> On 2012-03-23 05:17:01, Vinod Kone wrote:
> > hadoop/hadoop-0.20.205.0.patch, line 1354
> > <https://reviews.apache.org/r/4448/diff/2/?file=94645#file94645line1354>
> >
> >     kill ws here and everywhere else

Done (throughout the entire file!).


> On 2012-03-23 05:17:01, Vinod Kone wrote:
> > include/mesos/mesos.proto, line 91
> > <https://reviews.apache.org/r/4448/diff/2/?file=94646#file94646line91>
> >
> >     why are we removing the default executor? is there a jira i can look at?

We're removing the default executor to keep things simple. Now every TaskInfo 
requires either a CommandInfo or an ExecutorInfo.


> On 2012-03-23 05:17:01, Vinod Kone wrote:
> > src/common/type_utils.hpp, line 213
> > <https://reviews.apache.org/r/4448/diff/2/?file=94648#file94648line213>
> >
> >     why are you only wrapping left.resources() but not right? because 
> > Resources class can compare directly with protobuf? probably its easier to 
> > grok if you just wrap them both?

Agreed, I just have to line wrap now! ;)


> On 2012-03-23 05:17:01, Vinod Kone wrote:
> > src/master/master.cpp, line 1318
> > <https://reviews.apache.org/r/4448/diff/2/?file=94659#file94659line1318>
> >
> >     it would be cleaner to just extract task.executor() into a variable

There are lots of places where we don't have this style, so I'm not going to do 
it here for now.


> On 2012-03-23 05:17:01, Vinod Kone wrote:
> > src/master/master.cpp, line 1345
> > <https://reviews.apache.org/r/4448/diff/2/?file=94659#file94659line1345>
> >
> >     s/task's/tasks

Done, thanks!


> On 2012-03-23 05:17:01, Vinod Kone wrote:
> > src/master/master.cpp, line 1361
> > <https://reviews.apache.org/r/4448/diff/2/?file=94659#file94659line1361>
> >
> >     when does this happen? a different slave has an executor with the same 
> > id?

This was actually to catch for other tasks being processed simultaneously 
(i.e., in the same collection from SchedulerDriver::launchTasks) but it looks 
like we actually call Master::launchTask after each task rather than at the end 
so it doesn't look like we need this extra logic!


> On 2012-03-23 05:17:01, Vinod Kone wrote:
> > src/master/master.cpp, line 1706
> > <https://reviews.apache.org/r/4448/diff/2/?file=94659#file94659line1706>
> >
> >     This could be fixed when we merge with the new slave restart code

How nice!


> On 2012-03-23 05:17:01, Vinod Kone wrote:
> > src/slave/slave.hpp, line 297
> > <https://reviews.apache.org/r/4448/diff/2/?file=94667#file94667line297>
> >
> >     can it have both?

No, nor can it have neither. Changed to:

CHECK(task.has_executor() != task.has_command());


> On 2012-03-23 05:17:01, Vinod Kone wrote:
> > src/tests/resource_offers_tests.cpp, line 218
> > <https://reviews.apache.org/r/4448/diff/2/?file=94671#file94671line218>
> >
> >     wait..what happened to task1's updates?

Nothing, I just don't care about them, I only care about the updates that error!


> On 2012-03-23 05:17:01, Vinod Kone wrote:
> > src/tests/resource_offers_tests.cpp, line 221
> > <https://reviews.apache.org/r/4448/diff/2/?file=94671#file94671line221>
> >
> >     this should be a string constant.

While I agree that would be nice, none of the error strings are pulled out into 
constants so I'll wait for a full sweep to do that.


- Benjamin


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/4448/#review6269
-----------------------------------------------------------


On 2012-03-23 04:42:17, Benjamin Hindman wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/4448/
> -----------------------------------------------------------
> 
> (Updated 2012-03-23 04:42:17)
> 
> 
> Review request for mesos, Matei Zaharia, John Sirois, and Vinod Kone.
> 
> 
> Summary
> -------
> 
> See summary.
> 
> 
> Diffs
> -----
> 
>   hadoop/hadoop-0.20.205.0.patch 909a41e 
>   include/mesos/mesos.proto 478f561 
>   include/mesos/scheduler.hpp 34018f5 
>   src/common/type_utils.hpp efa1a84 
>   src/examples/java/TestExceptionFramework.java 6859c91 
>   src/examples/java/TestFramework.java fcfac40 
>   src/examples/java/TestMultipleExecutorsFramework.java 03a104b 
>   src/examples/long_lived_framework.cpp 9ca49e5 
>   src/examples/no_executor_framework.cpp f6aa601 
>   src/examples/python/test_framework.py 388f7c0 
>   src/examples/test_framework.cpp f97d080 
>   src/java/jni/org_apache_mesos_MesosSchedulerDriver.cpp 3857be3 
>   src/java/src/org/apache/mesos/MesosSchedulerDriver.java 590f9a2 
>   src/master/http.cpp 628112b 
>   src/master/master.cpp 6ae8aef 
>   src/python/native/mesos_executor_driver_impl.cpp 6e8fb9c 
>   src/python/native/mesos_scheduler_driver_impl.cpp 4b142ce 
>   src/python/native/module.hpp add021c 
>   src/python/native/module.cpp 59d74a4 
>   src/python/native/proxy_executor.hpp aabb6c4 
>   src/python/native/proxy_scheduler.hpp a348cbb 
>   src/sched/sched.cpp 43d9717 
>   src/slave/slave.hpp b7bc45a 
>   src/tests/exception_tests.cpp 11214c2 
>   src/tests/fault_tolerance_tests.cpp 130218d 
>   src/tests/master_tests.cpp 256945f 
>   src/tests/resource_offers_tests.cpp 303a933 
>   src/tests/utils.hpp 8c038ce 
>   src/webui/master/framework.tpl 9e384c1 
> 
> Diff: https://reviews.apache.org/r/4448/diff
> 
> 
> Testing
> -------
> 
> make check && make hadoop
> 
> 
> Thanks,
> 
> Benjamin
> 
>

Reply via email to