Uptick in nightly failures of TestLatLonLine*Queries.testRandomBig

[Chris Hostetter <ho...@fucit.org>]

I don't know what might be the cause/corrolation of this, and i haven't 
tried to investigate the logs in depth, but i thought i should point it out...

In the past week, the failure rate of these 2 nightly tests has increased 
~20-25% as compared to the prior 3 (calendar) weeks...

Class: org.apache.lucene.document.TestLatLonLineShapeQueries
Method: testRandomBig
Historic Failures: 22.35% (19 / 85)
Recent Failures: 46.67% (7 / 15)

Class: org.apache.lucene.document.TestLatLonPolygonShapeQueries
Method: testRandomBig
Historic Failures: 14.10% (11 / 78)
Recent Failures: 35.71% (5 / 14)

...all of the recent failures have come from sarowe's jenkins box ... but 
i'm not sure if that indicates a problem with running these tests on his 
machine, or if he's just running nightly tests more often then other 
jenkins hosts?

http://fucit.org/solr-jenkins-reports/suspicious-failure-report.html



-Hoss
http://www.lucidworks.com/

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org

Re: Uptick in nightly failures of TestLatLonLine*Queries.testRandomBig

[Dawid Weiss <da...@gmail.com>]

> I looked at the logs and these all are suite timeouts; other Lucene tests also affected, e.g. Test2BPostings and TestGeo3DPoint.

These are extremely long sometimes... Maybe we can lower their complexity a bit?

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org

Re: Uptick in nightly failures of TestLatLonLine*Queries.testRandomBig

[Steve Rowe <sa...@gmail.com>]

> On Dec 14, 2018, at 1:05 PM, Chris Hostetter <ho...@fucit.org> wrote:
> ...all of the recent failures have come from sarowe's jenkins box ... but i'm not sure if that indicates a problem with running these tests on his machine, or if he's just running nightly tests more often then other jenkins hosts?

Thanks for the heads-up Hoss.

I looked at the logs and these all are suite timeouts; other Lucene tests also affected, e.g. Test2BPostings and TestGeo3DPoint.  My Jenkins runs module testing in parallel, so I'd guess in these cases that there are (way) more active threads than CPU cores, which would slow tests down.

I also use the box my Jenkins runs on for other tasks, which could be another cause of resourse starvation here.

I'll take a look at how the Nightly jobs run on my Jenkins and see if I can improve the situation.

Steve


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org