go - "No filesystem found for scheme gs"在谷歌云平台运行数据流时

Question

我在谷歌云平台(GCP)中运行我的谷歌数据流作业。
当我在本地运行此作业时，它运行良好，但是在 GCP 上运行时，出现此错误
“java.lang.IllegalArgumentException:未找到方案 gs 的文件系统”。
我可以访问那个谷歌云 URI，我可以将我的 jar 文件上传到那个 URI，我可以看到我的本地工作的一些临时文件。

我在 GCP 中的工作 ID:

2019-08-08_21_47_27-162804342585245230(梁版:2.12.0)

2019-08-09_16_41_15-11728697820819900062(梁版:2.14.0)

我尝试过 2.12.0 和 2.14.0 的梁版本，它们都有相同的错误。

java.lang.IllegalArgumentException: No filesystem found for scheme gs
    at org.apache.beam.sdk.io.FileSystems.getFileSystemInternal(FileSystems.java:456)
    at org.apache.beam.sdk.io.FileSystems.matchNewResource(FileSystems.java:526)
    at org.apache.beam.sdk.io.gcp.bigquery.BigQueryHelpers.resolveTempLocation(BigQueryHelpers.java:689)
    at org.apache.beam.sdk.io.gcp.bigquery.BigQuerySourceBase.extractFiles(BigQuerySourceBase.java:125)
    at org.apache.beam.sdk.io.gcp.bigquery.BigQuerySourceBase.split(BigQuerySourceBase.java:148)
    at org.apache.beam.runners.dataflow.worker.WorkerCustomSources.splitAndValidate(WorkerCustomSources.java:284)
    at org.apache.beam.runners.dataflow.worker.WorkerCustomSources.performSplitTyped(WorkerCustomSources.java:206)
    at org.apache.beam.runners.dataflow.worker.WorkerCustomSources.performSplitWithApiLimit(WorkerCustomSources.java:190)
    at org.apache.beam.runners.dataflow.worker.WorkerCustomSources.performSplit(WorkerCustomSources.java:169)
    at org.apache.beam.runners.dataflow.worker.WorkerCustomSourceOperationExecutor.execute(WorkerCustomSourceOperationExecutor.java:78)
    at org.apache.beam.runners.dataflow.worker.BatchDataflowWorker.executeWork(BatchDataflowWorker.java:412)
    at org.apache.beam.runners.dataflow.worker.BatchDataflowWorker.doWork(BatchDataflowWorker.java:381)
    at org.apache.beam.runners.dataflow.worker.BatchDataflowWorker.getAndPerformWork(BatchDataflowWorker.java:306)
    at org.apache.beam.runners.dataflow.worker.DataflowBatchWorkerHarness$WorkerThread.doWork(DataflowBatchWorkerHarness.java:135)
    at org.apache.beam.runners.dataflow.worker.DataflowBatchWorkerHarness$WorkerThread.call(DataflowBatchWorkerHarness.java:115)
    at org.apache.beam.runners.dataflow.worker.DataflowBatchWorkerHarness$WorkerThread.call(DataflowBatchWorkerHarness.java:102)
    at java.util.concurrent.FutureTask.run(FutureTask.java:266)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
    at java.lang.Thread.run(Thread.java:745)

Answer 1

如果您构建一个捆绑所有依赖项的“胖 jar”，这可能是由几个问题引起的。

您必须包含依赖项 org.apache.beam:google-cloud-platform-core 拥有 Beam GCS 文件系统。

在远 jar 中，您必须保留 META-INF/services/org.apache.beam.sdk.io.FileSystemRegistrar带一行的文件 org.apache.beam.sdk.extensions.gcp.storage.GcsFileSystemRegistrar .您可以在步骤 1 的 jar 中找到此文件。您的依赖项中可能有许多同名文件，注册了不同的 Beam 文件系统。您需要配置 maven 或 gradle 以将它们组合为构建的一部分，否则它们会相互覆盖而无法正常工作。