[ 
https://issues.apache.org/jira/browse/BEAM-13073?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17448193#comment-17448193
 ] 

Valentyn Tymofieiev commented on BEAM-13073:
--------------------------------------------

Thanks, [~labianchin] for making a PR. I am trying to setup A/B test.

Looking at the Job logs in 
https://ci-beam.apache.org/job/beam_LoadTests_Java_ParDo_Dataflow_V2_Batch_Java11/159/consoleFull
 to fish out a gradle command, and will try to run the command from the same 
image that we use to run tests on Jenkins, to minimize environment issues:

{noformat}
gcloud compute instances create valentyn-jenkins-image 
--project=apache-beam-testing --zone us-central1-b 
--image-family=jenkins-worker-boot-image --machine-type=n1-highmem-4
gcloud compute ssh valentyn-jenkins-image --project=apache-beam-testing --zone 
us-central1-b         
{noformat}

Need to see if the tests rebuild the container from HEAD or I also need to 
rebuild the container.
 

> Unexpected GC when using Java 11
> --------------------------------
>
>                 Key: BEAM-13073
>                 URL: https://issues.apache.org/jira/browse/BEAM-13073
>             Project: Beam
>          Issue Type: Bug
>          Components: sdk-java-harness
>            Reporter: Luis
>            Assignee: Kenneth Knowles
>            Priority: P1
>              Labels: java11, java9, performance
>             Fix For: 2.35.0
>
>         Attachments: perf_regression_java_11.png
>
>          Time Spent: 50m
>  Remaining Estimate: 0h
>
> Beam SDK has been supporting Java 11 for a while (I guess the support was 
> introduced here https://issues.apache.org/jira/browse/BEAM-2530). 
> Unfortunately, in Spotify we are still experiencing performance issues when 
> using Beam SDK 2.32, Google Dataflow and Java 11.
> Thanks to [~emilyye] and [~iht], they confirmed JVM 11 is using SerialGC, 
> while Java 8 uses ParallelGC. It sounds like ParallelGC is a good option for 
> high throughput / low latency jobs. For Java11 we'd expect to use G1GC or 
> ParallelGC.
> This SO question [1] clarifies that JVM chooses SerialGC when it treats the 
> machine as a "client". It looks like the Java SDK container could benefit 
> from using `-XX:+AlwaysActAsServerClassMachine`. Is that correct?
> Let me know if the ticket needs further context or adjustment. (It is my 
> first time creating a ticket here).
>  [1] 
> [https://stackoverflow.com/questions/52474162/why-is-serialgc-chosen-over-g1gc]



--
This message was sent by Atlassian Jira
(v8.20.1#820001)

Reply via email to