使用云功能执行数据流模板时出错

Question

尝试使用 Google Cloud 功能执行自定义数据流模板时出现以下错误。

Error:"problem running dataflow template, error was: { Error: A Forbidden error was returned while trying to retrieve an access token for the Compute Engine built-in service account. 这可能是因为 Compute Engine 实例没有指定了正确的权限范围。无法刷新访问令牌”。

我已尝试提供所有必需的权限和范围。请有人提出解决方案。

Answer 1

google-cloud 节点库尚不支持 Dataflow API，因此当前使用该 API 的方式是 googleapis 库。

按照那里的说明，我尝试使用 HTTP 触发的函数使用 Google-provided 模板启动 Dataflow 作业，但没有遇到任何问题：

const {google} = require('googleapis');
const project = "your-project-id"

exports.launchDataflowTemplate = (req, res) => {
    let result;
    google.auth.getApplicationDefault(function(err, authClient, projectId) {
            if (err) {
                throw err;
            }
            if (authClient.createScopedRequired && authClient.createScopedRequired()) {
                authClient = authClient.createScoped([
                    'https://www.googleapis.com/auth/cloud-platform',
                    'https://www.googleapis.com/auth/compute',
                    'https://www.googleapis.com/auth/compute.readonly',
                    'https://www.googleapis.com/auth/userinfo.email'
                ]);
            }
            var dataflow = google.dataflow({
                version: "v1b3",
                auth: authClient
            });

            var launchParams = {
                "inputFilePattern": "gs://your-input-bucket/*.gz",
                "outputDirectory": "gs://your-result-bucket/",
                "outputFailureFile": "gs://your-logs-bucket/error.csv"
            };

            var env = {
               "tempLocation": "gs://your-staging-bucket/temp",
               "zone": "us-central1-f"
            }


            var opts = {
                projectId: project,
                gcsPath: "gs://dataflow-templates/latest/Bulk_Decompress_GCS_Files",
                resource: {
                    parameters: launchParams,
                    environment: env
                }
            };

            dataflow.projects.templates.launch(opts, (err, result) => {
                if (err) {
                    throw err;
                }
                res.send(result.data);
            });
    });
};