You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mesos.apache.org by Benjamin Hindman <be...@berkeley.edu> on 2014/07/15 00:30:29 UTC
Review Request 23463: Serialize Containerizer::launch,wait invocations.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/23463/
-----------------------------------------------------------
Review request for mesos, Ian Downes and Jie Yu.
Bugs: MESOS-1527
https://issues.apache.org/jira/browse/MESOS-1527
Repository: mesos-git
Description
-------
The current contract between slave and containerizer is that the slave expects the containerizer to "know" about a container and handle the 'wait' method gettign called after 'launch' has been called but before the 'launch' has completed. These are a weird semantics and require awkward containerizer implementations that require storing information about a potential container launch before it has been successful in case an in between 'wait' occurs. Instead, this commit changes the semantics so that we don't do a 'wait' until after a 'launch' has completed.
Diffs
-----
src/slave/slave.cpp e81abb2e6371d052151253172a4abde7169cb72f
Diff: https://reviews.apache.org/r/23463/diff/
Testing
-------
make check
Thanks,
Benjamin Hindman
Re: Review Request 23463: Serialize Containerizer::launch,wait invocations.
Posted by Jie Yu <yu...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/23463/#review47921
-----------------------------------------------------------
Ship it!
Ship It!
- Jie Yu
On July 14, 2014, 10:30 p.m., Benjamin Hindman wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/23463/
> -----------------------------------------------------------
>
> (Updated July 14, 2014, 10:30 p.m.)
>
>
> Review request for mesos, Ian Downes and Jie Yu.
>
>
> Bugs: MESOS-1527
> https://issues.apache.org/jira/browse/MESOS-1527
>
>
> Repository: mesos-git
>
>
> Description
> -------
>
> The current contract between slave and containerizer is that the slave expects the containerizer to "know" about a container and handle the 'wait' method gettign called after 'launch' has been called but before the 'launch' has completed. These are a weird semantics and require awkward containerizer implementations that require storing information about a potential container launch before it has been successful in case an in between 'wait' occurs. Instead, this commit changes the semantics so that we don't do a 'wait' until after a 'launch' has completed.
>
>
> Diffs
> -----
>
> src/slave/slave.cpp e81abb2e6371d052151253172a4abde7169cb72f
>
> Diff: https://reviews.apache.org/r/23463/diff/
>
>
> Testing
> -------
>
> make check
>
>
> Thanks,
>
> Benjamin Hindman
>
>
Re: Review Request 23463: Serialize Containerizer::launch,wait invocations.
Posted by Ian Downes <ia...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/23463/#review47926
-----------------------------------------------------------
Ship it!
Is there cleanup to be done in the containerizers to remove the awkwardness? Can TODOs be added if so.
- Ian Downes
On July 14, 2014, 3:30 p.m., Benjamin Hindman wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/23463/
> -----------------------------------------------------------
>
> (Updated July 14, 2014, 3:30 p.m.)
>
>
> Review request for mesos, Ian Downes and Jie Yu.
>
>
> Bugs: MESOS-1527
> https://issues.apache.org/jira/browse/MESOS-1527
>
>
> Repository: mesos-git
>
>
> Description
> -------
>
> The current contract between slave and containerizer is that the slave expects the containerizer to "know" about a container and handle the 'wait' method gettign called after 'launch' has been called but before the 'launch' has completed. These are a weird semantics and require awkward containerizer implementations that require storing information about a potential container launch before it has been successful in case an in between 'wait' occurs. Instead, this commit changes the semantics so that we don't do a 'wait' until after a 'launch' has completed.
>
>
> Diffs
> -----
>
> src/slave/slave.cpp e81abb2e6371d052151253172a4abde7169cb72f
>
> Diff: https://reviews.apache.org/r/23463/diff/
>
>
> Testing
> -------
>
> make check
>
>
> Thanks,
>
> Benjamin Hindman
>
>