Migrating Data from a Bring-Your-Own Nginx Ingress to a LoadBalancer Ingress

Example Scenarios

A company is using a CCE cluster on Huawei Cloud to run their services. They have set up ingress forwarding rules using their NGINX Ingress Controller, which includes domain-based and URL-based forwarding rules. Their services are accessible externally through a DNS domain name. When a client requests access to the domain name example.com, Nginx Ingress forwards the requests to the pod associated with the Service based on the forwarding rules created on the ingress.

To develop services, the company must transfer their data from Nginx ingress to a LoadBalancer ingress. To maintain service stability, they want to keep the DNS domain name and IP addresses of their backend servers unchanged. To meet this requirement, you can configure LoadBalancer ingress rules in the CCE cluster to preserve the previous forwarding rules. Then, configure weighted DNS records to transfer data from the Nginx ingress to the LoadBalancer ingress.

• Migrate data during off-peak hours.
• ELB is billed. For details, see Billing Overview.

Configuring the Target LoadBalancer Ingress

1. Create an ingress for migration and route it to the original backend service.
   In this example, an existing load balancer is used to create an ingress, and the elbingress.yaml is for reference only. You can also choose to automatically create a load balancer associated with the ingress. For more information, see Creating a LoadBalancer Ingress Using kubectl.

   apiVersion: networking.k8s.io/v1
   kind: Ingress 
   metadata: 
     name: ingress-test
     annotations: 
       kubernetes.io/elb.id: <your_elb_id>  # Replace it with the ID of your existing load balancer.
       kubernetes.io/elb.ip: <your_elb_ip>  # Replace it with the IP of your existing load balancer.
       kubernetes.io/elb.class: performance  # A dedicated load balancer is used.
       kubernetes.io/elb.port: '80' 
   spec:
     rules: 
     - host: 'example.com'            # Replace it with your domain name.
       http: 
         paths: 
         - path: '/'
           backend: 
             service:
               name: <your_service_name>  # Replace it with the name of your target Service, which is the same as the Service associated with the original Nginx ingress.
               port: 
                 number: 8080             # Replace 8080 with the port number of your target Service.
           property:
             ingress.beta.kubernetes.io/url-match-mode: STARTS_WITH
           pathType: ImplementationSpecific
     ingressClassName: cce  

2. Run the following command to create the ingress:

   kubectl apply -f elbingress.yaml

Redirecting Traffic to the LoadBalancer Ingress

Before redirecting traffic, ensure you have created record A to map the service domain name to the public IP address of the company's Nginx ingress. You can configure weighted DNS records to transfer data from the Nginx ingress to the LoadBalancer ingress.

Figure 1 Using DNS to redirect traffic

To redirect traffic, do as follows:

1. Go to Public Zones.
2. On the Public Zones page, click the domain name (example.com) of the public zone.
3. The Record Sets page is displayed. You can find an A record, which contains the public IP address of the Nginx ingress. Change the weight of the A record to 9.
4. Add an A record set. To do so, click Add Record Set, and set Value to the public IP address of the target LoadBalancer ingress and Weight to 1.
   Figure 2 Adding a record set
   

5. Gradually increase the weight of the DNS record for the LoadBalancer ingress to 100% without affecting any services.

Deleting Redundant Resources

After all persistent connections to the Nginx ingress are released and there is no more traffic being forwarded to it, you can safely release any unnecessary resources such as the Nginx ingress, NGINX ingress Controller, and webhook after observing for a period of time.