We’ve migrated another guide from Red Hat Satellite to Foreman: Configuring Smart Proxies with a Load Balancer, available at http://docs.theforeman.org/
This guide shows you how to distribute client requests and network load between your Smart Proxies. It includes steps to install an HAProxy load balancer on a Red Hat Enterprise Linux server, but you can install any suitable load balancer that supports TCP forwarding and sticky sessions.
Thanks a ton @tahliar for bringing this guide upstream!
As one of the original authors of the architecture behind the guide, I’d like to point out that the central feature of this setup is delivering (RPM) content to the clients. So you’ll need not only Foreman, but Katello (and the respective proxy features) to follow the guide. And given it’s RPM-centric but Katello also supports DEB packages, that part was neither tested nor documented (it should work, but you never know, right).
I wonder if we have a way to mark the guide as “Katello” and “RPM” so that users who find it won’t be confused if it does not work/apply to their setup?
@tahliar we could add a banner at the top and also we could update the book title to have (Katello/RPM) since there are no restrictions in the upstream?
If anyone would contribute the equivalent Debian steps, I’d happily add them in to the upstream version.
We’re glad to have you here @tahliar! And thank you.
We have been marking these areas with a sentence. However, isn’t the guide still useful for Puppet load balancing? I was reviewing the upstreaming process of this one and I thought it’s a good one to publish as-is.
Generally, we are trying to be more opinionated in the new documentation and the Katello deployment is the recommended one, we don’t even mention how to install Foreman-core only deployment type at the moment. Honestly, I am hoping that the deployment procedure can be simplified and unified (single installer) so we won’t need to do separate content because that increases complexity of maintaining but more importantly it requires users to do hard decisions at the early stage when they probably don’t have an answer yet.
