AWS example fixes

[donovanmuller]

When trying the AWS example I encountered a few issues.
Below are the fixes required to successfully bring up a cluster using the example resources:

0b96df5 - Current Docker (releases.rancher.com/install-docker/1.12.sh) install fails with:

module.nodes.aws_instance.rke-node[0] (remote-exec): + sudo -E sh -c 'sleep 3; apt-get install -y -q dirmngr'
module.nodes.aws_instance.rke-node[0] (remote-exec): Reading package lists...

module.nodes.aws_instance.rke-node[0] (remote-exec): Building dependency tree...
module.nodes.aws_instance.rke-node[0] (remote-exec): Reading state information...
module.nodes.aws_instance.rke-node[0] (remote-exec): Some packages could not be installed. This may mean that you have
module.nodes.aws_instance.rke-node[0] (remote-exec): requested an impossible situation or if you are using the unstable
module.nodes.aws_instance.rke-node[0] (remote-exec): distribution that some required packages have not yet been created
module.nodes.aws_instance.rke-node[0] (remote-exec): or been moved out of Incoming.
module.nodes.aws_instance.rke-node[0] (remote-exec): The following information may help to resolve the situation:

module.nodes.aws_instance.rke-node[0] (remote-exec): The following packages have unmet dependencies:
module.nodes.aws_instance.rke-node[0] (remote-exec):  dirmngr : Depends: libassuan0 (>= 2.4.0) but it is not installable
module.nodes.aws_instance.rke-node[0] (remote-exec):            Depends: libnpth0 (>= 0.90) but it is not installable
module.nodes.aws_instance.rke-node[0] (remote-exec): E: Unable to correct problems, you have held broken packages.
module.nodes.aws_instance.rke-node[0] (remote-exec): usermod: group 'docker' does not exist


Error: error executing "/tmp/terraform_2106265970.sh": Process exited with status 6

bumping to https://releases.rancher.com/install-docker/18.09.sh (as per https://rancher.com/docs/rancher/v2.x/en/installation/requirements/installing-docker/).

e003b67 - Due to rancher/rke#1725 we need to include the node internal_address. Additionally, this has to be the IP and not internal DNS, due to etcd-io/etcd#9575

d9d69f4 - Upgrade to rke v1.0.2 to prevent kube-apiserver not starting due to rancher/rke#1805

[rawmind0]

Hi @donovanmuller . Thanks for the feedback and the PR.

Could you please take a look to changes requested??

[rawmind0]

One question @donovanmuller, what golang version are you using?? I got some differences executing GO111MODULE=on go mod vendor command.

[donovanmuller]

$ go version
go version go1.13.4 darwin/amd64

[rawmind0]

@donovanmuller , btm provider is developed using golang 1.12. Modules and vendor updated should be done using same version. That's why i'm encountering diffs on vendor updated done by golang 1.13.

I've submitted a PR just including modules and vendor update to support rke v1.0.2 #148 . Could you please remove your commit that is updating go modules and vendor files and just let the aws examples modifications??

[donovanmuller]

@rawmind0 Removed commit efc3294 👍

[rawmind0]

Thanks @donovanmuller , last detail. PR #148 is already merged, could you please rebase and remove go.mod and go.sum files from your PR??

[rawmind0]

LGTM