mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Greg Mann <g...@mesosphere.io>
Subject Re: Review Request 66143: Refactored agent task launch for better composition [2/2].
Date Wed, 21 Mar 2018 23:22:46 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/66143/#review199705
-----------------------------------------------------------




src/slave/slave.cpp
Line 2202 (original), 2203-2204 (patched)
<https://reviews.apache.org/r/66143/#comment280125>

    You can keep these on a single line:
    
    ```
    Future<Nothing> taskLaunch = collect(unschedules)
      .recover(defer(self(), onUnscheduleGCFailure))
    
      etc...
    ```



src/slave/slave.cpp
Lines 2348 (patched)
<https://reviews.apache.org/r/66143/#comment280114>

    Nit:
    s/_task/task_/
    
    Preceding underscores are only used to avoid shadowing in function parameters. Here and
elsewhere.



src/slave/slave.cpp
Lines 2320-2321 (original), 2369-2371 (patched)
<https://reviews.apache.org/r/66143/#comment280116>

    Would recommend:
    
    ```
      return collect(authorizations)
        .recover(defer(self(),
        
        etc...
    ```



src/slave/slave.cpp
Lines 2325-2327 (original), 2375-2377 (patched)
<https://reviews.apache.org/r/66143/#comment280128>

    Let's do:
    
    ```
              string error =
                "Failed to authorize " + taskOrTaskGroup(task, taskGroup) +
                ": " + future.failure();
    ```



src/slave/slave.cpp
Lines 2386 (patched)
<https://reviews.apache.org/r/66143/#comment280118>

    Let's make this variable a reference instead to avoid the copy:
    
    ```
    list<bool>& authorizations = future.get();
    ```



src/slave/slave.cpp
Lines 2399-2400 (patched)
<https://reviews.apache.org/r/66143/#comment280119>

    Indentation:
    
    ```
                  } else if (executorInfo.has_command() &&
                             executorInfo.command().has_user()) {
    ```
    
    or
    
    ```
                  } else if (
                      executorInfo.has_command() &&
                      executorInfo.command().has_user()) {
    ```



src/slave/slave.cpp
Lines 2404-2406 (patched)
<https://reviews.apache.org/r/66143/#comment280120>

    Let's do:
    ```
                  string error =
                    "Task '" + stringify(_task.task_id()) + "'"
                    " is not authorized to launch as user '" + user + "'";
    ```



src/tests/slave_tests.cpp
Line 4137 (original), 4140 (patched)
<https://reviews.apache.org/r/66143/#comment280136>

    s/tie reaching the critical moment when to kill the task to `_run`./so that we can control
when the task is killed./
    
    Here and below.


- Greg Mann


On March 20, 2018, 9:47 p.m., Meng Zhu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/66143/
> -----------------------------------------------------------
> 
> (Updated March 20, 2018, 9:47 p.m.)
> 
> 
> Review request for mesos, Chun-Hung Hsiao and Greg Mann.
> 
> 
> Bugs: MESOS-8617 and MESOS-8624
>     https://issues.apache.org/jira/browse/MESOS-8617
>     https://issues.apache.org/jira/browse/MESOS-8624
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> This helps to encapsulate a task launch into a single
> future which will come in handy when enforcing the task
> launch order.
> 
> Affected tests are also updated.
> 
> 
> Diffs
> -----
> 
>   src/slave/slave.hpp 37f0361251524e63d02d251e8a03901812b8affb 
>   src/slave/slave.cpp 915d4f1bbd642ca5d2587cfdd846262cb7e08788 
>   src/tests/mock_slave.hpp 8ec2b65a75f774d970ef6507df8b0c02fda5f2fd 
>   src/tests/mock_slave.cpp 8ec55b6aa84dd8b1c3743cbcdc246311585d0f63 
>   src/tests/slave_tests.cpp f76500ebdb67f131a57a3b5aaae8c952d019e354 
> 
> 
> Diff: https://reviews.apache.org/r/66143/diff/3/
> 
> 
> Testing
> -------
> 
> make check
> 
> 
> Thanks,
> 
> Meng Zhu
> 
>


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