aurora-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Joshua Cohen <jco...@apache.org>
Subject Re: Review Request 44685: Add scheduler support for running tasks using the mesos Docker containerizer.
Date Fri, 11 Mar 2016 16:13:35 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/44685/#review123156
-----------------------------------------------------------




src/main/java/org/apache/aurora/scheduler/app/AppModule.java (lines 91 - 93)
<https://reviews.apache.org/r/44685/#comment185331>

    I feel like this is confusingly named. The scheduler already accepts tasks that use the
Docker containerizer. What this patch actually does is allow for tasks to run in the Docker
containerizer without the need for an executor to coordinate their execution, right?
    
    Is the term "Docker Containerizer" somehow overloaded and leading to my confusion?
    
    How would you feel about renaming this to something like `require_docker_use_executor`
(defaulting to True)?
    
    Open to other ideas as well, but figured I'd throw one out to start.



src/main/java/org/apache/aurora/scheduler/configuration/ConfigurationManager.java (lines 281
- 282)
<https://reviews.apache.org/r/44685/#comment185332>

    This read a little bit better to me the previous way it was written. Maybe something like:
`This scheduler is not configured to allow the container type: " + containerType.get().toString()`?



src/main/java/org/apache/aurora/scheduler/mesos/MesosTaskFactory.java (line 164)
<https://reviews.apache.org/r/44685/#comment185333>

    Instead of building the `TaskInfo` above only to clear it all out in this case, how about
just extracting the logic to build the TaskInfo for tasks w/ an executor, and invoking that
in the branches where it's used?


- Joshua Cohen


On March 11, 2016, 2:24 a.m., Bill Farner wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/44685/
> -----------------------------------------------------------
> 
> (Updated March 11, 2016, 2:24 a.m.)
> 
> 
> Review request for Aurora, Joshua Cohen and John Sirois.
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> This is currently labeled as experimental.
> 
> Only the most basic wiring is added here, and assumes that the provided image
> includes an ENTRYPOINT.  Unlike Docker support via the thermos executor, this
> approach allows containers with an entrypoint, and does not impose environment
> requirements on the image (e.g. python interpreter, libmesos dependencies).
> 
> Note that when using this, other familiar Aurora facilities that relate to the
> thermos executor will not work.  For example, browsing task logs is not
> supported.
> 
> Support for exercising this from the client will come shortly.
> 
> 
> Diffs
> -----
> 
>   NEWS da3e4cea8ca688b6b7c5bafae67133df065d9255 
>   src/jmh/java/org/apache/aurora/benchmark/ThriftApiBenchmarks.java 60746383fccb107ca27925a91aa1803e2cf0fd85

>   src/main/java/org/apache/aurora/scheduler/app/AppModule.java a0d2a717534bbb2e85a556721cc53c1e4b743461

>   src/main/java/org/apache/aurora/scheduler/base/TaskTestUtil.java 1de6966565d2fbd9abd220ad8162b624b109959a

>   src/main/java/org/apache/aurora/scheduler/configuration/ConfigurationManager.java e700fa3550312bfa9c8a3adb25d135f6f500c4b5

>   src/main/java/org/apache/aurora/scheduler/mesos/MesosTaskFactory.java a34af4d2fb3863ab8197bcdce942c513d629621b

>   src/test/java/org/apache/aurora/scheduler/configuration/ConfigurationManagerTest.java
11062e3a097e490c61bfd4dc84990903275521a3 
>   src/test/java/org/apache/aurora/scheduler/mesos/MesosTaskFactoryImplTest.java 3db531b52fb2bd94b4b5ce62e6554b5a85ed3ea8

> 
> Diff: https://reviews.apache.org/r/44685/diff/
> 
> 
> Testing
> -------
> 
> Via additional hacking, i successfully ran the stock [hello world](https://hub.docker.com/_/hello-world/)
image.  Within the sandbox, i observed the expected output in the `stdout` file.  Status updates
for the task exiting worked as expected.
> 
> 
> Thanks,
> 
> Bill Farner
> 
>


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