Some news, I found a way to reach a bit further my goal.
Instead of deploying Ingress by the Helm chart, I re-deployed it manually :
kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/master/deploy/mandatory.yaml
Then for last step I edited the service-nodeport.yaml by adding the ips of my nodes :
curl https://raw.githubusercontent.com/kubernetes/ingress-nginx/master/deploy/provider/baremetal/service-nodeport.yaml -o service-nodeport.yaml
kubectl get -o jsonpath='{.status.addresses[?(@.type=="InternalIP")].address}' $(kubectl get nodes -o name)
Add to service-nodeport.yaml :
externalIPs:
- X.X.X.X
- X.X.X.X
Then :
kubectl apply -f service-nodeport.yaml
It worked and the Ingress listens on all my nodes on 80 and 433 now.
But I think this way may not be the cleanest one ... Any advice/remark ?
Thanks,