At my Kubernetes environment, I cannot ping pods from other pods. Pods cannot access services, either. Looks like there is a configuration problem. Pod to pod and pod to service communications fail.
I followed the CoreOS + Kubernetes manual steps to install the kubernetes environment (Calico is not installed). https://coreos.com/kubernetes/docs/1.6.1/getting-started.html
Do you have any suggestion how to investigate the problem?
Additional Information----------------------------
Services created:
system # /home/core/kubernetes/kubectl get services -o wide
NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE SELECTOR
kubernetes 10.3.0.1 <none> 443/TCP 35d <none>
upper-rest 10.3.0.195 <nodes> 5000:32001/TCP 12h app=upper-restPods:
system # /home/core/kubernetes/kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE
upper-rest-4004779333-h7fc2 1/1 Running 0 12h 10.2.58.4 10.201.101.238
upper-rest-4004779333-j444t 1/1 Running 0 12h 10.2.44.2 10.201.101.236if I ping the service, I get no results
system # ping 10.3.0.195
PING 10.3.0.195 (10.3.0.195) 56(84) bytes of data.
^C
--- 10.3.0.195 ping statistics ---
24 packets transmitted, 0 received, 100% packet loss, time 23555ms
Similarly when I run the ping inside the pod to access a service or other pod's IP, again no reply back.
system # /home/core/kubernetes/kubectl exec upper-rest-4004779333-h7fc2 -- ping 10.3.0.195
PING 10.3.0.195 (10.3.0.195): 56 data bytes
^C
system # /home/core/kubernetes/kubectl exec upper-rest-4004779333-h7fc2 -- ping 10.2.44.2
PING 10.2.44.2 (10.2.44.2): 56 data bytes
^C
that i have no enough reputation to add comment, I think your iptables is old record, maybe you have reset your k8s ever.
I recommanded that you can reset all the cluster, include docker, and clear the iptable and ip link
iptables -P INPUT ACCEPT
iptables -P FORWARD ACCEPT
iptables -P OUTPUT ACCEPT
iptables -t nat -F
iptables -t mangle -F
iptables -F
iptables -X
and
ip link delete cni0
ip link delete flannel.1try with this,
Good luck