You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@geode.apache.org by Dan Smith <ds...@pivotal.io> on 2016/01/26 21:47:09 UTC

Review Request 42822: GEODE-864: Waiting for conflation thread in conflation tests

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/42822/
-----------------------------------------------------------

Review request for geode, Barry Oglesby and xiaojian zhou.


Repository: geode


Description
-------

For parallel queues, there is a actually a separate conflation thread
that removes entries from the queue on conflation. The tests need to
wait for this thread to finish it's work.

As part of this change, I refactored the ParallelWANConflationDUnitTest
to remove a bunch of duplicate code, replaced the VM.invoke(String)
calls with lambdas, and scaled the test down to create fewer buckets and
do fewer puts so that it runs in 1/4 of the time.


Diffs
-----

  gemfire-wan/src/test/java/com/gemstone/gemfire/internal/cache/wan/WANTestBase.java 230c8d8ce6f903cf5e4b9c8bf42b87ca04271647 
  gemfire-wan/src/test/java/com/gemstone/gemfire/internal/cache/wan/parallel/ParallelWANConflationDUnitTest.java 4acaaf4c5f24123db335e39ea56592717c9ea17c 

Diff: https://reviews.apache.org/r/42822/diff/


Testing
-------


Thanks,

Dan Smith


Re: Review Request 42822: GEODE-864: Waiting for conflation thread in conflation tests

Posted by Dan Smith <ds...@pivotal.io>.

> On Jan. 26, 2016, 11:26 p.m., xiaojian zhou wrote:
> > gemfire-wan/src/test/java/com/gemstone/gemfire/internal/cache/wan/WANTestBase.java, line 2165
> > <https://reviews.apache.org/r/42822/diff/1/?file=1221832#file1221832line2165>
> >
> >     There is a pauseSender() and pauseSenderAndWaitForDispatcherToPause(). You just made the 2 methods identical. 
> >     
> >     Was that purposely using pauseSender() here?
> >     
> >     Others look good.

Could catch! Before my change, the test was calling pauseSender and then waiting for 3 seconds, so I think the intent was to wait for it to pause. I'll merge those duplicate methods.


- Dan


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/42822/#review116478
-----------------------------------------------------------


On Jan. 26, 2016, 8:47 p.m., Dan Smith wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/42822/
> -----------------------------------------------------------
> 
> (Updated Jan. 26, 2016, 8:47 p.m.)
> 
> 
> Review request for geode, Barry Oglesby and xiaojian zhou.
> 
> 
> Repository: geode
> 
> 
> Description
> -------
> 
> For parallel queues, there is a actually a separate conflation thread
> that removes entries from the queue on conflation. The tests need to
> wait for this thread to finish it's work.
> 
> As part of this change, I refactored the ParallelWANConflationDUnitTest
> to remove a bunch of duplicate code, replaced the VM.invoke(String)
> calls with lambdas, and scaled the test down to create fewer buckets and
> do fewer puts so that it runs in 1/4 of the time.
> 
> 
> Diffs
> -----
> 
>   gemfire-wan/src/test/java/com/gemstone/gemfire/internal/cache/wan/WANTestBase.java 230c8d8ce6f903cf5e4b9c8bf42b87ca04271647 
>   gemfire-wan/src/test/java/com/gemstone/gemfire/internal/cache/wan/parallel/ParallelWANConflationDUnitTest.java 4acaaf4c5f24123db335e39ea56592717c9ea17c 
> 
> Diff: https://reviews.apache.org/r/42822/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Dan Smith
> 
>


Re: Review Request 42822: GEODE-864: Waiting for conflation thread in conflation tests

Posted by xiaojian zhou <zh...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/42822/#review116478
-----------------------------------------------------------




gemfire-wan/src/test/java/com/gemstone/gemfire/internal/cache/wan/WANTestBase.java (line 2165)
<https://reviews.apache.org/r/42822/#comment177495>

    There is a pauseSender() and pauseSenderAndWaitForDispatcherToPause(). You just made the 2 methods identical. 
    
    Was that purposely using pauseSender() here?
    
    Others look good.


- xiaojian zhou


On Jan. 26, 2016, 8:47 p.m., Dan Smith wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/42822/
> -----------------------------------------------------------
> 
> (Updated Jan. 26, 2016, 8:47 p.m.)
> 
> 
> Review request for geode, Barry Oglesby and xiaojian zhou.
> 
> 
> Repository: geode
> 
> 
> Description
> -------
> 
> For parallel queues, there is a actually a separate conflation thread
> that removes entries from the queue on conflation. The tests need to
> wait for this thread to finish it's work.
> 
> As part of this change, I refactored the ParallelWANConflationDUnitTest
> to remove a bunch of duplicate code, replaced the VM.invoke(String)
> calls with lambdas, and scaled the test down to create fewer buckets and
> do fewer puts so that it runs in 1/4 of the time.
> 
> 
> Diffs
> -----
> 
>   gemfire-wan/src/test/java/com/gemstone/gemfire/internal/cache/wan/WANTestBase.java 230c8d8ce6f903cf5e4b9c8bf42b87ca04271647 
>   gemfire-wan/src/test/java/com/gemstone/gemfire/internal/cache/wan/parallel/ParallelWANConflationDUnitTest.java 4acaaf4c5f24123db335e39ea56592717c9ea17c 
> 
> Diff: https://reviews.apache.org/r/42822/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Dan Smith
> 
>


Re: Review Request 42822: GEODE-864: Waiting for conflation thread in conflation tests

Posted by Dan Smith <ds...@pivotal.io>.

> On Jan. 26, 2016, 11:06 p.m., Barry Oglesby wrote:
> > I think you can remove the 'WANTestBase' from the lambda calls too since WANTestBase is a super class of the test class.
> > 
> > from: vm4.invoke(WANTestBase.class, "enableConflation", new Object[] { "ln" });
> > to: vm4.invoke(() -> WANTestBase.enableConflation( "ln" ));
> > to: vm4.invoke(() -> enableConflation( "ln" ));
> > 
> > I'm surprised you left the 'KBKBKB' log messages in there. :-)

Ok, I'll make those changes and push it. Thanks!


- Dan


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/42822/#review116471
-----------------------------------------------------------


On Jan. 26, 2016, 8:47 p.m., Dan Smith wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/42822/
> -----------------------------------------------------------
> 
> (Updated Jan. 26, 2016, 8:47 p.m.)
> 
> 
> Review request for geode, Barry Oglesby and xiaojian zhou.
> 
> 
> Repository: geode
> 
> 
> Description
> -------
> 
> For parallel queues, there is a actually a separate conflation thread
> that removes entries from the queue on conflation. The tests need to
> wait for this thread to finish it's work.
> 
> As part of this change, I refactored the ParallelWANConflationDUnitTest
> to remove a bunch of duplicate code, replaced the VM.invoke(String)
> calls with lambdas, and scaled the test down to create fewer buckets and
> do fewer puts so that it runs in 1/4 of the time.
> 
> 
> Diffs
> -----
> 
>   gemfire-wan/src/test/java/com/gemstone/gemfire/internal/cache/wan/WANTestBase.java 230c8d8ce6f903cf5e4b9c8bf42b87ca04271647 
>   gemfire-wan/src/test/java/com/gemstone/gemfire/internal/cache/wan/parallel/ParallelWANConflationDUnitTest.java 4acaaf4c5f24123db335e39ea56592717c9ea17c 
> 
> Diff: https://reviews.apache.org/r/42822/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Dan Smith
> 
>


Re: Review Request 42822: GEODE-864: Waiting for conflation thread in conflation tests

Posted by Barry Oglesby <bo...@pivotal.io>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/42822/#review116471
-----------------------------------------------------------


Ship it!




I think you can remove the 'WANTestBase' from the lambda calls too since WANTestBase is a super class of the test class.

from: vm4.invoke(WANTestBase.class, "enableConflation", new Object[] { "ln" });
to: vm4.invoke(() -> WANTestBase.enableConflation( "ln" ));
to: vm4.invoke(() -> enableConflation( "ln" ));

I'm surprised you left the 'KBKBKB' log messages in there. :-)

- Barry Oglesby


On Jan. 26, 2016, 8:47 p.m., Dan Smith wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/42822/
> -----------------------------------------------------------
> 
> (Updated Jan. 26, 2016, 8:47 p.m.)
> 
> 
> Review request for geode, Barry Oglesby and xiaojian zhou.
> 
> 
> Repository: geode
> 
> 
> Description
> -------
> 
> For parallel queues, there is a actually a separate conflation thread
> that removes entries from the queue on conflation. The tests need to
> wait for this thread to finish it's work.
> 
> As part of this change, I refactored the ParallelWANConflationDUnitTest
> to remove a bunch of duplicate code, replaced the VM.invoke(String)
> calls with lambdas, and scaled the test down to create fewer buckets and
> do fewer puts so that it runs in 1/4 of the time.
> 
> 
> Diffs
> -----
> 
>   gemfire-wan/src/test/java/com/gemstone/gemfire/internal/cache/wan/WANTestBase.java 230c8d8ce6f903cf5e4b9c8bf42b87ca04271647 
>   gemfire-wan/src/test/java/com/gemstone/gemfire/internal/cache/wan/parallel/ParallelWANConflationDUnitTest.java 4acaaf4c5f24123db335e39ea56592717c9ea17c 
> 
> Diff: https://reviews.apache.org/r/42822/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Dan Smith
> 
>