[jira] [Commented] (FLINK-5006) SystemProcessingTimeServiceTest.testTimerSorting fails
[
https://issues.apache.org/jira/browse/FLINK-5006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15672075#comment-15672075
]
ASF GitHub Bot commented on FLINK-5006:
---
Github user asfgit closed the pull request at:
https://github.com/apache/flink/pull/2785
> SystemProcessingTimeServiceTest.testTimerSorting fails
> --
>
> Key: FLINK-5006
> URL: https://issues.apache.org/jira/browse/FLINK-5006
> Project: Flink
> Issue Type: Improvement
> Components: Streaming
>Reporter: Ufuk Celebi
> Labels: test-stability
>
> {code}
> testTimerSorting(org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest)
> Time elapsed: 0.023 sec <<< FAILURE!
> java.lang.AssertionError: expected:<1478173518115> but was:<1478173518122>
> 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.junit.Assert.assertEquals(Assert.java:631)
> at
> org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest.testTimerSorting(SystemProcessingTimeServiceTest.java:298)
> {code}
> Failed in a private branch with unrelated changes (the test is very much self
> contained).
> https://s3.amazonaws.com/archive.travis-ci.org/jobs/172910645/log.txt
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (FLINK-5006) SystemProcessingTimeServiceTest.testTimerSorting fails
[
https://issues.apache.org/jira/browse/FLINK-5006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15671485#comment-15671485
]
ASF GitHub Bot commented on FLINK-5006:
---
Github user StephanEwen commented on the issue:
https://github.com/apache/flink/pull/2785
Looks good, merging this!
Thanks, @BorisOsipov for bearing with us on this one through the long
discussion!
> SystemProcessingTimeServiceTest.testTimerSorting fails
> --
>
> Key: FLINK-5006
> URL: https://issues.apache.org/jira/browse/FLINK-5006
> Project: Flink
> Issue Type: Improvement
> Components: Streaming
>Reporter: Ufuk Celebi
> Labels: test-stability
>
> {code}
> testTimerSorting(org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest)
> Time elapsed: 0.023 sec <<< FAILURE!
> java.lang.AssertionError: expected:<1478173518115> but was:<1478173518122>
> 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.junit.Assert.assertEquals(Assert.java:631)
> at
> org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest.testTimerSorting(SystemProcessingTimeServiceTest.java:298)
> {code}
> Failed in a private branch with unrelated changes (the test is very much self
> contained).
> https://s3.amazonaws.com/archive.travis-ci.org/jobs/172910645/log.txt
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (FLINK-5006) SystemProcessingTimeServiceTest.testTimerSorting fails
[
https://issues.apache.org/jira/browse/FLINK-5006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15666568#comment-15666568
]
ASF GitHub Bot commented on FLINK-5006:
---
Github user BorisOsipov commented on the issue:
https://github.com/apache/flink/pull/2785
@aljoscha thank you for additional research and comments.
@StephanEwen I think we can remove it as I proposed earlier.
I'v removed test and added javadoc to registerTimer method. Please look at
this.
Is it ok?
> SystemProcessingTimeServiceTest.testTimerSorting fails
> --
>
> Key: FLINK-5006
> URL: https://issues.apache.org/jira/browse/FLINK-5006
> Project: Flink
> Issue Type: Improvement
> Components: Streaming
>Reporter: Ufuk Celebi
> Labels: test-stability
>
> {code}
> testTimerSorting(org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest)
> Time elapsed: 0.023 sec <<< FAILURE!
> java.lang.AssertionError: expected:<1478173518115> but was:<1478173518122>
> 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.junit.Assert.assertEquals(Assert.java:631)
> at
> org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest.testTimerSorting(SystemProcessingTimeServiceTest.java:298)
> {code}
> Failed in a private branch with unrelated changes (the test is very much self
> contained).
> https://s3.amazonaws.com/archive.travis-ci.org/jobs/172910645/log.txt
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (FLINK-5006) SystemProcessingTimeServiceTest.testTimerSorting fails
[
https://issues.apache.org/jira/browse/FLINK-5006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15657226#comment-15657226
]
ASF GitHub Bot commented on FLINK-5006:
---
Github user StephanEwen commented on the issue:
https://github.com/apache/flink/pull/2785
Okay, if that is the case, then we should probably simply remove the "test
timer sorting" test and note down that this does not give any guarantees on the
order of timers.
@BorisOsipov What do you think about this?
> SystemProcessingTimeServiceTest.testTimerSorting fails
> --
>
> Key: FLINK-5006
> URL: https://issues.apache.org/jira/browse/FLINK-5006
> Project: Flink
> Issue Type: Improvement
> Components: Streaming
>Reporter: Ufuk Celebi
> Labels: test-stability
>
> {code}
> testTimerSorting(org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest)
> Time elapsed: 0.023 sec <<< FAILURE!
> java.lang.AssertionError: expected:<1478173518115> but was:<1478173518122>
> 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.junit.Assert.assertEquals(Assert.java:631)
> at
> org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest.testTimerSorting(SystemProcessingTimeServiceTest.java:298)
> {code}
> Failed in a private branch with unrelated changes (the test is very much self
> contained).
> https://s3.amazonaws.com/archive.travis-ci.org/jobs/172910645/log.txt
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (FLINK-5006) SystemProcessingTimeServiceTest.testTimerSorting fails
[
https://issues.apache.org/jira/browse/FLINK-5006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15657138#comment-15657138
]
ASF GitHub Bot commented on FLINK-5006:
---
Github user aljoscha commented on the issue:
https://github.com/apache/flink/pull/2785
@StephanEwen It's used to drive the automatic watermark generation and the
inactivity timer in the `BucketingSink`, but those are cases where the order
doesn't matter.
> SystemProcessingTimeServiceTest.testTimerSorting fails
> --
>
> Key: FLINK-5006
> URL: https://issues.apache.org/jira/browse/FLINK-5006
> Project: Flink
> Issue Type: Improvement
> Components: Streaming
>Reporter: Ufuk Celebi
> Labels: test-stability
>
> {code}
> testTimerSorting(org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest)
> Time elapsed: 0.023 sec <<< FAILURE!
> java.lang.AssertionError: expected:<1478173518115> but was:<1478173518122>
> 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.junit.Assert.assertEquals(Assert.java:631)
> at
> org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest.testTimerSorting(SystemProcessingTimeServiceTest.java:298)
> {code}
> Failed in a private branch with unrelated changes (the test is very much self
> contained).
> https://s3.amazonaws.com/archive.travis-ci.org/jobs/172910645/log.txt
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (FLINK-5006) SystemProcessingTimeServiceTest.testTimerSorting fails
[
https://issues.apache.org/jira/browse/FLINK-5006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15657114#comment-15657114
]
ASF GitHub Bot commented on FLINK-5006:
---
Github user StephanEwen commented on the issue:
https://github.com/apache/flink/pull/2785
@aljoscha Okay, if that is the only thing that we use it for, then order is
indeed not important for that service. Are we sure it is not used in other
places?
> SystemProcessingTimeServiceTest.testTimerSorting fails
> --
>
> Key: FLINK-5006
> URL: https://issues.apache.org/jira/browse/FLINK-5006
> Project: Flink
> Issue Type: Improvement
> Components: Streaming
>Reporter: Ufuk Celebi
> Labels: test-stability
>
> {code}
> testTimerSorting(org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest)
> Time elapsed: 0.023 sec <<< FAILURE!
> java.lang.AssertionError: expected:<1478173518115> but was:<1478173518122>
> 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.junit.Assert.assertEquals(Assert.java:631)
> at
> org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest.testTimerSorting(SystemProcessingTimeServiceTest.java:298)
> {code}
> Failed in a private branch with unrelated changes (the test is very much self
> contained).
> https://s3.amazonaws.com/archive.travis-ci.org/jobs/172910645/log.txt
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (FLINK-5006) SystemProcessingTimeServiceTest.testTimerSorting fails
[
https://issues.apache.org/jira/browse/FLINK-5006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15657076#comment-15657076
]
ASF GitHub Bot commented on FLINK-5006:
---
Github user aljoscha commented on the issue:
https://github.com/apache/flink/pull/2785
@StephanEwen I think it's not a problem that these physical, low-level
timers don't fire in order of their timestamp because the ordering of timers is
maintained in the implementation of `InternalTimerService` (in a queue in the
current implementation). Also, under normal circumstances we only have one
active timer at the `ProcessingTimeService` (`SystemProcessingTimeService`).
> SystemProcessingTimeServiceTest.testTimerSorting fails
> --
>
> Key: FLINK-5006
> URL: https://issues.apache.org/jira/browse/FLINK-5006
> Project: Flink
> Issue Type: Improvement
> Components: Streaming
>Reporter: Ufuk Celebi
> Labels: test-stability
>
> {code}
> testTimerSorting(org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest)
> Time elapsed: 0.023 sec <<< FAILURE!
> java.lang.AssertionError: expected:<1478173518115> but was:<1478173518122>
> 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.junit.Assert.assertEquals(Assert.java:631)
> at
> org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest.testTimerSorting(SystemProcessingTimeServiceTest.java:298)
> {code}
> Failed in a private branch with unrelated changes (the test is very much self
> contained).
> https://s3.amazonaws.com/archive.travis-ci.org/jobs/172910645/log.txt
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (FLINK-5006) SystemProcessingTimeServiceTest.testTimerSorting fails
[
https://issues.apache.org/jira/browse/FLINK-5006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15656884#comment-15656884
]
ASF GitHub Bot commented on FLINK-5006:
---
Github user StephanEwen commented on the issue:
https://github.com/apache/flink/pull/2785
It's not about timing guarantees here, only about order of running timers.
The problem is that they are sorted by the delay which is affected by
races/stalls, rather than by proper timestamp.
Since Flink actually relies on timer triggering order, I think we need to
implement our own timer service, which sorts by actual timestamp, rather than
by delay.
> SystemProcessingTimeServiceTest.testTimerSorting fails
> --
>
> Key: FLINK-5006
> URL: https://issues.apache.org/jira/browse/FLINK-5006
> Project: Flink
> Issue Type: Improvement
> Components: Streaming
>Reporter: Ufuk Celebi
> Labels: test-stability
>
> {code}
> testTimerSorting(org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest)
> Time elapsed: 0.023 sec <<< FAILURE!
> java.lang.AssertionError: expected:<1478173518115> but was:<1478173518122>
> 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.junit.Assert.assertEquals(Assert.java:631)
> at
> org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest.testTimerSorting(SystemProcessingTimeServiceTest.java:298)
> {code}
> Failed in a private branch with unrelated changes (the test is very much self
> contained).
> https://s3.amazonaws.com/archive.travis-ci.org/jobs/172910645/log.txt
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (FLINK-5006) SystemProcessingTimeServiceTest.testTimerSorting fails
[
https://issues.apache.org/jira/browse/FLINK-5006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15654890#comment-15654890
]
ASF GitHub Bot commented on FLINK-5006:
---
Github user BorisOsipov commented on the issue:
https://github.com/apache/flink/pull/2785
Yes it is. The ThreadPoolExecutor is not claiming to be a real-time system
with any sort of timing guarantees.
The only thing it will guarantee is that it doesn't run tasks too early.
> SystemProcessingTimeServiceTest.testTimerSorting fails
> --
>
> Key: FLINK-5006
> URL: https://issues.apache.org/jira/browse/FLINK-5006
> Project: Flink
> Issue Type: Improvement
> Components: Streaming
>Reporter: Ufuk Celebi
> Labels: test-stability
>
> {code}
> testTimerSorting(org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest)
> Time elapsed: 0.023 sec <<< FAILURE!
> java.lang.AssertionError: expected:<1478173518115> but was:<1478173518122>
> 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.junit.Assert.assertEquals(Assert.java:631)
> at
> org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest.testTimerSorting(SystemProcessingTimeServiceTest.java:298)
> {code}
> Failed in a private branch with unrelated changes (the test is very much self
> contained).
> https://s3.amazonaws.com/archive.travis-ci.org/jobs/172910645/log.txt
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (FLINK-5006) SystemProcessingTimeServiceTest.testTimerSorting fails
[
https://issues.apache.org/jira/browse/FLINK-5006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15654640#comment-15654640
]
ASF GitHub Bot commented on FLINK-5006:
---
Github user StephanEwen commented on the issue:
https://github.com/apache/flink/pull/2785
I guess the problem is that when the tasks get registered at the
`ScheduledThreadPoolExecutor`, multiple tasks may have a delay of zero, which
makes the order of triggering undefined.
What do you think?
> SystemProcessingTimeServiceTest.testTimerSorting fails
> --
>
> Key: FLINK-5006
> URL: https://issues.apache.org/jira/browse/FLINK-5006
> Project: Flink
> Issue Type: Improvement
> Components: Streaming
>Reporter: Ufuk Celebi
> Labels: test-stability
>
> {code}
> testTimerSorting(org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest)
> Time elapsed: 0.023 sec <<< FAILURE!
> java.lang.AssertionError: expected:<1478173518115> but was:<1478173518122>
> 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.junit.Assert.assertEquals(Assert.java:631)
> at
> org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest.testTimerSorting(SystemProcessingTimeServiceTest.java:298)
> {code}
> Failed in a private branch with unrelated changes (the test is very much self
> contained).
> https://s3.amazonaws.com/archive.travis-ci.org/jobs/172910645/log.txt
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (FLINK-5006) SystemProcessingTimeServiceTest.testTimerSorting fails
[
https://issues.apache.org/jira/browse/FLINK-5006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15654613#comment-15654613
]
ASF GitHub Bot commented on FLINK-5006:
---
GitHub user BorisOsipov opened a pull request:
https://github.com/apache/flink/pull/2785
[FLINK-5006] SystemProcessingTimeServiceTest.testTimerSorting fails
Hello!
I'v reproduced this issue on my environment several times.
I found that the test uses ScheduledThreadPoolExecutor as threadFactory
(actually
[SystemProcessingTimeService](https://github.com/apache/flink/blob/master/flink-streaming-java/src/main/java/org/apache/flink/streaming/runtime/tasks/SystemProcessingTimeService.java)
uses). ScheduledThreadPoolExecutor does not guarantee that tasks will be
executed immediately after delay:
> Delayed tasks execute no sooner than they are enabled, but without any
real-time guarantees about when, after they are enabled, they will commence.
See Oracle [docs ](http://bit.ly/2fobpXv)
Thus, order of executions and adding timestamps to ArrayBlockingQueue isn't
defined. So it looks like the test is not correct.
In changes I increased intervals between tasks, it helps to avoid falling.
If I am correct, maybe it makes sense to get rid of this test?
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/BorisOsipov/flink FLINK-5006
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/flink/pull/2785.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #2785
commit d53100c8361551a5eac29b6955ecb0fa93769a53
Author: Boris Osipov
Date: 2016-11-10T15:52:56Z
[FLINK-5006] SystemProcessingTimeServiceTest.testTimerSorting fails
Increases test stability
> SystemProcessingTimeServiceTest.testTimerSorting fails
> --
>
> Key: FLINK-5006
> URL: https://issues.apache.org/jira/browse/FLINK-5006
> Project: Flink
> Issue Type: Improvement
> Components: Streaming
>Reporter: Ufuk Celebi
> Labels: test-stability
>
> {code}
> testTimerSorting(org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest)
> Time elapsed: 0.023 sec <<< FAILURE!
> java.lang.AssertionError: expected:<1478173518115> but was:<1478173518122>
> 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.junit.Assert.assertEquals(Assert.java:631)
> at
> org.apache.flink.streaming.runtime.tasks.SystemProcessingTimeServiceTest.testTimerSorting(SystemProcessingTimeServiceTest.java:298)
> {code}
> Failed in a private branch with unrelated changes (the test is very much self
> contained).
> https://s3.amazonaws.com/archive.travis-ci.org/jobs/172910645/log.txt
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
