[
https://issues.apache.org/jira/browse/YARN-7654?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16401192#comment-16401192
]
Eric Yang commented on YARN-7654:
---------------------------------
[~jlowe] Thank you for the review. This patch requires YARN-7221 to apply. As
you can see the struggle with flipping code for execv, this is the reason that
this patch takes a long time to develop. If we want to separate execv call
from getting a working version, then patch 001 would be reasonable to commit.
However, my current workspace is deeply committed to get execv working and I
don't want to get another CVE to add to my list of Hadoop security problems.
Therefore, I am likely to stay on course to finish execv version. I also found
that we only did docker inspect once to get pid. There is a fairly high chance
that docker inspect isn't ready and no pid would be produced. I have refactor
the code to loop through docker inspect n times before giving up. This
property needs to be configurable or detect the first child process has exited.
Docker image download can take sometime, and the inspect command might need a
lot of retires before giving up. Shane provided some good technique to check
proc in existing code. I should be able to reuse it for detection part.
{quote}
Why does the code not check for an execv error and instead exits with a success
code? It should log the strerror of the execv failure and exit with a failure
code instead.
{quote}
Good point. This will be fixed in the next version. Thanks again.
> 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
> Attachments: YARN-7654.001.patch, YARN-7654.002.patch
>
>
> 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]
