[
https://issues.apache.org/jira/browse/ACCUMULO-4110?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15378430#comment-15378430
]
Michael Wall commented on ACCUMULO-4110:
----------------------------------------
After some research, some testing and some discussion on IRC (thanks
[~elserj]), I am going to say this is a timing issue. When enough mutations
are written, the TabletServerLogger calls createLogger which is synchronized
and calls startLogMaker which starts a thread that is constantly trying to open
the next WAL so we have a standy. When the test fails as shown above, we just
happen to be counting the logs before the standby is finished being created.
Inspecting zookeeper after these failures show there is actually 3 WALs, we
just counted 2.
Going to put a loop around counting these on error try again. Josh, feel free
to correct anything I have wrong or add more details as necessary
> WALSunnyDayIT is repeatedly failing
> -----------------------------------
>
> Key: ACCUMULO-4110
> URL: https://issues.apache.org/jira/browse/ACCUMULO-4110
> Project: Accumulo
> Issue Type: Bug
> Components: test
> Reporter: Josh Elser
> Assignee: Michael Wall
> Priority: Blocker
> Fix For: 1.8.0
>
>
> {code}
> Tests run: 1, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 13.035 sec
> <<< FAILURE! - in org.apache.accumulo.test.functional.WALSunnyDayIT
> test(org.apache.accumulo.test.functional.WALSunnyDayIT) Time elapsed: 13.034
> sec <<< FAILURE!
> java.lang.AssertionError: should have 3 WALs after roll expected:<3> but
> was:<2>
> at org.junit.Assert.fail(Assert.java:88)
> at org.junit.Assert.failNotEquals(Assert.java:834)
> at org.junit.Assert.assertEquals(Assert.java:645)
> at
> org.apache.accumulo.test.functional.WALSunnyDayIT.test(WALSunnyDayIT.java:113)
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
