional, for checking the manifests - kubectl build .
$ kubectl apply -k .

Everything works fine from a networking POV:

$ kubectl get pod -A -o wide
NAMESPACE     NAME                              READY   STATUS    RESTARTS     
  AGE     IP            NODE      NOMINATED NODE   READINESS GATES
kube-system   coredns-fc8b57f45-r7kc2           1/1     Running   0            
  11m     10.32.0.3     kubic-0   <none>           <none>
kube-system   coredns-fc8b57f45-wmshw           1/1     Running   0            
  11m     10.32.0.2     kubic-0   <none>           <none>
kube-system   etcd-kubic-0                      1/1     Running   0            
  11m     10.16.0.208   kubic-0   <none>           <none>
kube-system   kube-apiserver-kubic-0            1/1     Running   0            
  11m     10.16.0.208   kubic-0   <none>           <none>
kube-system   kube-controller-manager-kubic-0   1/1     Running   0            
  11m     10.16.0.208   kubic-0   <none>           <none>
kube-system   kube-proxy-rrhkb                  1/1     Running   0            
  11m     10.16.0.208   kubic-0   <none>           <none>
kube-system   kube-scheduler-kubic-0            1/1     Running   0            
  11m     10.16.0.208   kubic-0   <none>           <none>
kube-system   weave-net-wg4hq                   2/2     Running   1 (4m14s ago)
  4m16s   10.16.0.208   kubic-0   <none>           <none>

The warnings about failing to load modules are still there, but they don't seem
to matter:

$ kubectl -n kube-system logs daemonset/weave-net -c weave-init
modprobe: can't load module nfnetlink (kernel/net/netfilter/nfnetlink.ko.zst):
invalid module format
Ignore the error if "xt_set" is built-in in the kernel

In case it matters, I am creating the VMs using a patched version of
https://github.com/kubic-project/kubic-terraform-kvm/ .