You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@beam.apache.org by Gareth Western <ga...@garethwestern.com> on 2017/01/23 20:21:53 UTC

Debugging a failed Dataflow job

I'm having trouble running my pipeline using the dataflow runner. The 
job is submitted successfully:

Dataflow SDK version: 0.4.0
Submitted job: 2017-01-23_12_17_20-13351949104581541182

but in the Dataflow console it simply says " workflow failed". I can't 
seem to find any more details regarding the cause of the failure. Where 
should I be looking?


Re: Debugging a failed Dataflow job

Posted by Gareth Western <ga...@garethwestern.com>.
Thanks Chris. That's useful to know.

I cross-posted the question to SO [1] and was told that the error 
message was because I had forgotten to authorise the dataflow API for 
this project in the Cloud Console (PEBCAK).

1. 
https://stackoverflow.com/questions/41815519/cloud-dataflow-job-failed-without-reason/


On 23. jan. 2017 23:34, Chris Fei wrote:
> I believe that documentation refers to the 1.X releases of the Google 
> Dataflow SDK. BlockingDataflowPipelineRunner exists there, but it 
> doesn't exist in Beam 0.4.0. Take a look at the breaking changes for 
> the 2.X releases at 
> https://cloud.google.com/dataflow/release-notes/release-notes-java-2, 
> which outline how to migrate away from BlockingDataflowPipelineRunner.
>
> You should just be able to specify DataflowRunner in your options, and 
> then add the --blockOnRun option.
>
> Chris
>
>
> On Mon, Jan 23, 2017, at 05:03 PM, Gareth Western wrote:
>>
>> Thank you. How does one use the BlockingDataflowPipelineRunner with 
>> Beam?
>>
>> Specifying it in the PipelineOptions results in an Exception with a 
>> message stating that it is not one of the supported pipeline runners:
>>
>> "Exception in thread "main" java.lang.IllegalArgumentException: Class 
>> 'com.google.cloud.dataflow.sdk.runners.BlockingDataflowPipelineRunner' 
>> does not implement PipelineRunner. Supported pipeline runners 
>> [DataflowRunner, DirectRunner, FlinkRunner, TestFlinkRunner]"
>>
>>
>> On 23. jan. 2017 22:31, Lukasz Cwik wrote:
>>> Please take a look at 
>>> https://cloud.google.com/dataflow/pipelines/logging#monitoring-pipeline-logs
>>>
>>> On Mon, Jan 23, 2017 at 12:21 PM, Gareth Western 
>>> <gareth@garethwestern.com <ma...@garethwestern.com>> wrote:
>>>
>>>     I'm having trouble running my pipeline using the dataflow
>>>     runner. The job is submitted successfully:
>>>
>>>     Dataflow SDK version: 0.4.0
>>>     Submitted job: 2017-01-23_12_17_20-13351949104581541182
>>>
>>>     but in the Dataflow console it simply says " workflow failed". I
>>>     can't seem to find any more details regarding the cause of the
>>>     failure. Where should I be looking?
>>>
>


Re: Debugging a failed Dataflow job

Posted by Chris Fei <cf...@gmail.com>.
I believe that documentation refers to the 1.X releases of the Google
Dataflow SDK. BlockingDataflowPipelineRunner exists there, but it
doesn't exist in Beam 0.4.0. Take a look at the breaking changes for the
2.X releases at
https://cloud.google.com/dataflow/release-notes/release-notes-java-2,
which outline how to migrate away from BlockingDataflowPipelineRunner.


You should just be able to specify DataflowRunner in your options, and
then add the --blockOnRun option.


Chris





On Mon, Jan 23, 2017, at 05:03 PM, Gareth Western wrote:

> Thank you. How does one use the BlockingDataflowPipelineRunner
> with Beam?
> Specifying it in the PipelineOptions results in an Exception with a
> message stating that it is not one of the supported pipeline runners:
> "Exception in thread "main" java.lang.IllegalArgumentException: Class
> 'com.google.cloud.dataflow.sdk.runners.BlockingDataflowPipelineRunner'
> does not implement PipelineRunner. Supported pipeline runners
> [DataflowRunner, DirectRunner, FlinkRunner, TestFlinkRunner]"
> 



> On 23. jan. 2017 22:31, Lukasz Cwik wrote:

>> Please take a look at
>> https://cloud.google.com/dataflow/pipelines/logging#monitoring-pipeline-logs
>> 

>> On Mon, Jan 23, 2017 at 12:21 PM, Gareth Western
>> <ga...@garethwestern.com> wrote:
>> 

>>> I'm having trouble running my pipeline using the dataflow runner.
>>> The job is submitted successfully:
>>> 

>>>  Dataflow SDK version: 0.4.0

>>>  Submitted job: 2017-01-23_12_17_20-13351949104581541182

>>> 

>>>  but in the Dataflow console it simply says " workflow failed". I
>>>  can't seem to find any more details regarding the cause of the
>>>  failure. Where should I be looking?
>>> 



Re: Debugging a failed Dataflow job

Posted by Gareth Western <ga...@garethwestern.com>.
Thank you. How does one use the BlockingDataflowPipelineRunner with Beam?

Specifying it in the PipelineOptions results in an Exception with a 
message stating that it is not one of the supported pipeline runners:

"Exception in thread "main" java.lang.IllegalArgumentException: Class 
'com.google.cloud.dataflow.sdk.runners.BlockingDataflowPipelineRunner' 
does not implement PipelineRunner. Supported pipeline runners 
[DataflowRunner, DirectRunner, FlinkRunner, TestFlinkRunner]"


On 23. jan. 2017 22:31, Lukasz Cwik wrote:
> Please take a look at 
> https://cloud.google.com/dataflow/pipelines/logging#monitoring-pipeline-logs
>
> On Mon, Jan 23, 2017 at 12:21 PM, Gareth Western 
> <gareth@garethwestern.com <ma...@garethwestern.com>> wrote:
>
>     I'm having trouble running my pipeline using the dataflow runner.
>     The job is submitted successfully:
>
>     Dataflow SDK version: 0.4.0
>     Submitted job: 2017-01-23_12_17_20-13351949104581541182
>
>     but in the Dataflow console it simply says " workflow failed". I
>     can't seem to find any more details regarding the cause of the
>     failure. Where should I be looking?
>
>


Re: Debugging a failed Dataflow job

Posted by Lukasz Cwik <lc...@google.com>.
Please take a look at
https://cloud.google.com/dataflow/pipelines/logging#monitoring-pipeline-logs

On Mon, Jan 23, 2017 at 12:21 PM, Gareth Western <ga...@garethwestern.com>
wrote:

> I'm having trouble running my pipeline using the dataflow runner. The job
> is submitted successfully:
>
> Dataflow SDK version: 0.4.0
> Submitted job: 2017-01-23_12_17_20-13351949104581541182
>
> but in the Dataflow console it simply says " workflow failed". I can't
> seem to find any more details regarding the cause of the failure. Where
> should I be looking?
>
>