I’m doing the HA install of Rancher 2.x using cert-manager and nginx-ingress.
The steps I did:
- Create DO cluster with 3 nodes
- Install cert-manager
- Install nginx-ingress
- Install rancher
All those installs were based on the helm charts.
All works great, I can see the “local” cluster in Rancher interface. The service is stable: I can see all 3 hosts as healthy in the load balancer (the DO load-balancer).
Moving forward, the next step:
- Add another DO k8s cluster to Rancher (using the import feature)
After doing this, Rancher starts to fail the health checks and the k8s API stops working correctly (“kubectl get all” fails). Trying to use the rancher interface fails with intermittent 5xx responses.
Do you have any ideas of what might be causing this?