Oleksandr Vorona

As a work-around I had to re-install hcloud CCM every time I add worker ``` hetzner-kube cluster addon uninstall hcloud-controller-manager -n CLUSTER-NAME hetzner-kube cluster addon install hcloud-controller-manager -n CLUSTER-NAME ```

@JohnnyQQQQ I got ingress working via ``` helm install stable/nginx-ingress --name ingress --namespace ingress --set rbac.create=true,controller.kind=DaemonSet,controller.service.type=ClusterIP,controller.service.externalIPs='{1.2.3.4,5.6.7.8}',controller.stats.enabled=true,controller.metrics.enabled=true ``` not a `.[0]` @cornelius-keller what about using [aledbf/kube-keepalived-vip](https://github.com/aledbf/kube-keepalived-vip) as base image, it looks...

@kaosmonk You can try [hetzner-failover-ip helm-chart](https://github.com/exocode/helm-charts/tree/master/hetzner-failover-ip), example NodeSelector with multiple IPs and nginx-ingress.

@kaosmonk it's a number of keepalived pods to be spawned across k8s nodes. I don't see the reason to use any other value than 2 . If you need faster...

@monofone > There is an option --iface to flanneld which can take an Interface name. I tried `--iface wg0` without success, no internal traffic worked. Didn't dig into it unfortunately.

Looks like I found a work-around * Add `--iface wg0` (2 lines) to flannel command line in kube-system flannel DaemonSet, kill all flannel pods to apply ``` echo 'spec: template:...

>I like your approach. Could you explain what's happening here? As @monofone mentioned, there is enough to specify `--iface wg0` arg to `flanneld` to get flannel to work via `wg0`...

It runs on every `flannel.*` interface creation. Thus it persists after reboots too, as flannel starts on every k8s node boot and creates it's interface.

IMHO it would be great to have an ability to specify custom post-install script to run on all servers. This may provide required flexibility and simplicity.

Let me describe our expirience how to get synced node w/o issues. Key point is: * Use *low-latency* disks(or terabytes of RAM for cache), 1.5TB+. This is **really** important. Generic...