i running them in a ethernet,
usering flannel's node vip , three nodes ping each other with no problem.
my software version: k8s 1.6.6, nginx-ingress-controller -0.9.0-beta.8
I0622 09:23:29.778408 7 controller.go:418] ingress backend successfully reloaded...
2017/06/22 09:24:59 [error] 993#993: *1 upstream timed out (110: Connection timed out) while connecting to upstream, client: 192.168.1.232, server: 0.0.0.0:8000, upstream: "10.1.17.2:8000", bytes from/to client:0/0, bytes from/to upstream:0/0
2017/06/22 09:25:59 [error] 993#993: *1 upstream timed out (110: Connection timed out) while connecting to upstream, client: 192.168.1.232, server: 0.0.0.0:8000, upstream: "10.1.28.4:8000", bytes from/to client:0/0, bytes from/to upstream:0/0
@liyulai2017 ping between the nodes is not enough. Please check you can enter in the ingress pod and run curl to one of the endpoints:
kubectl exec <ing pod> -it -- curl -v http://10.1.28.4:8000
If you get an error it means you have issues with flannel.
Closing. Please reopen if you still have issues once the flannel routing is working ok.
@aledbf thank you very much ! it's really a problem with flannel routing.
@liyulai2017 How have you solved this flannel routing issue?
Most helpful comment
@liyulai2017 How have you solved this flannel routing issue?