-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/27467/#review59655
-----------------------------------------------------------


It seems like exponential backoff is warranted here - in the case that the 
master is completely overloaded there may be confirmation status updates in the 
queue that we haven't seen yet.

Also, do you know off hand what the behavior of the mesos API is to a killTask 
that the master has forgotten about/we've already acked? Will it send a 
TASK_LOST?

- Kevin Sweeney


On Nov. 3, 2014, 1:25 p.m., Bill Farner wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/27467/
> -----------------------------------------------------------
> 
> (Updated Nov. 3, 2014, 1:25 p.m.)
> 
> 
> Review request for Aurora, Kevin Sweeney and Maxim Khutornenko.
> 
> 
> Bugs: AURORA-910
>     https://issues.apache.org/jira/browse/AURORA-910
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> Retry Driver.killTask periodically until the task changes states.
> 
> 
> Diffs
> -----
> 
>   src/main/java/org/apache/aurora/scheduler/async/AsyncModule.java 
> df7e56a955fef21d48e75797bb5761453ae1e57a 
>   src/main/java/org/apache/aurora/scheduler/async/KillRetry.java PRE-CREATION 
>   
> src/test/java/org/apache/aurora/scheduler/async/JobUpdateHistoryPrunerTest.java
>  748aac83db7c470dc7aed7936e629400c3e6c2d8 
>   src/test/java/org/apache/aurora/scheduler/async/KillRetryTest.java 
> PRE-CREATION 
>   
> src/test/java/org/apache/aurora/scheduler/testing/FakeScheduledExecutor.java 
> 1688a339a1a9c4072d7f834963cb5ccad6fe0736 
> 
> Diff: https://reviews.apache.org/r/27467/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Bill Farner
> 
>

Reply via email to