No balancing when node goes down

[vinid223]

Hello,

I have an issue when one node is suddently not available and the load balancer is just not working for any deployments on the cluster. Even if the deployments run on other nodes, they are unreachable until the node is back up.

I am running a k3s cluster of 5 Raspberry Pi.

I don't know what to provide for the issue, but I am available to diagnose

[cawoodm]

We are seeing the same issue with 2 K3S nodes (1 master, 1 worker). If the worker goes down and we ping a service on the UP node we are seeing 50% failures as ingress continues directing traffic round-robin to both nodes. We have all our pods (incl coredns and traefik) on both nodes so we expect 100% availability when a node goes down.

StackOverflow