[ 
https://issues.apache.org/jira/browse/BEAM-8345?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16944894#comment-16944894
 ] 

Valentyn Tymofieiev commented on BEAM-8345:
-------------------------------------------

I think we are missing Flink/Spark runner validations in automated scripts. 
When interacting with Flink/Spark, we can run Beam SDK harness in a container 
or local process. Container execution can also help validate  the  container 
images we release. 

I manually tried validating Flink/Spark using commands at the end of this 
message. However that sequence of commands may not be very robust since the 
output of wordcount is available only inside a container 
(https://issues.apache.org/jira/browse/BEAM-7857), and the container seems to 
be garbage-collected shortly after execution. Also, we need to pay attention to 
which version of Flink we are using. I tried 1.5, but we may need to test 
against all versions we support.   

Related issue that we may need to address: 
https://issues.apache.org/jira/browse/BEAM-7858.

{noformat}
git checkout tags/v2.16.0-RC1

./gradlew :runners:spark:job-server:runShadow   or ./gradlew 
:runners:flink:1.5:job-server:runShadow

In a separate terminal: 

docker pull apachebeam/python3.5_sdk:2.16.0_rc1
docker tag apachebeam/python3.5_sdk:2.16.0_rc1 apachebeam/python3.5_sdk:2.16.0

./gradlew :sdks:python:test-suites:portable:py35:portableWordCountBatch  
-PjobEndpoint=localhost:8099 -PenvirionmentType=DOCKER 

As soon as that is done:
docker ps -a  | grep apachebeam
Note the container ID of a running container. It stays around only for a minute 
or so, and gets garbage-collected...
docker exec 866fb8932207 /bin/bash -c 'cat /tmp/py-wordcount*'

<wordcount output is printed>
{noformat}

> Add missing validations to run_rc_validation.sh
> -----------------------------------------------
>
>                 Key: BEAM-8345
>                 URL: https://issues.apache.org/jira/browse/BEAM-8345
>             Project: Beam
>          Issue Type: Sub-task
>          Components: testing
>            Reporter: Mark Liu
>            Assignee: Mark Liu
>            Priority: Major
>          Time Spent: 10m
>  Remaining Estimate: 0h
>




--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to