Having issues configuring an AWS ELB for use with the v1.2 HA deployment.
The docs indicate to expose port 8080 on the rancher/server container, so I’ve tried configuring the ELB for HTTPS/HTTP (443/8080) which appears to work as long as ProxyProtocol is not enable enabled for the ELB, but then my agents won’t connect. If I enable ProxyProtocol, then the agents connect but I can’t get to the web page for Rancher server (it returns 400 Bad Request).
Furthermore, the section on “Basic SSL Configuration” retains the previous configurations of using ports 80/81 and 443/444 for the server when using an ELB.
It appears there is a serious misalignment of the docs for v1.2 when it comes to deploying in a HA setup.
You want the newer ALB (Application Load Balancer) for 1.2, It supports WebSockets properly so you don’t need TCP listeners and PROXY protocol and all that shenanigans.
@denise docs need updating --^
Thanks vincent. That appears to have worked.
I see the Install Docs are updated to include a bit about using a ALB, however in trying to set one up I’ve been unsuccessful.
We have the additional requirement that all HTTP requests are re-directed to HTTPS, so everything is always over SSL. In the past we set this up by running a NGINX container withs upport for PROXY protocol (see GitHub - Demandbase/docker-nginx-https-redirect: A simple nginx container that redirects all http requests to https) and it worked well with an ELB.
Is there a guide on setting up Rancher to use an ALB in AWS? We have a ALB setup with a Target Group pointing to 8080 on the Rancher master server, with a health check of 8080:/, but it never comes into service.
I also like to figure out how to use a HTTP 301 re-direct with an ALB, but that’s not really a Rancher issue.
Using /v1/token or /v2-beta/token as the health check path works for a AWS ALB.
