Mailing List Archive

Hi,

Actually this is the default JVM, so its not OpenJ9 or another EA
release.It could be one of the tests haging, but we can't figure that out.

P.S.: Jenkins kills jobs, if they take longer than usual it kills it (it
has no hard limit, it takes the average time of previous runs and if one
takes much longer it kills).

Uwe

Am 14.11.2023 um 11:06 schrieb Michael McCandless:
> Hmm build timed out -- not sure why it's taking so long to run tests:
> Build timed out (after 137 minutes). Marking the build as aborted.
> Build timed out (after 137 minutes). Marking the build as failed.
> Mike McCandless
>
> http://blog.mikemccandless.com
>
>
> On Tue, Nov 14, 2023 at 1:39?AM Policeman Jenkins Server
> <jenkins@thetaphi.de> wrote:
>
> Build: https://jenkins.thetaphi.de/job/Lucene-main-Linux/45536/
> Java: 64bit/hotspot/jdk-17.0.9 -XX:+UseCompressedOops
> -XX:+UseShenandoahGC
>
> All tests passed
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: builds-unsubscribe@lucene.apache.org
> For additional commands, e-mail: builds-help@lucene.apache.org
>
--
Uwe Schindler
Achterdiek 19, D-28357 Bremen
https://www.thetaphi.de
eMail:uwe@thetaphi.de

Thanks Uwe.

OK so this might just be a high-sigma outlier-ish event due to unluckily
slow seed selection?

I wonder whether the distribution of total run time of each full "./gradlew
test" on each JVM configuration is roughly Gaussian-ish?

Mike McCandless

http://blog.mikemccandless.com


On Tue, Nov 14, 2023 at 5:40?AM Uwe Schindler <uwe@thetaphi.de> wrote:

> Hi,
>
> Actually this is the default JVM, so its not OpenJ9 or another EA
> release.It could be one of the tests haging, but we can't figure that out.
>
> P.S.: Jenkins kills jobs, if they take longer than usual it kills it (it
> has no hard limit, it takes the average time of previous runs and if one
> takes much longer it kills).
>
> Uwe
>
> Am 14.11.2023 um 11:06 schrieb Michael McCandless:
>
> Hmm build timed out -- not sure why it's taking so long to run tests:
>
> Build timed out (after 137 minutes). Marking the build as aborted.
> Build timed out (after 137 minutes). Marking the build as failed.
>
> Mike McCandless
>
> http://blog.mikemccandless.com
>
>
> On Tue, Nov 14, 2023 at 1:39?AM Policeman Jenkins Server <
> jenkins@thetaphi.de> wrote:
>
>> Build: https://jenkins.thetaphi.de/job/Lucene-main-Linux/45536/
>> Java: 64bit/hotspot/jdk-17.0.9 -XX:+UseCompressedOops -XX:+UseShenandoahGC
>>
>> All tests passed
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: builds-unsubscribe@lucene.apache.org
>> For additional commands, e-mail: builds-help@lucene.apache.org
>
> --
> Uwe Schindler
> Achterdiek 19, D-28357 Bremenhttps://www.thetaphi.de
> eMail: uwe@thetaphi.de
>
>

Hi,

It could also be a JVM bug! This happened the second time with Temurin
Hotspot JDK 17.0.9:
https://jenkins.thetaphi.de/job/Lucene-main-Linux/45532/ and
https://jenkins.thetaphi.de/job/Lucene-main-Linux/45536/

Both times it hung in phonetic tests...

Uwe

Am 14.11.2023 um 11:46 schrieb Michael McCandless:
> Thanks Uwe.
>
> OK so this might just be a high-sigma outlier-ish event due to
> unluckily slow seed selection?
>
> I wonder whether the distribution of total run time of each full
> "./gradlew test" on each JVM configuration is roughly Gaussian-ish?
>
> Mike McCandless
>
> http://blog.mikemccandless.com
>
>
> On Tue, Nov 14, 2023 at 5:40?AM Uwe Schindler <uwe@thetaphi.de> wrote:
>
> Hi,
>
> Actually this is the default JVM, so its not OpenJ9 or another EA
> release.It could be one of the tests haging, but we can't figure
> that out.
>
> P.S.: Jenkins kills jobs, if they take longer than usual it kills
> it (it has no hard limit, it takes the average time of previous
> runs and if one takes much longer it kills).
>
> Uwe
>
> Am 14.11.2023 um 11:06 schrieb Michael McCandless:
>> Hmm build timed out -- not sure why it's taking so long to run
>> tests:
>> Build timed out (after 137 minutes). Marking the build as aborted.
>> Build timed out (after 137 minutes). Marking the build as failed.
>> Mike McCandless
>>
>> http://blog.mikemccandless.com
>>
>>
>> On Tue, Nov 14, 2023 at 1:39?AM Policeman Jenkins Server
>> <jenkins@thetaphi.de> wrote:
>>
>> Build: https://jenkins.thetaphi.de/job/Lucene-main-Linux/45536/
>> Java: 64bit/hotspot/jdk-17.0.9 -XX:+UseCompressedOops
>> -XX:+UseShenandoahGC
>>
>> All tests passed
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: builds-unsubscribe@lucene.apache.org
>> For additional commands, e-mail: builds-help@lucene.apache.org
>>
> --
> Uwe Schindler
> Achterdiek 19, D-28357 Bremen
> https://www.thetaphi.de
> eMail:uwe@thetaphi.de
>
--
Uwe Schindler
Achterdiek 19, D-28357 Bremen
https://www.thetaphi.de
eMail:uwe@thetaphi.de

Hi,

Sorry, we do not know where it hangs. Phonetic was already finished. As
tasks may run in parallel, we have no idea where it hung.

Uwe

Am 14.11.2023 um 11:56 schrieb Uwe Schindler:
>
> Hi,
>
> It could also be a JVM bug! This happened the second time with Temurin
> Hotspot JDK 17.0.9:
> https://jenkins.thetaphi.de/job/Lucene-main-Linux/45532/ and
> https://jenkins.thetaphi.de/job/Lucene-main-Linux/45536/
>
> Both times it hung in phonetic tests...
>
> Uwe
>
> Am 14.11.2023 um 11:46 schrieb Michael McCandless:
>> Thanks Uwe.
>>
>> OK so this might just be a high-sigma outlier-ish event due to
>> unluckily slow seed selection?
>>
>> I wonder whether the distribution of total run time of each full
>> "./gradlew test" on each JVM configuration is roughly Gaussian-ish?
>>
>> Mike McCandless
>>
>> http://blog.mikemccandless.com
>>
>>
>> On Tue, Nov 14, 2023 at 5:40?AM Uwe Schindler <uwe@thetaphi.de> wrote:
>>
>> Hi,
>>
>> Actually this is the default JVM, so its not OpenJ9 or another EA
>> release.It could be one of the tests haging, but we can't figure
>> that out.
>>
>> P.S.: Jenkins kills jobs, if they take longer than usual it kills
>> it (it has no hard limit, it takes the average time of previous
>> runs and if one takes much longer it kills).
>>
>> Uwe
>>
>> Am 14.11.2023 um 11:06 schrieb Michael McCandless:
>>> Hmm build timed out -- not sure why it's taking so long to run
>>> tests:
>>> Build timed out (after 137 minutes). Marking the build as aborted.
>>> Build timed out (after 137 minutes). Marking the build as failed.
>>> Mike McCandless
>>>
>>> http://blog.mikemccandless.com
>>>
>>>
>>> On Tue, Nov 14, 2023 at 1:39?AM Policeman Jenkins Server
>>> <jenkins@thetaphi.de> wrote:
>>>
>>> Build: https://jenkins.thetaphi.de/job/Lucene-main-Linux/45536/
>>> Java: 64bit/hotspot/jdk-17.0.9 -XX:+UseCompressedOops
>>> -XX:+UseShenandoahGC
>>>
>>> All tests passed
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: builds-unsubscribe@lucene.apache.org
>>> For additional commands, e-mail: builds-help@lucene.apache.org
>>>
>> --
>> Uwe Schindler
>> Achterdiek 19, D-28357 Bremen
>> https://www.thetaphi.de
>> eMail:uwe@thetaphi.de
>>
> --
> Uwe Schindler
> Achterdiek 19, D-28357 Bremen
> https://www.thetaphi.de
> eMail:uwe@thetaphi.de

--
Uwe Schindler
Achterdiek 19, D-28357 Bremen
https://www.thetaphi.de
eMail:uwe@thetaphi.de

> P.S.: Jenkins kills jobs, if they take longer than usual it kills it (it
> has no hard limit, it takes the average time of previous runs and if one
> takes much longer it kills).
>

From what I see here -

https://plugins.jenkins.io/build-timeout/

it should be possible to tweak jenkins to use an absolute timeout... and
even if not, provide a shell script to perhaps jps all java processes so
that there's more diagnostics?

D.

>

It is absolutely NOT wanted to have absolute timeouts on Policeman
Jenkins. This was hanging really, so it was killed for correct reasons.

Uwe

Am 14.11.2023 um 12:19 schrieb Dawid Weiss:
>
> P.S.: Jenkins kills jobs, if they take longer than usual it kills
> it (it has no hard limit, it takes the average time of previous
> runs and if one takes much longer it kills).
>
>
> From what I see here -
>
> https://plugins.jenkins.io/build-timeout/
>
> it should be possible to tweak jenkins to use an absolute timeout...
> and even if not, provide a shell script to perhaps jps all java
> processes so that there's more diagnostics?
>
> D.
>
--
Uwe Schindler
Achterdiek 19, D-28357 Bremen
https://www.thetaphi.de
eMail:uwe@thetaphi.de