It seems that Flink cluster does not retrieve program arguments correctly. For
reference, I sent the following request:
Method Type: POST
URL:
http://dbtpa05p.ch3.dev.i.com:9171/v1/jars/321febd8-a5e8-4255-858b-c221b49aef18_aa-lmx-etl-1.0.0.1-SNAPSHOT.jar/run
Body: {"programArgs" :
"--job=mediaSpent,--initialScopeId=b494c35d-4c37-4338-8d23-0fc947bef690,--integratedScopeId=91769bd8-df4d-436c-b8d0-2e23ce862859,--projectId=333,--log.path=../log"}
Content-Type: application/json
From: Chesnay Schepler <ches...@apache.org>
Sent: Πέμπτη, 14 Μαρτίου 2019 2:24 μμ
To: Papadopoulos, Konstantinos <konstantinos.papadopou...@iriworldwide.com>;
user@flink.apache.org
Subject: Re: ProgramInvocationException when trying to submit a job by running
a jar using Monitoring REST API
Please enable debug logging, re-submit the job, check the JobManager logs and
report back the logged exception.
On 14.03.2019 13:16, Papadopoulos, Konstantinos wrote:
Hi all,
As part of our projects, I experiment with Flink Monitoring REST API and,
especially, its capabilities of uploading and running jar files.
When I am trying to submit one of our jobs by running a jar previously uploaded
via '/jars/upload', I am getting an 500 Internal Server Error response with the
following body:
{ "errors": [ "org.apache.flink.client.program.ProgramInvocationException: The
main method caused an error." ] }
On the other hand, when I am trying the same with the 'Word Count' batch
example, execution succeeds.
It must be pointed out that I tried to execute the respective POST request both
via 'curl' and Postman, but both failed with the same error.
Do anyone have idea why this may happen?
Thanks in advance,
Konstantinos
