> On July 5, 2017, 4:48 p.m., Jie Yu wrote:
> > src/slave/containerizer/mesos/launch.cpp
> > Lines 799 (patched)
> > <https://reviews.apache.org/r/60280/diff/6/?file=1768561#file1768561line799>
> >
> >     This is still based on the assumption that `launchInfo` has 
> > `working_directory()` set. Can you do the following instead:
> >     ```
> >     Result<string> realpath = os::realpath(executable);
> >     if (!realpath.isSome()) {
> >       ...
> >     }
> >     ```
> 
> Jie Yu wrote:
>     I also realized that using realpath will make the search for PATH not 
> possible. This issue is indeed not trivial. The best way I can think of is to 
> test if executable is relative and see if it exists in the sandbox or not and 
> test if it is executable or not. Alternatively, we can add cwd to PATH, but 
> this will polute PATH i would rather avoid.
>     
>     Do you have other thoughts?

Is it possible to assume that any custom executor would always be in the 
sandbox? E.g. assume that a framework will always provide a URI to fetch the 
executor and have Mesos execute it in the usual way. If that's the case then 
would we need to search `PATH` at all?


- Aaron


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/60280/#review179650
-----------------------------------------------------------


On July 3, 2017, 6:35 p.m., Aaron Wood wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/60280/
> -----------------------------------------------------------
> 
> (Updated July 3, 2017, 6:35 p.m.)
> 
> 
> Review request for mesos, Jie Yu, James Peach, and Zhitao Li.
> 
> 
> Bugs: MESOS-7703
>     https://issues.apache.org/jira/browse/MESOS-7703
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> If a framework specifies use of its own executor and sets shell to false the 
> executor is never found. This provides the full path to the executor so that 
> the `execvp` or `execvpe` is successful.
> 
> 
> Diffs
> -----
> 
>   src/slave/containerizer/mesos/launch.cpp 162ca1c2e 
> 
> 
> Diff: https://reviews.apache.org/r/60280/diff/6/
> 
> 
> Testing
> -------
> 
> `cd build && cmake .. -DCMAKE_BUILD_TYPE=Release && make -j4`
> Also spun up a master and agent, connected and sent a task using the UCR 
> (both with and without the use of an OCI image) via our own framework, and 
> checked the sandbox to verify that things went accordingly.
> 
> 
> Thanks,
> 
> Aaron Wood
> 
>

Reply via email to