Run a Batch job using Workflows

Batch is a fully managed service that lets you schedule, queue, and execute batch processing workloads on Compute Engine virtual machine (VM) instances. Batch provisions resources and manages capacity on your behalf, allowing your batch workloads to run at scale.

Workflows allows you to execute the services you need in an order that you define described using the Workflows syntax.

In this tutorial, you use the Workflows connector for Batch to schedule and run a Batch job that executes six tasks in parallel on two Compute Engine VMs. Using both Batch and Workflows allows you to combine the advantages they offer and efficiently provision and orchestrate the entire process.

Create an Artifact Registry repository

Create a repository to store your Docker container image.

Console

  1. In the Google Cloud console, go to the Repositories page.

    Go to Repositories

  2. Click Create Repository.

  3. Enter containers as the repository name.

  4. For Format, choose Docker.

  5. For Location Type, choose Region.

  6. In the Region list, select us-central1.

  7. Click Create.

gcloud

Run the following command:

  gcloud artifacts repositories create containers \
    --repository-format=docker \
    --location=us-central1

You have created an Artifact Registry repository named containers in the us-central1 region. For more information about supported regions, see Artifact Registry locations.

Get the code samples

Google Cloud stores the application source code for this tutorial in GitHub. You can clone that repository or download the samples.

  1. Clone the sample app repository to your local machine:

    git clone https://github.com/GoogleCloudPlatform/batch-samples.git

    Alternatively, you can download the samples in the main.zip file and extract it.

  2. Change to the directory that contains the sample code:

    cd batch-samples/primegen

You now have the source code for the application in your development environment.

Build the Docker image using Cloud Build

The Dockerfile contains the information needed to build a Docker image using Cloud Build. Run the following command to build it:

gcloud builds submit \
  -t us-central1-docker.pkg.dev/PROJECT_ID/containers/primegen-service:v1 PrimeGenService/

Replace PROJECT_ID with your Google Cloud project ID.

When the build is complete, you should see output similar to the following:

DONE
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
ID: a54818cc-5d14-467b-bfda-5fc9590af68c
CREATE_TIME: 2022-07-29T01:48:50+00:00
DURATION: 48S
SOURCE: gs://project-name_cloudbuild/source/1659059329.705219-17aee3a424a94679937a7200fab15bcf.tgz
IMAGES: us-central1-docker.pkg.dev/project-name/containers/primegen-service:v1
STATUS: SUCCESS

Using a Dockerfile, you've built a Docker image named primegen-service and pushed the image to an Artifact Registry repository named containers.

Deploy a workflow that schedules and runs a Batch job

The following workflow schedules and runs a Batch job that runs a Docker container as six tasks in parallel on two Compute Engine VMs. The result is the generation of six batches of prime numbers, stored in a Cloud Storage bucket.

Console

  1. In the Google Cloud console, go to the Workflows page.

    Go to Workflows

  2. Click Create.

  3. Enter a name for the new workflow, such as batch-workflow.

  4. In the Region list, select us-central1.

  5. Select the Service account you previously created.

  6. Click Next.

  7. In the workflow editor, enter the following definition for your workflow:

    YAML

    main:
  params: [args]
  steps:
    - init:
        assign:
          - projectId: ${sys.get_env("GOOGLE_CLOUD_PROJECT_ID")}
          - region: "us-central1"
          - imageUri: ${region + "-docker.pkg.dev/" + projectId + "/containers/primegen-service:v1"}
          - jobId: ${"job-primegen-" + string(int(sys.now()))}
          - bucket: ${projectId + "-" + jobId}
    - createBucket:
        call: googleapis.storage.v1.buckets.insert
        args:
          query:
            project: ${projectId}
          body:
            name: ${bucket}
    - logCreateBucket:
        call: sys.log
        args:
          data: ${"Created bucket " + bucket}
    - logCreateBatchJob:
        call: sys.log
        args:
          data: ${"Creating and running the batch job " + jobId}
    - createAndRunBatchJob:
        call: googleapis.batch.v1.projects.locations.jobs.create
        args:
            parent: ${"projects/" + projectId + "/locations/" + region}
            jobId: ${jobId}
            body:
              taskGroups:
                taskSpec:
                  runnables:
                    - container:
                        imageUri: ${imageUri}
                      environment:
                        variables:
                          BUCKET: ${bucket}
                # Run 6 tasks on 2 VMs
                taskCount: 6
                parallelism: 2
              logsPolicy:
                destination: CLOUD_LOGGING
        result: createAndRunBatchJobResponse
    # You can delete the batch job or keep it for debugging
    - logDeleteBatchJob:
        call: sys.log
        args:
          data: ${"Deleting the batch job " + jobId}
    - deleteBatchJob:
        call: googleapis.batch.v1.projects.locations.jobs.delete
        args:
            name: ${"projects/" + projectId + "/locations/" + region + "/jobs/" + jobId}
        result: deleteResult
    - returnResult:
        return:
          jobId: ${jobId}
          bucket: ${bucket}

    JSON

    {
  "main": {
    "params": [
      "args"
    ],
    "steps": [
      {
        "init": {
          "assign": [
            {
              "projectId": "${sys.get_env(\"GOOGLE_CLOUD_PROJECT_ID\")}"
            },
            {
              "region": "us-central1"
            },
            {
              "imageUri": "${region + \"-docker.pkg.dev/\" + projectId + \"/containers/primegen-service:v1\"}"
            },
            {
              "jobId": "${\"job-primegen-\" + string(int(sys.now()))}"
            },
            {
              "bucket": "${projectId + \"-\" + jobId}"
            }
          ]
        }
      },
      {
        "createBucket": {
          "call": "googleapis.storage.v1.buckets.insert",
          "args": {
            "query": {
              "project": "${projectId}"
            },
            "body": {
              "name": "${bucket}"
            }
          }
        }
      },
      {
        "logCreateBucket": {
          "call": "sys.log",
          "args": {
            "data": "${\"Created bucket \" + bucket}"
          }
        }
      },
      {
        "logCreateBatchJob": {
          "call": "sys.log",
          "args": {
            "data": "${\"Creating and running the batch job \" + jobId}"
          }
        }
      },
      {
        "createAndRunBatchJob": {
          "call": "googleapis.batch.v1.projects.locations.jobs.create",
          "args": {
            "parent": "${\"projects/\" + projectId + \"/locations/\" + region}",
            "jobId": "${jobId}",
            "body": {
              "taskGroups": {
                "taskSpec": {
                  "runnables": [
                    {
                      "container": {
                        "imageUri": "${imageUri}"
                      },
                      "environment": {
                        "variables": {
                          "BUCKET": "${bucket}"
                        }
                      }
                    }
                  ]
                },
                "taskCount": 6,
                "parallelism": 2
              },
              "logsPolicy": {
                "destination": "CLOUD_LOGGING"
              }
            }
          },
          "result": "createAndRunBatchJobResponse"
        }
      },
      {
        "logDeleteBatchJob": {
          "call": "sys.log",
          "args": {
            "data": "${\"Deleting the batch job \" + jobId}"
          }
        }
      },
      {
        "deleteBatchJob": {
          "call": "googleapis.batch.v1.projects.locations.jobs.delete",
          "args": {
            "name": "${\"projects/\" + projectId + \"/locations/\" + region + \"/jobs/\" + jobId}"
          },
          "result": "deleteResult"
        }
      },
      {
        "returnResult": {
          "return": {
            "jobId": "${jobId}",
            "bucket": "${bucket}"
          }
        }
      }
    ]
  }
}

  8. Click Deploy.

gcloud

  1. Create a source code file for your workflow:

    touch batch-workflow.JSON_OR_YAML

    Replace JSON_OR_YAML with yaml or json depending on the format of your workflow.

  2. In a text editor, copy the following workflow to your source code file:

    YAML

    main:
  params: [args]
  steps:
    - init:
        assign:
          - projectId: ${sys.get_env("GOOGLE_CLOUD_PROJECT_ID")}
          - region: "us-central1"
          - imageUri: ${region + "-docker.pkg.dev/" + projectId + "/containers/primegen-service:v1"}
          - jobId: ${"job-primegen-" + string(int(sys.now()))}
          - bucket: ${projectId + "-" + jobId}
    - createBucket:
        call: googleapis.storage.v1.buckets.insert
        args:
          query:
            project: ${projectId}
          body:
            name: ${bucket}
    - logCreateBucket:
        call: sys.log
        args:
          data: ${"Created bucket " + bucket}
    - logCreateBatchJob:
        call: sys.log
        args:
          data: ${"Creating and running the batch job " + jobId}
    - createAndRunBatchJob:
        call: googleapis.batch.v1.projects.locations.jobs.create
        args:
            parent: ${"projects/" + projectId + "/locations/" + region}
            jobId: ${jobId}
            body:
              taskGroups:
                taskSpec:
                  runnables:
                    - container:
                        imageUri: ${imageUri}
                      environment:
                        variables:
                          BUCKET: ${bucket}
                # Run 6 tasks on 2 VMs
                taskCount: 6
                parallelism: 2
              logsPolicy:
                destination: CLOUD_LOGGING
        result: createAndRunBatchJobResponse
    # You can delete the batch job or keep it for debugging
    - logDeleteBatchJob:
        call: sys.log
        args:
          data: ${"Deleting the batch job " + jobId}
    - deleteBatchJob:
        call: googleapis.batch.v1.projects.locations.jobs.delete
        args:
            name: ${"projects/" + projectId + "/locations/" + region + "/jobs/" + jobId}
        result: deleteResult
    - returnResult:
        return:
          jobId: ${jobId}
          bucket: ${bucket}

    JSON

    {
  "main": {
    "params": [
      "args"
    ],
    "steps": [
      {
        "init": {
          "assign": [
            {
              "projectId": "${sys.get_env(\"GOOGLE_CLOUD_PROJECT_ID\")}"
            },
            {
              "region": "us-central1"
            },
            {
              "imageUri": "${region + \"-docker.pkg.dev/\" + projectId + \"/containers/primegen-service:v1\"}"
            },
            {
              "jobId": "${\"job-primegen-\" + string(int(sys.now()))}"
            },
            {
              "bucket": "${projectId + \"-\" + jobId}"
            }
          ]
        }
      },
      {
        "createBucket": {
          "call": "googleapis.storage.v1.buckets.insert",
          "args": {
            "query": {
              "project": "${projectId}"
            },
            "body": {
              "name": "${bucket}"
            }
          }
        }
      },
      {
        "logCreateBucket": {
          "call": "sys.log",
          "args": {
            "data": "${\"Created bucket \" + bucket}"
          }
        }
      },
      {
        "logCreateBatchJob": {
          "call": "sys.log",
          "args": {
            "data": "${\"Creating and running the batch job \" + jobId}"
          }
        }
      },
      {
        "createAndRunBatchJob": {
          "call": "googleapis.batch.v1.projects.locations.jobs.create",
          "args": {
            "parent": "${\"projects/\" + projectId + \"/locations/\" + region}",
            "jobId": "${jobId}",
            "body": {
              "taskGroups": {
                "taskSpec": {
                  "runnables": [
                    {
                      "container": {
                        "imageUri": "${imageUri}"
                      },
                      "environment": {
                        "variables": {
                          "BUCKET": "${bucket}"
                        }
                      }
                    }
                  ]
                },
                "taskCount": 6,
                "parallelism": 2
              },
              "logsPolicy": {
                "destination": "CLOUD_LOGGING"
              }
            }
          },
          "result": "createAndRunBatchJobResponse"
        }
      },
      {
        "logDeleteBatchJob": {
          "call": "sys.log",
          "args": {
            "data": "${\"Deleting the batch job \" + jobId}"
          }
        }
      },
      {
        "deleteBatchJob": {
          "call": "googleapis.batch.v1.projects.locations.jobs.delete",
          "args": {
            "name": "${\"projects/\" + projectId + \"/locations/\" + region + \"/jobs/\" + jobId}"
          },
          "result": "deleteResult"
        }
      },
      {
        "returnResult": {
          "return": {
            "jobId": "${jobId}",
            "bucket": "${bucket}"
          }
        }
      }
    ]
  }
}

  3. Deploy the workflow by entering the following command:

    gcloud workflows deploy batch-workflow \
  --source=batch-workflow.yaml \
  --location=us-central1 \
  --service-account=SERVICE_ACCOUNT_NAME@PROJECT_ID.iam.gserviceaccount.com

    Replace SERVICE_ACCOUNT_NAME with the name of the service account you previously created.

Execute the workflow

Executing a workflow runs the current workflow definition associated with the workflow.

Console

  1. In the Google Cloud console, go to the Workflows page.

    Go to Workflows

  2. On the Workflows page, click the batch-workflow workflow to go to its details page.

  3. On the Workflow details page, click Execute.

  4. Click Execute again.

    The workflow execution should take a few minutes.

  5. View the results of the workflow in the Output pane.

    The results should look similar to the following:

    {
  "bucket": "project-name-job-primegen-TIMESTAMP",
  "jobId": "job-primegen-TIMESTAMP"
}

gcloud

  1. Execute the workflow:

    gcloud workflows run batch-workflow \
  --location=us-central1

    The workflow execution should take a few minutes.

  2. You can check the status of a long running execution.

  3. To get the status of the last completed execution, run the following command:

    gcloud workflows executions describe-last

    The results should be similar to the following:

    name: projects/PROJECT_NUMBER/locations/us-central1/workflows/batch-workflow/executions/EXECUTION_ID
result: '{"bucket":"project-name-job-primegen-TIMESTAMP","jobId":"job-primegen-TIMESTAMP"}'
startTime: '2022-07-29T16:08:39.725306421Z'
state: SUCCEEDED
status:
  currentSteps:
  - routine: main
    step: returnResult
workflowRevisionId: 000001-9ba

List the objects in the output bucket

You can confirm that the results are as expected by listing the objects in your Cloud Storage output bucket.

Console

  1. In the Google Cloud console, go to the Cloud Storage Buckets page.

    Go to Buckets

  2. In the bucket list, click on the name of the bucket whose contents you want to view.

    The results should be similar to the following, with six files in total, and each listing a batch of 10,000 prime numbers:

    primes-1-10000.txt
primes-10001-20000.txt
primes-20001-30000.txt
primes-30001-40000.txt
primes-40001-50000.txt
primes-50001-60000.txt

gcloud

  1. Retrieve your output bucket name:

    gcloud storage ls

    The output is similar to the following: 

    gs://PROJECT_ID-job-primegen-TIMESTAMP/

  2. List the objects in your output bucket:

    gcloud storage ls gs://PROJECT_ID-job-primegen-TIMESTAMP/** --recursive

    Replace TIMESTAMP with the timestamp returned by the previous command.

    The output should be similar to the following, with six files in total, and each listing a batch of 10,000 prime numbers:

    gs://project-name-job-primegen-TIMESTAMP/primes-1-10000.txt
gs://project-name-job-primegen-TIMESTAMP/primes-10001-20000.txt
gs://project-name-job-primegen-TIMESTAMP/primes-20001-30000.txt
gs://project-name-job-primegen-TIMESTAMP/primes-30001-40000.txt
gs://project-name-job-primegen-TIMESTAMP/primes-40001-50000.txt
gs://project-name-job-primegen-TIMESTAMP/primes-50001-60000.txt