mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Vinod Kone" <vinodk...@gmail.com>
Subject Re: Review Request 34970: Cleaned up and generalized NoExecutorScheduler to be more configurable.
Date Fri, 05 Jun 2015 21:37:04 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/34970/#review86846
-----------------------------------------------------------

Ship it!



src/examples/no_executor_framework.cpp
<https://reviews.apache.org/r/34970/#comment138944>

    I think it is easier to follow if this is written as:
    
    ```
    while(remaining.contains(taskResources) &&
          (totalTasks.isNone() || tasksLaunched < totalTasks.get())) {
          
    }
    ```



src/examples/no_executor_framework.cpp
<https://reviews.apache.org/r/34970/#comment138946>

    LOG(FATAL)? because this is not possible with a command executor?



src/examples/no_executor_framework.cpp
<https://reviews.apache.org/r/34970/#comment138948>

    s/active/launched/ ?



src/examples/no_executor_framework.cpp
<https://reviews.apache.org/r/34970/#comment138943>

    Also mention that, if this is not specified, as many tasks as possible are launched?



src/examples/no_executor_framework.cpp
<https://reviews.apache.org/r/34970/#comment138942>

    Only set this if flags.task_revocable_resources.isSome() ?


- Vinod Kone


On June 5, 2015, 7:56 p.m., Ben Mahler wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/34970/
> -----------------------------------------------------------
> 
> (Updated June 5, 2015, 7:56 p.m.)
> 
> 
> Review request for mesos, Jie Yu and Vinod Kone.
> 
> 
> Bugs: MESOS-2655
>     https://issues.apache.org/jira/browse/MESOS-2655
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Now the NoExecutorScheduler can take a custom command, as well as the resources for the
task and how many tasks to run.
> 
> This allows us to continue using it for an end-to-end test as part of make check, but
also allows it to be used against a cluster in a long-lived fashion.
> 
> 
> Diffs
> -----
> 
>   src/examples/no_executor_framework.cpp 1fb853d6e4a3deb3c36e6e661689697c8ebf898f 
>   src/tests/no_executor_framework_test.sh d2d395595b778bc543e1baaa0fd415dc622b647f 
>   src/tests/script.cpp 515e3141e1d517da4cfc1421d5301d0a3bd6ad51 
> 
> Diff: https://reviews.apache.org/r/34970/diff/
> 
> 
> Testing
> -------
> 
> The existing no executor framework test picks this up.
> 
> Since we do not have an non-zero estimator yet, I modified the slave to send revocable
resources in order to manually test the revocable case.
> 
> 
> Thanks,
> 
> Ben Mahler
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message