For some reason I'm unable to curl my service or pod exposing the service. Below setup used to work - not anymore.
$ kubectl create deployment nginx --image=nginx
$ kubectl expose deployment nginx --port=80
$ kubectl get svc,pods -o wide
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE SELECTOR
service/nginx ClusterIP 10.109.212.238 <none> 80/TCP 12m app=nginx
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
pod/nginx-866795cfb8-ff5zv 1/1 Running 0 13m 10.47.0.2 nebula-web01 <none> <none>
Getting application content from Service IP or Pod IP results with this error.
$ curl 10.109.212.238
curl: (7) Failed connect to 10.109.212.238:80; No route to host
$ curl 10.47.0.2
curl: (7) Failed connect to 10.47.0.2:80; No route to hostFunny enough I discovered that the error appeared only on nebula-web01 node. If I deployed the same application on different node e.g. nebula-web02 and curl Service IP or Pod IP then the application returned its content. This made me think that there has to be something wrong with nebula-web01 setup.
What fixed my problem was to restarting docker and kubelet daemon on nebula-web01 node.
$ systemctl restart docker
$ systemctl restart kubeletAdditonaly, you might disable firewall and flush iptables according to this instructions.