Re: [rdf4j-dev] Jenkins build became unstable: rdf4j-develop-deploy #68

[Håvard Ottestad <hmottestad@xxxxxxxxx>]

Hi Jeen,

Thank you for finding this. 

We’ve had problems with this test before. Last time it was due to the NativeStore sharing models between connections to ensure serializable, which was causing an issue with the TreeModel because it will lazily creating indexes which would result in a concurrent modification exception when done in parallel. 

I’ll look at this. Hopefully I can reproduce it on my laptop. 

Håvard

On 30 Apr 2020, at 01:13, Jeen Broekstra <jeen.broekstra@xxxxxxxxx> wrote:



Havard, this one is actually somewhat interesting, as it fails on a test in the SHACL Sail somewhere. See https://ci.eclipse.org/rdf4j/job/rdf4j-develop-deploy/org.eclipse.rdf4j$rdf4j-shacl/68/testReport/junit/org.eclipse.rdf4j.sail.shacl/MultithreadedNativeStoreTest/testLotsOfValidationFailuresSerializable/.

I haven't looked in detail but if that is a test that fails intermittently (and clearly it doesn't always fail as Github actions is not reporting a problem on develop), there might be some sort of race condition in how you handle multithreaded access in the SHACL Sail.

Cheers,

Jeen

On Thu, Apr 30, 2020 at 6:08 AM <ci-admin@xxxxxxxxxxx> wrote:

See <https://ci.eclipse.org/rdf4j/job/rdf4j-develop-deploy/68/display/redirect>

_______________________________________________
rdf4j-dev mailing list
rdf4j-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/rdf4j-dev

_______________________________________________
rdf4j-dev mailing list
rdf4j-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/rdf4j-dev