mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Guangya Liu <gyliu...@gmail.com>
Subject Re: Review Request 50841: Added GPU scheduling logic to docker containerizer process.
Date Sun, 09 Oct 2016 05:16:32 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/50841/#review151477
-----------------------------------------------------------




src/slave/containerizer/docker.cpp (lines 648 - 649)
<https://reviews.apache.org/r/50841/#comment219882>

    Put the blank space at the begining for the second sentence.
    ```
    return Failure("The 'allocateNvidiaGpus' function was called"
                   " without an 'NvidiaGpuAllocator' set");
    ```



src/slave/containerizer/docker.cpp (line 677)
<https://reviews.apache.org/r/50841/#comment219883>

    s/containers_[containerId]/containers_.at(containerId)



src/slave/containerizer/docker.cpp (lines 687 - 688)
<https://reviews.apache.org/r/50841/#comment219887>

    ```
    return Failure("The 'deallocateNvidiaGpus' function was called"
                   " without an 'NvidiaGpuAllocator' set");
    ```



src/slave/containerizer/docker.cpp (line 691)
<https://reviews.apache.org/r/50841/#comment219900>

    s/containers_[containerId]/containers_.at(containerId)



src/slave/containerizer/docker.cpp (line 703)
<https://reviews.apache.org/r/50841/#comment219884>

    s/containers_[containerId]/containers_.at(containerId)



src/slave/containerizer/docker.cpp (lines 1339 - 1358)
<https://reviews.apache.org/r/50841/#comment219892>

    What about moving this right before #1378?



src/slave/containerizer/docker.cpp (line 1350)
<https://reviews.apache.org/r/50841/#comment219885>

    s/scala/scalar



src/slave/containerizer/docker.cpp (lines 1554 - 1564)
<https://reviews.apache.org/r/50841/#comment220597>

    I think we do not need the update here as `destroy` will always be called even if the
container exit normally, relasing the GPU in `___destroy` is good enough.



src/slave/containerizer/docker.cpp (lines 2105 - 2109)
<https://reviews.apache.org/r/50841/#comment219895>

    Not yours, but I think that the comments can be removed as we have the logic of killing
those containers forcely in #2118 and #2123. You can fix this in a separate patch.



src/slave/containerizer/docker.cpp (lines 2162 - 2182)
<https://reviews.apache.org/r/50841/#comment220601>

    What about moving this to a new function named as `____destroy` and use continuation for
`deallocateNvidiaGpus` in #2159?


- Guangya Liu


On 九月 22, 2016, 6:16 a.m., Yubo Li wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/50841/
> -----------------------------------------------------------
> 
> (Updated 九月 22, 2016, 6:16 a.m.)
> 
> 
> Review request for mesos, Benjamin Mahler, Guangya Liu, Kevin Klues, and Rajat Phull.
> 
> 
> Bugs: MESOS-5795
>     https://issues.apache.org/jira/browse/MESOS-5795
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Added control logic to allocate/deallocate GPUs to GPU-related task
> when the task is started/terminated.
> 
> 
> Diffs
> -----
> 
>   src/slave/containerizer/docker.hpp 8da63101f951892e673612134770fc155d86112d 
>   src/slave/containerizer/docker.cpp 1d27761fcb3f310cf954d45ed41f4c89ecbd5982 
> 
> Diff: https://reviews.apache.org/r/50841/diff/
> 
> 
> Testing
> -------
> 
> make check
> 
> 
> Thanks,
> 
> Yubo Li
> 
>


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