Deploy a Go MSSQL API Endpoint in the Kyma Runtime

1. Copy the repository URL.

In your browser, go to kyma-runtime-extension-samples. This repository contains a collection of Kyma sample applications which will be used during the tutorial.

Choose the Code button and choose one of the options to download the code locally or simply run the following command within your CLI at your desired folder location:

git clone https://github.com/SAP-samples/kyma-runtime-extension-samples

1. Open the api-mssql-go directory in your desired editor.

2. Explore the content of the sample.

3. Within the cmd/api directory, you can find main.go which is the main entry point of the Go application.

4. The docker directory contains the Dockerfile used to generate the Docker image. The image is built in two stages to create an image with a small file size.

   In the first stage, a Go image is used. It copies the related content of the project into the image and builds the application. The built application is then copied into the Docker scratch image and exposed on port 8000. The scratch image is an empty image containing no other tools within it so obtaining a shell/bash session is not possible. If desired, the following lines could be commented out to build an image with more included tools, but this results in a larger image.

   Shell/Bash

   Copy

   FROM scratch
   WORKDIR /app
   COPY --from=builder /app/api-mssql-go /app/
   

5. The internal directory contains the rest of the Go application, which is broken down into three packages: api, config and db. Explore these contents to understand the structure and functionality.

6. Within the k8s directory you can find the Kubernetes/Kyma resources you will apply to your Kyma runtime.

7. Within the root you can find go.mod and go.sum files that are used to manage the dependencies the application uses.

Run the following commands from the api-mssql-go directory using your CLI.

1. The application expects that the following environment variables are set as they are required for the database connection. Make sure to adjust them for your environment. Review the the tutorial Deploying MSSQL in the Kyma Runtime for the different configurations to run the database.

1. To run the application, use the following command:

   Shell/Bash

   Copy

   go run ./cmd/api/main.go
   

2. The application is available at http://localhost:8000/orders. You can use a tool such as curl to test the different HTTP methods, for example:

   To GET data:

   Shell/Bash

   Copy

   curl -i -H "Accept: application/json" -H "Content-Type: application/json" -X GET http://localhost:8000/orders
   

   To POST data:

   Shell/Bash

   Copy

   curl --data "{\"order_id\":\"10000003\",\"description\":\"test from curl\"}" http://localhost:8000/orders
   

Run the following commands from the api-mssql-go directory within your CLI.

Make sure to replace the value of <your-docker-id> with your Docker account ID.

1. To build the Docker image, run this command:

   Shell/Bash

   Copy

   docker build -t <your-docker-id>/api-mssql-go -f docker/Dockerfile .
   

2. To push the Docker image to your Docker repository, run this command:

   Shell/Bash

   Copy

   docker push <your-docker-id>/api-mssql-go
   

Run the following commands from the api-mssql-go directory within your CLI.

Make sure to replace the value of <your-docker-id> with your Docker account ID.

1. Start the image locally by running the following command:

The command is expecting that the database is available at localhost:1433 on the host machine. This is denoted by the host.docker.internal value. Review the tutorial Deploying MSSQL in the Kyma Runtime for the different configurations to run/stop the database.

1. The application is available at http://localhost:8000/orders. You can use a tool such as curl to test the different HTTP methods, for example:

   To GET data:

   Shell/Bash

   Copy

   curl -i -H "Accept: application/json" -H "Content-Type: application/json" -X GET http://localhost:8000/orders
   

   To POST data - make sure to use a unique value for the order_id:

   Shell/Bash

   Copy

   curl --data "{\"order_id\":\"10000004\",\"description\":\"test from curl\"}" http://localhost:8000/orders
   

2. Stop the Docker container by running:

   Shell/Bash

   Copy

   docker stop api-mssql-go
   

3. Here are some additional commands that you can use:

   To restart the Docker container, run:

   Shell/Bash

   Copy

   docker start api-mssql-go
   

   To remove the Docker container, run:

   Shell/Bash

   Copy

   docker rm api-mssql-go
   

You can find the resource definitions in the k8s folder. If you performed any changes in the configuration, these files may also need to be updated. The folder contains the following files that are relevant to this tutorial:

• apirule.yaml: defines the API endpoint which exposes the application to the Internet. This endpoint does not define any authentication access strategy and should be disabled when not in use.
• configmap.yaml: defines the name of the database, host, and port. The host value assumes that the service for the database is named mssql and is defined in the dev Namespace. Make sure to adjust this if you made any changes.
• deployment.yaml: defines the deployment definition for the Go API, as well as a service used for communication. This definition references both the secret.yaml, which was defined in the previous tutorial and also included in this directory, and the configmap.yaml by name.

1. Start by creating the dev Namespace if it doesn’t already exist:

   Shell/Bash

   Copy

   kubectl create namespace dev
   

   Namespaces separate objects inside a Kubernetes cluster. Choosing a different namespace will require adjustments to the provided samples.

2. Within the deployment.yaml, adjust the value of spec.template.spec.containers.image, commented with #change it to your image, to use your Docker image. Apply the Deployment which will cause an error which we will further explore:

   Shell/Bash

   Copy

   kubectl -n dev apply -f ./k8s/deployment.yaml
   

3. Check the status of the Pod by running:

   Shell/Bash

   Copy

   kubectl -n dev get po
   

   This command results in a table similar to the one below, showing a Pod with the name api-mssql-go- ending with a random hash. Make sure to adjust this value in the subsequent commands. Notice that STATUS is reporting CreateContainerConfigError.

   Shell/Bash

   Copy

   NAME                           READY   STATUS                       RESTARTS   AGE
   api-mssql-go-c694bc847-tkthc   1/2     CreateContainerConfigError   0          23s
   

   You can also see the logs that report the same STATUS CreateContainerConfigError:

   Shell/Bash

   Copy

   kubectl logs api-mssql-go-c694bc847-tkthc -c api-mssql-go -n dev
   

   To find the reason for the STATUS CreateContainerConfigError, review the Pod definition:

   Shell/Bash

   Copy

   kubectl describe pod api-mssql-go-c694bc847-tkthc -n dev
   

   Within the Events of the pods you should find a message Error: configmap "api-mssql-go" not found.

4. Apply the ConfigMap:

   Shell/Bash

   Copy

   kubectl -n dev apply -f ./k8s/configmap.yaml
   

5. Verify the status of the Pod by running:

   Shell/Bash

   Copy

   kubectl -n dev get po
   

   The Pod should now be running.

   Shell/Bash

   Copy

   NAME                           READY   STATUS    RESTARTS   AGE
   api-mssql-go-c694bc847-tkthc   2/2     Running   0          23m
   

6. Apply the APIRule:

   Shell/Bash

   Copy

   kubectl -n dev apply -f ./k8s/apirule.yaml
   

To access the API we can use the APIRule we created in the previous step.

1. Open the Kyma runtime console

2. From the menu, choose Namespaces

3. Choose the dev Namespace.

4. From the menu, choose Discovery and Network > API Rules.

5. Choose the Host entry for the api-mssql-go APIRule to open the application in the browser which will produce a 404 error. Append /orders to the end of the URL and refresh the page to successfully access the API. The URL should be similar to:

   https://api-mssql-go.<cluster>.kyma.ondemand.com/orders

   A tool such as curl can be used to test the various HTTP methods of the API.