Document Kubefed on-prem AWS Route53 configuration (#7834)

* Initial Kubefed AWS Route53 on-prem documentation

* Add flag requirement for kubefed init

* conform to style guidelines

* fix controller-manager wording

* move cut-off section from CoreDNS support
This commit is contained in:
Doug Holt 2018-04-12 16:25:58 -06:00 committed by k8s-ci-robot
parent 629f3010d3
commit 6715e08ec9
1 changed files with 37 additions and 0 deletions

View File

@ -374,6 +374,43 @@ kubefed init fellowship \
For more information see
[Setting up CoreDNS as DNS provider for Cluster Federation](/docs/tasks/federation/set-up-coredns-provider-federation/).
#### AWS Route53 support
It is possible to utilize AWS Route53 as a cloud DNS provider when the
federation controller-manager is run on-premise. The controller-manager
Deployment must be configured with AWS credentials since it cannot implicity
gather them from a VM running on AWS.
Currently, `kubefed init` does not read AWS Route53 credentials from the
`--dns-provider-config` flag, so a patch must be applied.
Specify AWS Route53 as your DNS provider when initializing your on-premise
federation controller-manager by passing the flag `--dns-provider="aws-route53"`
to `kubefed init`.
Create a patch file with your AWS credentials:
```yaml
spec:
template:
spec:
containers:
- name: controller-manager
env:
- name: AWS_ACCESS_KEY_ID
value: "ABCDEFG1234567890"
- name: AWS_SECRET_ACCESS_KEY
value: "ABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890"
```
Patch the Deployment:
```shell
kubectl -n federation-system patch deployment controller-manager --patch "$(cat <patch-file-name>.yml)"
```
Where `<patch-file-name>` is the name of the file you created above.
## Adding a cluster to a federation
After you've deployed a federation control plane, you'll need to make that control plane aware of the clusters it should manage.