This example shows how to run a C++ application on Cloud Batch job using C++.

If you are not familiar with the Batch API, we recommend you first read the API overview before starting this guide.

The following steps are included:

1. Create a docker image
2. Upload it to Artifact registry
3. Create the job
4. Poll until the job finishes

1. Install the gcloud CLI.
2. Install docker.

The instructions are here.

1. [If it does not already exist] Create the artifact repository
2. Build the image locally
3. Tag and push the image to the artifact repository

To run this example, replace the [PROJECT ID] placeholder with the id of your project:

Authorize via gcloud cli

PROJECT_ID=[PROJECT_ID]
LOCATION="us-central1"
REPOSITORY="application-repo"

gcloud auth login
gcloud config set project ${PROJECT_ID}
# Create the repository
gcloud artifacts repositories create ${REPOSITORY} \
    --repository-format=docker \
    --location=${LOCATION} \
    --description="Store the example C++ application" \
    --async

application-repo            DOCKER  STANDARD_REPOSITORY  Store the example C++ application  us-central1          Google-managed key  2024-05-13T20:07:11  2024-05-13T20:07:11  0

cd batch/cpp_application/application
docker build --tag=application-image:latest .

cd batch/cpp_application/application
gcloud builds submit --region=${LOCATION} --tag ${LOCATION}-docker.pkg.dev/${PROJECT_ID}/${REPOSITORY}/application-image:latest

# Tag the image
docker tag application-image:latest ${LOCATION}-docker.pkg.dev/${PROJECT_ID}/${REPOSITORY}/application-image:latest

# Push the image
docker push ${LOCATION}-docker.pkg.dev/${PROJECT_ID}/${REPOSITORY}/application-image:latest

You can do either of the following:

1. Use the C++ client libraries to create and poll for the job until completion
2. Use the gcloud CLI to create the job

This project uses vcpkg to install its dependencies. Clone vcpkg in your $HOME:

git clone -C $HOME https://github.com/microsoft/vcpkg.git

Install the typical development tools, on Ubuntu you would use:

apt update && apt install -y build-essential cmake git ninja-build pkg-config g++ curl tar zip unzip

In this directory compile the dependencies and the code, this can take as long as an hour, depending on the performance of your workstation:

cd cpp-samples/batch/cpp_application
cmake -S . -B .build -DCMAKE_BUILD_TYPE=Release \
    -DCMAKE_TOOLCHAIN_FILE=$HOME/vcpkg/scripts/buildsystems/vcpkg.cmake
cmake --build .build

Run the example, replace the [PROJECT ID] placeholder with the id of your project:

.build/driver [PROJECT ID] us-central1 cpp-application-run application.json application-repo

This submits the batch job and then polls until the job is complete.

1. Replace the imageURI field in application.json

  "runnables": [
      {
          "container": {
              "imageUri": "${LOCATION_ID}-docker.pkg.dev/${PROJECT_ID}/{REPOSITORY}/application-image:latest",
          }
      }
  ],

2. Submit the job

gcloud batch jobs submit cpp-application-cli-run \
    --config=application.json \
    --location=us-central1

3. Check on the job status

gcloud batch jobs describe cpp-application-cli-run  --location=us-central1