In this video, you will see how to provision a load balancer. Before we create a load balancer, we need to have the backend servers in place. I already have two compute instances created as Web server 1 and Web server 2. I have installed a basic web server there, httpd server is installed. They are both public compute instances with a public IP. If you reach the public IP of each of those individual web servers, you get that response with the static page that is given. This is just for the demonstration purposes. If you made the application reachable directly on the compute instance host, you will know that only one response will come. In the ideal scenario, both responses should be the same response. But for the purpose of this demonstration, to identify which web server is running, we have given a different default page for both the nodes. Now, let me go and provision a load balancer here. I am not going to implement SSL here, but I will tell you how you implement the SSL when we come into those specific pages and load balancer of type Layer 7 is what I'm going to use. When we use this, I'm going to use a public load balancer. I will call it as public load balancer. Get an ephemeral IP address. I can choose a standard shape of minimal bandwidth because we're not going to have a heavy load. Choose the VCM and a public subnet to provision it. Under Advanced Options, you can add tags as required. You need to add security rules to offer communication which you can use by network security groups if you have, or in the next screen, you will have the ability for OCI to add rules automatically so that communication between the load balancer and the backend servers is allowed, as well as accessing the load balancer on the listener port over the Internet because this is a public load balancer. I will choose the default weighted round robin method of load balancing policy, and add the two compute instances as the backend servers. Both backend servers would ideally be distributed across multiple ADs, so that you have a highly available application. You choose the port in which the application is available in each load balancer so that the health check will be automatically configured based on that, and I'm using a standard HTTP based web server, so this is enough. I will use the default health check policies that come along. If I want to enable SSL communication between the load balancer and the backend sets, I can use the SSL certificate here. This is something we will see later, but right now, you have the ability here. The collection of backend servers that I've put in can be grouped into a backend set. You can ask OCI to automatically add rules or you can manually add them by choosing this option. If you want session persistence so that a given client connection should always be routed to the same backend server, you can use it by enabling cookies. I'm not going to use that in this implementation. In the last screen, you choose a listener wherein the load balancer will listen on this port. I have a basic HTTP load balancer implementation that I'm doing, but in case you want client-side encryption, wherein communication from the end clients to the load balancer to be encrypted, then you can upload an SSL certificate here. We'll come to that later on. Let me choose a standard HTTP based load balancer which will listen on port 80, and under Advanced Options you can set a timeout as to when it will throw back an error. With that, I will provision the load balancer, and as part of the load balancer provisioning, the health check is shown. We will pause here and come back to the next video once the load balancer is provisioned so that we can go ahead with further topics related to this chapter.
