accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Christopher Tubbs (JIRA)" <j...@apache.org>
Subject [jira] [Resolved] (ACCUMULO-4430) FastBulkImportIT does not always pass
Date Wed, 04 Apr 2018 21:57:00 GMT

     [ https://issues.apache.org/jira/browse/ACCUMULO-4430?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Christopher Tubbs resolved ACCUMULO-4430.
-----------------------------------------
       Resolution: Cannot Reproduce
    Fix Version/s:     (was: 1.9.0)
                       (was: 2.0.0)

Haven't seen this in awhile

> FastBulkImportIT does not always pass
> -------------------------------------
>
>                 Key: ACCUMULO-4430
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-4430
>             Project: Accumulo
>          Issue Type: Bug
>          Components: test
>            Reporter: Christopher Tubbs
>            Priority: Major
>
> I occasionally get the following failure:
> {code}
> -------------------------------------------------------------------------------
> Test set: org.apache.accumulo.test.performance.metadata.FastBulkImportIT
> -------------------------------------------------------------------------------
> Tests run: 1, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 64.649 sec <<<
FAILURE! - in org.apache.accumulo.test.performance.metadata.FastBulkImportIT
> test(org.apache.accumulo.test.performance.metadata.FastBulkImportIT)  Time elapsed: 64.648
sec  <<< FAILURE!
> java.lang.AssertionError
>         at org.apache.accumulo.test.performance.metadata.FastBulkImportIT.test(FastBulkImportIT.java:109)
> {code}
> The test is checking that the bulk import completes in less than 30 seconds with 5 assignment/process
threads, no compactions, and no WAL. That number seems arbitrary, and dependent on many conditions
which the test cannot control for. A hard-coded expectation of less than 30 seconds seems
unreasonable to me.
> Is there a better way we can test that the TServer doesn't re-read bulk-loaded flags,
perhaps with a mocked unit test?



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Mime
View raw message