I am trying to deploye fluentd, elasticsearch & kibana into k8s. I am facing the difficulty to bring fluentd up. I am using the scripts from https://github.com/kubernetes/kubernetes/blob/master/cluster/addons/fluentd-elasticsearch/fluentd-es-ds.yaml, no no custom changes made.

I have one k8s master, and two nodes. After kubectl apply -f yamlfile one node have fluentd pod up, but another fail, with message:

{"log":"2018-03-13 03:45:15 +0000 [error]: unexpected error error_class=Errno::EHOSTUNREACH error=\"Failed to open TCP connection to 10.96.0.1:443 (No route to host - connect(2)     for \\\"10.96.0.1\\\" port 443)\"\n","stream":"stdout","time":"2018-03-13T03:45:15.815733611Z"}

I get this message from /var/log/containers/fluentd_xxx.log I don't know where does this "10.96.0.1:443" come from, and why one

fluentd-es-v2.0.4-9pgcv 1/1 Running 0 19m fluentd-es-v2.0.4-tkxvv 0/1 CrashLoopBackOff 7 19m

any hints would appreciate.