气流异常: DataFlow失败，返回代码1

Question

我试图通过气流脚本执行数据流jar。为此，我正在使用DataFlowJavaOperator。在param中，当我试图运行以下作业时，我将传递本地system.But中显示的可执行jar文件的路径。

{gcp_dataflow_hook.py:108} INFO - Start waiting for DataFlow process to complete.
[2017-09-12 16:59:38,225] {models.py:1417} ERROR - DataFlow failed with return code 1
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/airflow/models.py", line 1374, in run
    result = task_copy.execute(context=context)
  File "/usr/lib/python2.7/site-packages/airflow/contrib/operators/dataflow_operator.py", line 116, in execute
    hook.start_java_dataflow(self.task_id, dataflow_options, self.jar)
  File "/usr/lib/python2.7/site-packages/airflow/contrib/hooks/gcp_dataflow_hook.py", line 146, in start_java_dataflow
    task_id, variables, dataflow, name, ["java", "-jar"])
  File "/usr/lib/python2.7/site-packages/airflow/contrib/hooks/gcp_dataflow_hook.py", line 138, in _start_dataflow
    _Dataflow(cmd).wait_for_done()
  File "/usr/lib/python2.7/site-packages/airflow/contrib/hooks/gcp_dataflow_hook.py", line 119, in wait_for_done
    self._proc.returncode))
Exception: DataFlow failed with return code 1`

我的气流脚本是：

from airflow.contrib.operators.dataflow_operator import DataFlowJavaOperator
from airflow.contrib.hooks.gcs_hook import GoogleCloudStorageHook
from airflow.models import BaseOperator
from airflow.utils.decorators import apply_defaults
from datetime import datetime, timedelta


default_args = {
'owner': 'airflow',
'start_date': datetime(2017, 03, 16),
'email': [<EmailID>],

'dataflow_default_options': {
        'project': '<ProjectId>',
       # 'zone': 'europe-west1-d', (i am not sure what should i pass here)
        'stagingLocation': 'gs://spark_3/staging/'
    }
 }

 dag = DAG('Dataflow',schedule_interval=timedelta(minutes=2), 
 default_args=default_args)

 dataflow1 = DataFlowJavaOperator(
 task_id='dataflow_example',
 jar ='/root/airflow_scripts/csvwriter.jar',
 gcp_conn_id  = 'GCP_smoke', 
 dag=dag)

我不知道我犯了什么错误，有人能帮我摆脱这件事吗？

Note :I am creating this jar while selecting option as Runnable JAR file by packaging all the external dependencies.

Answer 1

问题是我用的那个罐子。在使用jar之前，请确保jar按预期执行。

示例：(如果您的jar是dataflow_job1.jar )，使用

java -jar dataflow_job_1.jar --parameters_if_any

一旦您的jar成功运行，继续在气流DataflowJavaOperator jar中使用jar。

此外，，，如果您遇到与编码器相关的错误，您可能需要自己的编码器来执行代码。例如，我对TableRow类有一个问题，因为它没有默认的编码器，因此我不得不这样做：

TableRowCoder：

public class TableRowCoder extends Coder<TableRow> {
private static final long serialVersionUID = 1L;
private static final Coder<TableRow> tableRow = TableRowJsonCoder.of();
@Override
public void encode(TableRow value, OutputStream outStream) throws CoderException, IOException {
    tableRow.encode(value, outStream);

}
@Override
public TableRow decode(InputStream inStream) throws CoderException, IOException {
    return new TableRow().set("F1", tableRow.decode(inStream));
}
@Override
public List<? extends Coder<?>> getCoderArguments() {
    // TODO Auto-generated method stub
    return null;
}
@Override
public void verifyDeterministic() throws org.apache.beam.sdk.coders.Coder.NonDeterministicException {


}
}

然后在您的代码中注册此编码器

pipeline.getCoderRegistry().registerCoderForClass(TableRow.class, new TableRowCoder())

如果仍然存在错误(与编码器无关)，请导航到：

*.jar\META-INF\services\FileSystemRegistrar 

并添加可能发生的任何依赖项。

例如，可能存在一个暂存错误，因为：

Unable to find registrar for gs

我必须添加以下一行才能使其工作。

org.apache.beam.sdk.extensions.gcp.storage.GcsFileSystemRegistrar