mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Gastón Kleiman <gas...@mesosphere.io>
Subject Re: Review Request 61575: Added a test for verifying signal escalation on the default executor.
Date Wed, 16 Aug 2017 00:38:49 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/61575/#review183019
-----------------------------------------------------------




src/tests/default_executor_tests.cpp
Lines 599 (patched)
<https://reviews.apache.org/r/61575/#comment259008>

    Please use `ASSERT_FALSE(offers->offers().empty())` instead.



src/tests/default_executor_tests.cpp
Lines 609 (patched)
<https://reviews.apache.org/r/61575/#comment259010>

    Remove the empty line.



src/tests/default_executor_tests.cpp
Lines 621 (patched)
<https://reviews.apache.org/r/61575/#comment259014>

    I think we should make the grace period a bit longer to be able to check if a `TASK_KILLING`
status update is sent.



src/tests/default_executor_tests.cpp
Lines 625-626 (patched)
<https://reviews.apache.org/r/61575/#comment259011>

    I think that it'd be more readable to remove this from here, and to use the following
instead of the `CALL` block that follows:
    
    ```
      v1::Offer::Operation launchGroup = v1::LAUNCH_GROUP(
          executorInfo,
          v1::createTaskGroupInfo({task1, task2}));
    
      mesos.send(v1::createCallAccept(frameworkId, offer, {launchGroup}));
    ```



src/tests/default_executor_tests.cpp
Lines 629-630 (patched)
<https://reviews.apache.org/r/61575/#comment259012>

    If you  use the following, you can remove the acknowledge block bellow:
    
    ```
      EXPECT_CALL(*scheduler, update(_, _))
        .WillOnce(DoAll(
            FutureArg<1>(&update1),
            v1::scheduler::SendAcknowledge(
                frameworkId,
                offer.agent_id())));
    ```



src/tests/default_executor_tests.cpp
Lines 675 (patched)
<https://reviews.apache.org/r/61575/#comment259013>

    s/signify/signal/



src/tests/default_executor_tests.cpp
Lines 721 (patched)
<https://reviews.apache.org/r/61575/#comment259015>

    This checks that the process finally got a `SIGKILL`, but we can't be sure if it got a
`SIGTERM` before that.
    
    I couldn't find a test that checks this with the other executors, but I think that we
should at least add a TODO.


- Gastón Kleiman


On Aug. 15, 2017, 10:11 p.m., Anand Mazumdar wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/61575/
> -----------------------------------------------------------
> 
> (Updated Aug. 15, 2017, 10:11 p.m.)
> 
> 
> Review request for mesos, Gastón Kleiman, Jie Yu, and Vinod Kone.
> 
> 
> Bugs: MESOS-7879
>     https://issues.apache.org/jira/browse/MESOS-7879
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> This also required a small modification to the kill test helper binary
> to write a marker file signifying that the signal handlers have been
> set up correctly.
> 
> 
> Diffs
> -----
> 
>   src/tests/default_executor_tests.cpp b9776314a8781963b92ba9ac297654f61a443bc8 
>   src/tests/kill_policy_test_helper.hpp 29651102ec46b477e6e797c6e6bdef5b10afa665 
>   src/tests/kill_policy_test_helper.cpp a1880595ff015475f1ba49437d49f7397da19422 
> 
> 
> Diff: https://reviews.apache.org/r/61575/diff/3/
> 
> 
> Testing
> -------
> 
> make check
> 
> 
> Thanks,
> 
> Anand Mazumdar
> 
>


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