mesos-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Niklas Nielsen" <...@qni.dk>
Subject Re: Review Request 28069: Propagated slave shutdown grace period to Executor and CommandExecutor.
Date Thu, 20 Nov 2014 20:39:40 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/28069/#review62277
-----------------------------------------------------------


Looking good! Some comments below


src/exec/exec.cpp
<https://reviews.apache.org/r/28069/#comment104299>

    s/grace_shutdown/graceful_shutdown/g?



src/exec/exec.cpp
<https://reviews.apache.org/r/28069/#comment104300>

    s/in a sense/as/



src/exec/exec.cpp
<https://reviews.apache.org/r/28069/#comment104304>

    Still don't like the notion of 'adjusting'. You get _another_ kind of timeout from the
first value (which describes the global and inner-most timeout). 
    
    Can we change it?



src/launcher/executor.cpp
<https://reviews.apache.org/r/28069/#comment104305>

    Same here - graceful_shutdown.hpp?



src/launcher/executor.cpp
<https://reviews.apache.org/r/28069/#comment104307>

    We switch between escalationTimeout and shutdownTimeout. We should be consistent.



src/launcher/executor.cpp
<https://reviews.apache.org/r/28069/#comment104308>

    Which is a problem in case of pid reuse? Worth mentioning here?



src/launcher/executor.cpp
<https://reviews.apache.org/r/28069/#comment104309>

    Same comment as above wrt 'adjusting'



src/launcher/executor.cpp
<https://reviews.apache.org/r/28069/#comment104310>

    Can you remind me again why you needed to move this?



src/slave/containerizer/containerizer.cpp
<https://reviews.apache.org/r/28069/#comment104314>

    I would wrap as you do below for consistency



src/slave/flags.hpp
<https://reviews.apache.org/r/28069/#comment104439>

    I would drop that paragraph. That's an implementation detail :)



src/slave/slave.cpp
<https://reviews.apache.org/r/28069/#comment104312>

    s/grace/graceful/?



src/tests/slave_tests.cpp
<https://reviews.apache.org/r/28069/#comment104463>

    Please align like we do here: https://github.com/apache/mesos/blob/master/src/tests/slave_tests.cpp#L803
    
    Here and below.



src/tests/slave_tests.cpp
<https://reviews.apache.org/r/28069/#comment104316>

    This is expected executor info behavior and not strictly related to graceful shutdown.
I would drop the comment.


- Niklas Nielsen


On Nov. 17, 2014, 4:14 a.m., Alexander Rukletsov wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/28069/
> -----------------------------------------------------------
> 
> (Updated Nov. 17, 2014, 4:14 a.m.)
> 
> 
> Review request for mesos, Ben Mahler, Niklas Nielsen, and Till Toenshoff.
> 
> 
> Bugs: MESOS-1571
>     https://issues.apache.org/jira/browse/MESOS-1571
> 
> 
> Repository: mesos-git
> 
> 
> Description
> -------
> 
> The configurable slave's executor_shutdown_grace_period flag is propagated to Executor
and CommandExecutor via CommandInfo and an environment variable. Shutdown timeout in ExecutorProcess
and signal escalation timeout in CommandExecutor are now dependent on this value.
> 
> 
> Diffs
> -----
> 
>   src/exec/exec.cpp e15f834 
>   src/launcher/executor.cpp 6175bf5 
>   src/slave/constants.hpp fd1c1ab 
>   src/slave/constants.cpp 2a99b11 
>   src/slave/containerizer/containerizer.cpp f234835 
>   src/slave/flags.hpp fee79e0 
>   src/slave/slave.cpp 06b2e18 
>   src/tests/gc_tests.cpp 8618ae1 
>   src/tests/slave_tests.cpp 18ff8fe 
> 
> Diff: https://reviews.apache.org/r/28069/diff/
> 
> 
> Testing
> -------
> 
> make check (Mac OS 10.9.4, Ubuntu 14.04)
> 
> 
> Thanks,
> 
> Alexander Rukletsov
> 
>


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