[ 
https://issues.apache.org/jira/browse/YARN-7654?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16380908#comment-16380908
 ] 

Eric Yang commented on YARN-7654:
---------------------------------

[[email protected]]
{quote}
Will we eliminate setting the workdir in this model and leave that up to the 
image? This has been a complaint I've heard with the bring your own image model.
{quote}

We have two choices here.  

Choice 1: We can make bring your own image first class citizen.  User who would 
like to use file cache, and yarn localized data in workdir, they will have to 
define this via user bind-mount.  This approach will break compatibility with 
existing apps.

Choice 2: We can introduce a flag to skip bind-mount of workdir.  I am open to 
suggestions on how to describe the flag in a clean way to prevent breaking 
compatibility.

{quote}
What is the impact on bind mounts? Will this help eliminate any of the 
read-write mounts?
{quote}

The current code has good validations to control user defined read-write mounts 
and comply with security checks.  Therefore, all bind-mounts feature is same.

{quote}
Would it also be possible to support arguments passed into the container's 
entrypoint? For instance, the Dockerfile ENTRYPOINT is ["ping"], the CMD is  
["google.com"] and instead of pinging www.google.com, I want to "ping 
www.yahoo.com" - the docker run would be docker run ... -e K=V -e K=V 
[image-name] www.yahoo.com.
{quote}

Yes, this will be supported.

> Support ENTRY_POINT for docker container
> ----------------------------------------
>
>                 Key: YARN-7654
>                 URL: https://issues.apache.org/jira/browse/YARN-7654
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: yarn
>    Affects Versions: 3.1.0
>            Reporter: Eric Yang
>            Assignee: Eric Yang
>            Priority: Blocker
>
> Docker image may have ENTRY_POINT predefined, but this is not supported in 
> the current implementation.  It would be nice if we can detect existence of 
> {{launch_command}} and base on this variable launch docker container in 
> different ways:
> h3. Launch command exists
> {code}
> docker run [image]:[version]
> docker exec [container_id] [launch_command]
> {code}
> h3. Use ENTRY_POINT
> {code}
> docker run [image]:[version]
> {code}



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to