[jira] [Commented] (BEAM-6405) Improve PortableValidatesRunner test reliability on Jenkins

2019-01-16 Thread Maximilian Michels (JIRA)


[ 
https://issues.apache.org/jira/browse/BEAM-6405?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16744608#comment-16744608
 ] 

Maximilian Michels commented on BEAM-6405:
--

Streaming execution looks very good: 
https://builds.apache.org/job/beam_PostCommit_Java_PVR_Flink_Streaming/
Batch remains flaky due to memory problems: 
https://builds.apache.org/job/beam_PostCommit_Java_PVR_Flink_Batch/

This PR should get rid of the memory problems: 
https://github.com/apache/beam/pull/7515

> Improve PortableValidatesRunner test reliability on Jenkins
> ---
>
> Key: BEAM-6405
> URL: https://issues.apache.org/jira/browse/BEAM-6405
> Project: Beam
>  Issue Type: Test
>  Components: runner-flink
>Reporter: Maximilian Michels
>Assignee: Maximilian Michels
>Priority: Major
>  Time Spent: 5h
>  Remaining Estimate: 0h
>
> The PVR tests seem to be passing fine and then failing consecutively for no 
> reason: https://builds.apache.org/job/beam_PostCommit_Java_PVR_Flink_Batch/ 
> It looks like the outrageous parallelism, i.e. number of available cores, is 
> responsible for the flakiness if there is additional load on the build 
> slaves. We should lower the parallelism.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)


[jira] [Commented] (BEAM-6405) Improve PortableValidatesRunner test reliability on Jenkins

2019-01-10 Thread Maximilian Michels (JIRA)


[ 
https://issues.apache.org/jira/browse/BEAM-6405?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16739478#comment-16739478
 ] 

Maximilian Michels commented on BEAM-6405:
--

CC [~thw]

> Improve PortableValidatesRunner test reliability on Jenkins
> ---
>
> Key: BEAM-6405
> URL: https://issues.apache.org/jira/browse/BEAM-6405
> Project: Beam
>  Issue Type: Test
>  Components: runner-flink
>Reporter: Maximilian Michels
>Assignee: Maximilian Michels
>Priority: Major
>
> The PVR tests seem to be passing fine and then failing consecutively for no 
> reason: https://builds.apache.org/job/beam_PostCommit_Java_PVR_Flink_Batch/ 
> It looks like the outrageous parallelism, i.e. number of available cores, is 
> responsible for the flakiness if there is additional load on the build 
> slaves. We should lower the parallelism.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)