Merge pull request #801 from Bregor/features/cni/registry-proxy

Enable registry-proxy for CNI users

Author: mboersma

charts/workflow/values.yaml

@@ -57,6 +57,16 @@ global:
   # - true: deis-router will not be deployed. Workflow will not be usable until a Kubernetes ingress controller is installed.
   # - false: deis-router will be deployed (default).
   experimental_native_ingress: false
+  # If the Kubernetes cluster uses CNI
+  # use_cni: true
+  # Set the `listen` variable for registry-proxy's NGINX
+  #
+  # Valid values are:
+  # - 80: If the Kubernetes cluster run on GKE or AWS, or uses flannel or kubenet as a pod network
+  # - 127.0.0.1:5555: If the Kubernetes cluster uses CNI
+  #
+  # In case of CNI you can not use `hostPort` notation due to https://github.com/kubernetes/kubernetes/issues/23920
+  # registry_proxy_bind_addr: "80"
 
 
 s3:

src/managing-workflow/production-deployments.md

@@ -61,6 +61,10 @@ any deployed applications. To reduce the potential of this happening it is recom
 deis-router Deployment to run more than one router pod. This can be accomplished by running
 `kubectl --namespace=deis scale --replicas=2 deployment/deis-router`
 
+## Using on-cluster registry with CNI
+If you are using [CNI](https://github.com/containernetworking/cni) for managing container network, you cannot use `hostPort` notation due to [this issue](https://github.com/kubernetes/kubernetes/issues/23920).
+In this case you could enable CNI for `deis-registry-proxy` by setting `use_cni` variable to `true` inside `values.yaml` or by adding `--set global.use_cni=true` to `helm`'s args.
+
 [configuring object storage]: ../installing-workflow/configuring-object-storage.md
 [customizing controller]: tuning-component-settings.md#customizing-the-controller
 [customizing monitor]: tuning-component-settings.md#customizing-the-monitor