#6 - ensures better resiliency and moves joining to initialisation.

[ajohnstone]

Partially based on https://github.com/kubernetes/charts/tree/master/incubator/consul

fixes #6

[casa87]

Nice, thank you I was struggling after a reboot due to new IP and this fixed this part, but I am still getting the following error :
2017/02/15 23:19:22 [ERR] agent: coordinate update error: No cluster leader 2017/02/15 23:19:25 [ERR] raft: Failed to take snapshot: nothing new to snapshot 2017/02/15 23:19:33 [ERR] raft: Failed to take snapshot: nothing new to snapshot 2017/02/15 23:19:39 [ERR] raft: Failed to take snapshot: nothing new to snapshot

Any idea on how to fix this step would be awesome

[ajohnstone]

On complete cluster failure ips change and raft fails. The above solves a number of problems, however I have yet to solve this last part, one suggestion I had was to use a service for each pod.

Although I have yet to try this or find an alternative solution.

For reference: hashicorp/consul#457

[kelseyhightower]

Using Kubernetes auto joining as define in the docs.