> On March 15, 2016, 8:11 p.m., Ben Mahler wrote:
> > src/executor/executor.cpp, line 703
> > <https://reviews.apache.org/r/44655/diff/4/?file=1299805#file1299805line703>
> >
> >     I'm thinking now that we should trust the value in the environment if 
> > it's set, rather than overwriting, since the agent could set a smaller one. 
> > This would only happen if, rather than rejecting too large of timeouts, we 
> > bounded them down implicitly. There's also the case of the agent deciding 
> > it has less time, but that case would likely come at the time of the 
> > SHUTDOWN Event.
> >     
> >     Perhaps this logic should be: set if not found in environment. And we 
> > should document that the agent will use the ExecutorInfo to set the 
> > environment value.

Per offline discussion: the env var has the actual value. If the env var is not 
set, the agent is pre 0.28 and it doesn't support custom shutdown grace periods 
anyway, so even if `ExecutorInfo` specifies the custom grace period, we can 
ignore it.


- Alexander


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/44655/#review123731
-----------------------------------------------------------


On March 15, 2016, 3:49 p.m., Alexander Rukletsov wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/44655/
> -----------------------------------------------------------
> 
> (Updated March 15, 2016, 3:49 p.m.)
> 
> 
> Review request for mesos, Ben Mahler and Gilbert Song.
> 
> 
> Bugs: MESOS-4949
>     https://issues.apache.org/jira/browse/MESOS-4949
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> If `ExecutorInfo.shutdown_grace_period` is set, the executor
> driver uses it, otherwise it falls back to the environment
> variable `MESOS_EXECUTOR_SHUTDOWN_GRACE_PERIOD`.
> 
> 
> Diffs
> -----
> 
>   docs/configuration.md 739d4ff9aeeb1ba70ce11033168d63d37b6ef56b 
>   include/mesos/executor/executor.proto 
> ae211194a44e0bf2fadc79e833881e45ea3eb2c2 
>   include/mesos/mesos.proto deb9c0910a27afd67276f54b3f666a878212727b 
>   include/mesos/v1/executor/executor.proto 
> 36a2b3f9bc3aaa524f655b9e686a6d33512e6aaa 
>   include/mesos/v1/mesos.proto a981e750c24cfc48177bbc9ca56f0c3ecfae1a1b 
>   src/exec/exec.cpp 741786132f3a8cc43f5b9ced262429038832a946 
>   src/executor/executor.cpp 87db4e02cbaa778aab0173741bfe066fdee9a48d 
>   src/slave/flags.cpp 4d10818105627738e258116647ccada374e3d7b9 
>   src/slave/slave.cpp 840534ff0687e82ed063c386e36bbabada230697 
> 
> Diff: https://reviews.apache.org/r/44655/diff/
> 
> 
> Testing
> -------
> 
> The complete chain was tested. See https://reviews.apache.org/r/44662/.
> 
> 
> Thanks,
> 
> Alexander Rukletsov
> 
>

Reply via email to