Put tiller binary on the path in helm v2.X container image so that helm-tiller plugin will use it instead of duplicating the download of helm tarball and performing version check.
In trying to use the base quay.io/roboll/helmfile image to deploy a chart that depended on an external repository, I ran into this error:
```
Adding repo bitnami https://charts.bitnami.com/bitnami
in ./helmfile.yaml: in .helmfiles[0]: in dirac/helmfile.yaml: helm exited with status 1:
Error: Couldn't load repositories file (/root/.helm/repository/repositories.yaml).
You might need to run `helm init` (or `helm init --client-only` if tiller is already installed)
```
The minimum to work around that error would be approximately:
```
mkdir /root/.helm/repository && echo apiVersion: v1 > /root/.helm/repository/repositories.yaml
```
Though there might be other things that I'm unwittingly depending on in the base `helm init`, it should be more friendly if we just run `helm init --client-only` while building.
* Add kubectl and jq to Dockerfile
Signed-off-by: Jake Hill <jake@naphta.uk>
* Update kubectl to use checksum verification, and pin to a specific version
Signed-off-by: Jake Hill <jake@naphta.uk>
Resolves#792
Resolves#435 (Git as chart repository)
Resolves#220 (S3 as chart repository)
Resolves#436 (About bundling helm plugins)
A lot of thanks to @aslafy-z for authoring the awesome helm-git plugin and contributing it to the community!
* use golang 1.10.1 images everywhere
* introduce a `static-linux` target to build a statically linked 64-bit binary only
* use that binary for the official image and the circle build and tests
* use a multi-stage dockerfile for a consistent build environment