controller node fails to start periodically because flannel doesn't start.

[gamykla]

after kube-aws up the controller node fails to start properly because of some flannel error
cluster.yaml.txt
journalctl.txt

[peebs]

Hi Jelis,

Can you also provide which version of CoreOS this is happening on? Looking at the logs i'm not sure what is causing this but it could be from a recent OS change.

[gamykla]

@pbx0
$ kube-aws version
kube-aws version v0.8.2

from cluster.yaml
releaseChannel: stable
kubernetesVersion: v1.4.0_coreos.1

core@ip-10-0-0-50 ~ $ cat /etc/os-release
NAME=CoreOS
ID=coreos
VERSION=1122.2.0
VERSION_ID=1122.2.0
BUILD_ID=2016-09-06-1449
PRETTY_NAME="CoreOS 1122.2.0 (MoreOS)"
ANSI_COLOR="1;32"
HOME_URL="https://coreos.com/"
BUG_REPORT_URL="https://github.com/coreos/bugs/issues"

[gamykla]

For the hell of it, i tried to reboot the controller node and now the docker containers are up.

[gamykla]

@pbx0 i created another cluster today. This is 100% reproducible...
started cluster. logged into controller node. no containers running. sudo shutdown -r now ... log in again and its all good.....

[peebs]

Sorry I have not had the time to look into reproducing this but it could be the same issue as coreos/bugs#1393

This means that if flannel is failing to start even once (but then it might crashloop into a sucessful start) it could cause docker to cease running until it is manually restarted.