Multiple Discrete Instances with App-Only Image - no Load Balancer

Overview

 In our steps for setting up a multiple discrete instance of Yellowfin, we’ll create two Yellowfin instances  using the LoadBalancer service type in Kubernetes, so that each Yellowfin instance can be accessed on port 8080 of the provisioned load balancer.  

The Production instance of Yellowfin will be allocated 6GB of RAM, and the Development instance will be allocated 4GB of RAM. In a Kubernetes environment where a load balancer cannot be provisioned during the deployment (eg, a basic on-premises environment), the services specification can be modified to switch the ServiceType to NodePort. 

Before deploying your instances with these defaults, make sure you have already created a repository database and synced it with the same version of Yellowfin that will be used in the Yellowfin container. To do this, download the full application installer for Yellowfin, and install it on your workstation. This will create a Yellowfin repo DB as well as an instance of Yellowfin in a folder which can be deleted after configuring the containers.   

For a list of supported database types, see the database information on Install And Deploy Yellowfin.

In the steps below, we’ll show you how to deploy two discrete instances of Yellowfin. 

1. Install the full application installer version of Yellowfin on your workstation (this is temporary to ensure the repo DB is available for the containers to use)
2. Copy the web.xml file from this installation and save it as a backup to your preferred location (this acts as a reference for the Yellowfin credentials required to connect to your Yellowfin repo DB)
3. Ensure Kubernetes is running

4. Copy the following text and paste it into your preferred text editor:
   

   --- 
   ### Yellowfin Production Instance - Service ### 
   apiVersion: v1 
   kind: Service 
   metadata: 
     name: yellowfin-multi-instance-prod 
   spec: 
     ports: 
     - name: "web" 
       port: 8080 
       targetPort: 8080 
     selector: 
       app: yellowfin-multi-instance-prod 
     type: LoadBalancer 
   status: 
     loadBalancer: {} 
   --- 
   ### Yellowfin Development Instance - Service ### 
   apiVersion: v1 
   kind: Service 
   metadata: 
     name: yellowfin-multi-instance-dev 
   spec: 
     ports: 
     - name: "web" 
       port: 8080 
       targetPort: 8080 
     selector: 
       app: yellowfin-multi-instance-dev 
     type: LoadBalancer 
   status: 
     loadBalancer: {} 
   --- 
   ### Yellowfin Production Instance - Deployment ### 
   apiVersion: apps/v1 
   kind: Deployment 
   metadata: 
     namespace: default 
     labels: 
       app: yellowfin-multi-instance-prod 
     name: yellowfin-multi-instance-prod 
   spec: 
     replicas: 1 
     selector: 
       matchLabels: 
         app: yellowfin-multi-instance-prod 
     template: 
       metadata: 
         labels: 
           app: yellowfin-multi-instance-prod 
       spec: 
         containers: 
           - env: 
             - name: APP_MEMORY 
               value: "6144" 
             - name: JDBC_CLASS_NAME 
               value: INSERT_DATABASE_TYPE_1_HERE 
             - name: JDBC_CONN_ENCRYPTED 
               value: "true" 
             - name: JDBC_CONN_PASS 
               value: INSERT_JDBC_PASSWORD_1_HERE 
             - name: JDBC_CONN_URL 
               value: jdbc:INSERT_JDBC_CONNECTION_STRING_1_HERE 
             - name: JDBC_CONN_USER 
               value: INSERT_DATABASE_USER_1_HERE 
             name: yellowfin-multi-instance-prod 
             image: yellowfinbi/yellowfin-app-only:<RELEASE_VERSION_GOES_HERE> 
             ports: 
               - name: web 
                 containerPort: 8080 
   --- 
   ### Yellowfin Development Instance - Deployment ### 
   apiVersion: apps/v1 
   kind: Deployment 
   metadata: 
     namespace: default 
     labels: 
       app: yellowfin-multi-instance-dev 
     name: yellowfin-multi-instance-dev 
    
   spec: 
     replicas: 1 
     selector: 
       matchLabels: 
         app: yellowfin-multi-instance-dev 
     template: 
       metadata: 
         labels: 
           app: yellowfin-multi-instance-dev 
       spec: 
         containers: 
           - env: 
             - name: APP_MEMORY 
               value: "4096" 
             - name: JDBC_CLASS_NAME 
               value: INSERT_DATABASE_TYPE_2_HERE 
             - name: JDBC_CONN_ENCRYPTED 
               value: "true" 
             - name: JDBC_CONN_PASS 
               value: INSERT_JDBC_PASSWORD_2_HERE 
             - name: JDBC_CONN_URL 
               value: jdbc: INSERT_JDBC_CONNECTION_2_STRING_HERE 
             - name: JDBC_CONN_USER 
               value: INSERT_DATABASE_USER_2_HERE 
             name: yellowfin-multi-instance-prod 
             image: yellowfinbi/yellowfin-app-only:<RELEASE_VERSION_GOES_HERE> 
             ports: 
               - name: web 
                 containerPort: 8080 

   
   

5. Read through the above text and replace the database connection settings with your own configuration details (these are located in the web.xml file of the Yellowfin installation)
6. Save the text to a YAML file called yellowfin-multiple-instances.yml
7. Run the following command in a terminal to deploy Yellowfin:
   Kubectl apply –f yellowfin-multiple-instances.yml 
8. Start Yellowfin by typing your host URLon port 8080
9. Ensure that Yellowfin is running from your containers and that you can login (this confirms that your login credentials are correct, so you can safely delete the workstation instance of Yellowfin)
10. Delete the workstation instance of Yellowfin by removing the folder

Section navigation

This page is part of the Install And Deploy Yellowfin section of the wiki, which has these topics: