RareClusterStateIT#testDelayedMappingPropagationOnPrimary can fail with assertion error.

[jtibshirani]

I wasn't able to reproduce this error locally. Although it's only failed once, I opted to file an issue because there have been some recent changes to the test related to Zen2.

This looks related to #36813, feel free to close this issue if you'd prefer to track work there.

---

Link to the build: https://elasticsearch-ci.elastic.co/job/elastic+elasticsearch+master+matrix-java-periodic/ES_BUILD_JAVA=openjdk12,ES_RUNTIME_JAVA=zulu11,nodes=immutable&&linux&&docker/349/console

Command to reproduce:

./gradlew :server:integTest --tests "org.elasticsearch.cluster.coordination.RareClusterStateIT.testDelayedMappingPropagationOnPrimary" \
  -Dtests.seed=332AB212F4B698FB \
  -Dtests.security.manager=true \
  -Dtests.locale=kw-GB \
  -Dtests.timezone=America/North_Dakota/Center \
  -Dcompiler.java=12 \
  -Druntime.java=11

Relevant excerpt from the logs:

org.elasticsearch.cluster.coordination.RareClusterStateIT > testDelayedMappingPropagationOnPrimary FAILED
    java.util.concurrent.TimeoutException: Timeout waiting for task.
        at __randomizedtesting.SeedInfo.seed([332AB212F4B698FB:459F59389E75EA12]:0)
        at org.elasticsearch.common.util.concurrent.BaseFuture$Sync.get(BaseFuture.java:236)
        at org.elasticsearch.common.util.concurrent.BaseFuture.get(BaseFuture.java:65)
        at org.elasticsearch.cluster.coordination.RareClusterStateIT.lambda$testDelayedMappingPropagationOnPrimary$3(RareClusterStateIT.java:271)
        at org.elasticsearch.test.ESTestCase.assertBusy(ESTestCase.java:850)
        at org.elasticsearch.test.ESTestCase.assertBusy(ESTestCase.java:836)
        at org.elasticsearch.cluster.coordination.RareClusterStateIT.testDelayedMappingPropagationOnPrimary(RareClusterStateIT.java:269)

[elasticmachine]

Pinging @elastic/es-distributed

[ywelsch]

@original-brownbear as this test has been muted for a while and there is no recent failures with logs, and I can't reproduce this locally, should we enable the test again on master to get a recent failure?

[original-brownbear]

Sure I opened #41884

[jaymode]

This popped up today in a PR build, the build scan is https://gradle.com/s/nczuoemcggsl2.

[ywelsch]

This popped up today in a PR build, the build scan is https://gradle.com/s/nczuoemcggsl2.

This failed in a different place I think, and it does not make sense to me how it could fail there (it's an assertNotNull in an assertBusy that's failing, and the assertBusy does not seem to be repeated..., the test finishes in a few seconds, much faster than the assertBusy which typically takes 10 seconds to retry). We will need another failure to be reported here.

[original-brownbear]

@ywelsch see #42430 I think I fixed this one in there too (I had the replica version of the test assigned to myself).