On Mon, 26 Nov 2018 at 10:07, Axel Lebourhis <axel.lebourhis@linaro.org> wrote:
>
> Hi Milosz,
>
> You will find the patch attached.

Thanks. I tried applying it, but I think I have a more flexible
solution. Please check here:
https://review.linaro.org/#/c/qa/test-definitions/+/29209/

To use it add JAVA_OPTIONS to your lava job definition. Example below

- test:
  definitions:
    - from: git
      name: cts-part2-arm64-v8a
      params:
        ANDROID_VERSION: aosp-master
        FAILURES_PRINTED: '300'
        TEST_PARAMS: cts --abi arm64-v8a --disable-reboot
--skip-preconditions --skip-device-info
        TEST_URL: <some_url_to_download_cts.zip>
        JAVA_OPTIONS: "-Xmx700M"
      path: automated/android/noninteractive-tradefed/tradefed.yaml
      repository: https://git.linaro.org/qa/test-definitions.git
      namespace: tlxc
      timeout: {minutes: 505}

milosz

>
> Regards,
> Axel
>
> On Mon, 26 Nov 2018 at 09:39, Milosz Wasilewski <milosz.wasilewski@linaro.org> wrote:
>>
>> On Mon, 26 Nov 2018 at 09:37, Axel Lebourhis <axel.lebourhis@linaro.org> wrote:
>> >
>> > Hi everyone,
>> >
>> > I'm currently running CTS tests on imx8mq on Android 9.
>> > My test job is based on Linaro's test definition located here :
>> > https://git.linaro.org/qa/test-definitions.git/tree/automated/android/tradefed
>> >
>> > I had issue concerning the Java heap size, causing the test to abort and
>> > skip a lot of tests.
>> > So, I cloned this git tree and made modifications in "tradefed.sh" :
>> >
>> >
>> > export _JAVA_OPTIONS="-Xmx350M"
>> >
>> > replaced by :
>> >
>> > export _JAVA_OPTIONS="-Xmx700M"
>> >
>> > Since this modification, I'm able to run the full CTS 9.
>> > So, I think you should consider modify this script, so other users
>> > can benefit from it.
>>
>> Axel,
>>
>> Do you mind sending me a patch? I can merge it to the repository.
>>
>> milosz
>>
>> >
>> > Best regards,
>> >
>> > Axel
>> >
>> >
>> > _______________________________________________
>> > Lava-users mailing list
>> > Lava-users@lists.lavasoftware.org
>> > https://lists.lavasoftware.org/mailman/listinfo/lava-users