0

When releasing an update, it all works fine (pods are destroyed one by one and the new version fired up), but if I refresh the website during a rolling update sometime I get a server unreachable error. I guess it is due to the fact the load balancer still trying to send traffic to a pod which probably in the terminating state.

How can I resolve this issue? Should I add a retry policy to the Traefik Ingressroute or is there a way to force the pod to be removed from the load balancer before anything?

Sugafree
  • 631
  • 2
  • 14
  • 30
  • How do you perform rolling update? Do you have a deployment related to the service exposed by traefik? – Federico Paparoni Nov 19 '20 at 19:13
  • Yes, I have a statefulSet yaml file (5 replicas) with a simple ClusterIp service exposing it within the cluster and traefil has a LB with a ingressroute controller and a route that will direct the traffic to the specific ClusterIp type service – Sugafree Nov 19 '20 at 19:30
  • Ok for statefulset you can see this answer https://stackoverflow.com/a/54310540/4895267 – Federico Paparoni Nov 19 '20 at 20:13

0 Answers0