[
https://issues.apache.org/jira/browse/HBASE-19288?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16258430#comment-16258430
]
Eshcar Hillel commented on HBASE-19288:
---------------------------------------
I agree setting the in-memory flush factor to 0.25 in this test should fix the
problem.
With flush size 500, when the factor is 0.25 then in-memory flush is invoked at
125, with the new default it is invoked at 10 and hence the first assertion
failure.
Setting the counter to 0 at the beginning of the test avoids the assertions
upon retries.
+1
> Intermittent test failure in TestHStore.testRunDoubleMemStoreCompactors
> -----------------------------------------------------------------------
>
> Key: HBASE-19288
> URL: https://issues.apache.org/jira/browse/HBASE-19288
> Project: HBase
> Issue Type: Test
> Components: test
> Reporter: Ted Yu
> Assignee: Ted Yu
> Fix For: 2.0.0-beta-1
>
> Attachments: 19288.v1.txt, 19288.v2.txt,
> testRunDoubleMemStoreCompactors.out
>
>
> Here was one of the test failures:
> https://builds.apache.org/job/PreCommit-HBASE-Build/9812/testReport/junit/org.apache.hadoop.hbase.regionserver/TestHStore/testRunDoubleMemStoreCompactors/
>
> {code}
> [ERROR]
> org.apache.hadoop.hbase.regionserver.TestHStore.testRunDoubleMemStoreCompactors(org.apache.hadoop.hbase.regionserver.TestHStore)
> [ERROR] Run 1: TestHStore.testRunDoubleMemStoreCompactors:1500 expected:<2>
> but was:<3>
> [ERROR] Run 2: TestHStore.testRunDoubleMemStoreCompactors:1481 expected:<1>
> but was:<4>
> [ERROR] Run 3: TestHStore.testRunDoubleMemStoreCompactors:1481 expected:<1>
> but was:<5>
> {code}
> From the counts for second and third runs, we know that RUNNER_COUNT was not
> cleared in between the reruns, leading to failure at the 1st assertion.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
