-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/41226/#review109984
-----------------------------------------------------------


Can you take a stab at illustrating in the description how this problem 
manifests?  The current shape of the patch feels like it papers over a problem 
that we might be left to solve again, so thinking about the problem at a high 
level might help us think about a broader fix.

- Bill Farner


On Dec. 10, 2015, 5:33 p.m., Maxim Khutornenko wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/41226/
> -----------------------------------------------------------
> 
> (Updated Dec. 10, 2015, 5:33 p.m.)
> 
> 
> Review request for Aurora and Bill Farner.
> 
> 
> Bugs: AURORA-1506 and AURORA-1507
>     https://issues.apache.org/jira/browse/AURORA-1506
>     https://issues.apache.org/jira/browse/AURORA-1507
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> Task event processing delays may generate duplicate evaluation results in the 
> `InstanceActionHandler`. This leads to occasional stack traces in the log but 
> is generally benign as `JobUpdateEventSubscriber` catches these errors. 
> 
> The easiest fix is to check for the task existence before attempting to 
> kill/add instance.
> 
> 
> Diffs
> -----
> 
>   
> src/main/java/org/apache/aurora/scheduler/updater/InstanceActionHandler.java 
> d8686f1f1b53e5ff2791663489e24c342503831e 
>   src/test/java/org/apache/aurora/scheduler/updater/AddTaskTest.java 
> 0583a63a175880355a7296ebd1c6e6fb5dc99f38 
>   src/test/java/org/apache/aurora/scheduler/updater/KillTaskTest.java 
> PRE-CREATION 
> 
> Diff: https://reviews.apache.org/r/41226/diff/
> 
> 
> Testing
> -------
> 
> ./gradlew -Pq build
> 
> 
> Thanks,
> 
> Maxim Khutornenko
> 
>

Reply via email to