What is the proper way to set `use-forwarded-headers=true` for the nginx ingress controller?

Rancher
1

Greetings,
To set use-forwarded-headers: “true” For the NGINX Ingress Controller, I’ve navigated to System > ingress-nginx (namespace) and then edited the configMap nginx-configuration via the UI, adding:
use-forwarded-headers = “true”

Viewing the configMap as YAML, I see it’s added the entry under the data section (as I think it should) like so:

data:
  use-forwarded-headers: “true”

If I then step into one of the nginx-ingress-controller pods on one of the cluster nodes and grep /etc/nginx/nginx.conf for the setting, I see that the option value was not picked up:

use_forwarded_headers = false,

I’ve tried re-deploying all the pods but the custom setting is still not picked up from the configMap.

Per this documentation, the option should be supported, but I don’t see the option listed on this page, which seems to be Rancher’s fork that is being used in it’s clusters. Was omitting use_forwarded_headers from the list an oversight, or does Rancher’s fork not support this option?

Any help to get this option picked up by the nginx ingress controller would be greatly appreciated!

2

It doesn’t seem like the rancher fork of ingress controller supports the use-forwarded-headers option.

My goal is to have the client IP passed through to the service and it’s looking like the right way to accomplish this may be by enabling the proxy protocol between haproxy and nginx, e.g. use-proxy-protocol on the nginx side at least. It looks like haproxy needs either send-proxy or send-proxy-v2 in the backend definitions.