[
https://issues.apache.org/jira/browse/ZOOKEEPER-1932?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15398460#comment-15398460
]
Michael Han commented on ZOOKEEPER-1932:
----------------------------------------
Hi [~fpj] I saw you mentioned that we could remove the this test because the
code it is testing, the LeaderElection is deprecated. Should we just do that
for this test, or you prefer we keep this test for now and remove it along with
the LE code in future releases? What others think about this proposal?
Anyway, the curios part of me looked at the problem briefly today and it can be
a little bit easier reproducible if we increase the thread count, from 30 to
60, and can be hardly reproducible if I decrease the thread count from 30 to 5.
Thread dump indicates that one or more LE threads were in WAITING state which
implies they never bail out (that they never finds a leader for some reasons).
Sounds another threading related issue.
> org.apache.zookeeper.test.LETest.testLE fails once in a while
> -------------------------------------------------------------
>
> Key: ZOOKEEPER-1932
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-1932
> Project: ZooKeeper
> Issue Type: Bug
> Components: leaderElection
> Affects Versions: 3.5.0
> Reporter: Michi Mutsuzaki
> Assignee: Flavio Junqueira
> Fix For: 3.5.3, 3.6.0
>
> Attachments: TEST-org.apache.zookeeper.test.LETest.txt
>
>
> org.apache.zookeeper.test.LETest.testLE is failing on trunk once in a while.
> I'm not able to reproduce the failure on my box. I looked at the log, but I
> couldn't quite figure out what's going on.
> https://builds.apache.org/view/S-Z/view/ZooKeeper/job/ZooKeeper-trunk/2315/testReport/
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
