#6465
@zmerlynn

@dchen1107

LGTM

cc/ @derekwaynecarr

Sorry for breaking vagrant again. Derek, @ArtfulCoder did try to test the vagrant setup for etcd in pod, but failed even without his change by following kubernetes/docs/getting-started-guides/vagrant.md. We are going to definitely include vagrant into our test metrics, but need to clean up startup guide first. Can we have your help on this? Thanks!

@dchen1107: If you're looking for some Google-side help, @fabioy recently came back from vacation, and has setup Vagrant previously. He might be able to help.

@dchen1107 - i will test vagrant again now that this merged. i know it worked when i left on thurs, and was broken when i came back today, so something could have happened in interim aside from this.

@dchen1107 - so i tested after this PR merged and was able to bring up a k8s cluster on vagrant using a master and single minion. i will keep an eye out if there are any intermittent failures through week to see if there is something i can fix on my end.

@ArtfulCoder - thanks for the quick fix on this issue ;-)

@derekwaynecarr If this model works fine for you, I am fine with it. My concern is that we are keeping breaking vagrant setup, which we tried hard to avoid. The problem we couldn't avoid is lacking of automate system bringing a vagrant cluster for us to validate our changes. Thanks for quick validate the issues for us.