安装CoreOS-libvirt时kube-addons.service失败

Question

I have the following issue installing and provisioning my Kubernetes CoreOS-libvirt-based cluster.

When I'm logging on the master node, I see the following:

ssh core@192.168.10.1
Last login: Thu Dec 10 17:19:21 2015 from 192.168.10.254
CoreOS alpha (884.0.0)
Update Strategy: No Reboots
Failed Units: 1
  kube-addons.service

Trying to debug it, I run and receive the following:

core@kubernetes-master ~ $ systemctl status kube-addons.service
● kube-addons.service - Kubernetes addons
   Loaded: loaded (/etc/systemd/system/kube-addons.service; disabled; vendor preset: disabled)
   Active: failed (Result: exit-code) since Thu 2015-12-10 16:41:06 UTC; 41min ago
     Docs: https://github.com/GoogleCloudPlatform/kubernetes
  Process: 801 ExecStart=/opt/kubernetes/bin/kubectl create -f /opt/kubernetes/addons (code=exited, status=1/FAILURE)
  Process: 797 ExecStartPre=/bin/sleep 10 (code=exited, status=0/SUCCESS)
  Process: 748 ExecStartPre=/bin/bash -c while [[ "$(curl -s http://127.0.0.1:8080/healthz)" != "ok" ]]; do sleep 1; done (code=exited, status=0/SUCCESS)
 Main PID: 801 (code=exited, status=1/FAILURE)

Dec 10 16:40:53 kubernetes-master systemd[1]: Starting Kubernetes addons...
Dec 10 16:41:06 kubernetes-master kubectl[801]: replicationcontroller "skydns" created
Dec 10 16:41:06 kubernetes-master kubectl[801]: error validating "/opt/kubernetes/addons/skydns-svc.yaml": error validating data: found invalid field portalIP for v1.ServiceSpec; if you choose to ignore these errors, turn validation off with --validate=false
Dec 10 16:41:06 kubernetes-master systemd[1]: kube-addons.service: Main process exited, code=exited, status=1/FAILURE
Dec 10 16:41:06 kubernetes-master systemd[1]: Failed to start Kubernetes addons.
Dec 10 16:41:06 kubernetes-master systemd[1]: kube-addons.service: Unit entered failed state.
Dec 10 16:41:06 kubernetes-master systemd[1]: kube-addons.service: Failed with result 'exit-code'.

My etcd version is:

etcd --version
etcd version 0.4.9

But I have a etcd2 also:

etcd2 --version
etcd Version: 2.2.2
Git SHA: b4bddf6 
Go Version: go1.4.3
Go OS/Arch: linux/amd64

And at the current moment the second one is being runned:

ps aux | grep etcd 
etcd       731  0.5  8.4 329788 42436 ?        Ssl  16:40   0:16 /usr/bin/etcd2
root       874  0.4  7.4  59876 37804 ?        Ssl  17:19   0:02 /opt/kubernetes/bin/kube-apiserver --address=0.0.0.0 --port=8080 --etcd-servers=http://127.0.0.1:2379 --kubelet-port=10250 --service-cluster-ip-range=10.11.0.0/16
core       953  0.0  0.1   6740   876 pts/0    S+   17:27   0:00 grep --colour=auto etcd

What causes the issue and how can I solve it?

Thank you.

Answer 1

The relevant log line is:

/opt/kubernetes/addons/skydns-svc.yaml": error validating data: found invalid field portalIP for v1.ServiceSpec; if you choose to ignore these errors, turn validation off with --validate=false

You should figure out what's invalid about that IP or set the flag to ignore.