RFE: Boot-strapping etcd cluster + operator

[timothysc]

FEATURE REQUEST

This is one of several changes I'm proposing regarding kubeadm support for HA and I will cross link with the draft-doc I'd like to share tomorrow, which highlights the details, right now this is a placeholder to be cross-linked.

Versions

All

/cc @kubernetes/sig-cluster-lifecycle-feature-requests

[xiang90]

/cc @hongchaodeng

[timothysc]

xref: https://docs.google.com/document/d/1lH9OKkFZMSqXCApmSXemEDuy9qlINdm5MfWWGrK3JYc/edit#

[justinsb]

Tracking some issues here: coreos/tectonic-installer#257

[jamiehannaford]

Unless someone else has claimed it, I'd like to work on this feature if that's okay

[timothysc]

@jamiehannaford go for it. I'd be happy to help it along where I can.

The one constraint that is missing from the doc is we would like every

kubeadm init/(master join) to update the cluster vs some explicit migrate step that is listed in the doc.

[luxas]

@timothysc @jamiehannaford Is this tracking issue needed when we have #261?

[jamiehannaford]

I think we can close this since it's covered by #261 and #277

[luxas]

Ok, thanks @jamiehannaford