Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Flume, mail # dev - Review Request: TestLoadBalancingSinkProcessor.testRoundRobinBackoffFailureRecovery is flaky, fails every once in a while...


Copy link to this message
-
Review Request: TestLoadBalancingSinkProcessor.testRoundRobinBackoffFailureRecovery is flaky, fails every once in a while...
Ted Malaska 2012-09-11, 15:29

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/7020/
-----------------------------------------------------------

Review request for Flume.
Description
-------

It is possible that it is a clock issue.

because s2 has to wait 2000 mil/s because of the following line
state.restoreTime = now + Math.min(MAX_TIMEOUT, 1000 * (1 << state.sequentialFails));

But then the test waits 2000 mil/s before starting the loop again.  

There is a chance that we reach the following if statement at only 2000 mil/s passing

if (state.restoreTime < curTime) {

So I'm going to increase the wait to 2001 mil/s.  That should do it.
This addresses bug FLUME-1562.
    https://issues.apache.org/jira/browse/FLUME-1562
Diffs
-----

  flume-ng-core/src/test/java/org/apache/flume/sink/TestLoadBalancingSinkProcessor.java 981f88e

Diff: https://reviews.apache.org/r/7020/diff/
Testing
-------
Thanks,

Ted Malaska