[ 
https://issues.apache.org/jira/browse/YARN-10018?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16995001#comment-16995001
 ] 

Eric Badger commented on YARN-10018:
------------------------------------

bq. I have one suggestion. I believe it is better to use ERROR_FORKING_PROCESS 
everywhere, to support quick debugging. However, this means that you need to 
check if the calling functions handle this properly.

I think the important thing here is that we have 2 pieces of information when 
the container-executor fails. We need a pointer to what piece of code failed 
(e.g. docker exec, traffic control, etc.), and information on what kind of 
failure we encountered. Patch 001 does this via a log message for the forking 
error, and then the piece of code in the return code. Obviously, the other way 
is to log the piece of code via the ERRORFILE, and then specify the forking 
error via the return code. I'm not too awfully picky between the two approaches 
so long as both pieces of information are available when the container-executor 
fails. 

> container-executor: possible -1 return value of fork() is not always checked
> ----------------------------------------------------------------------------
>
>                 Key: YARN-10018
>                 URL: https://issues.apache.org/jira/browse/YARN-10018
>             Project: Hadoop YARN
>          Issue Type: Bug
>          Components: nodemanager
>            Reporter: Peter Bacsko
>            Assignee: Peter Bacsko
>            Priority: Major
>         Attachments: YARN-10018-001.patch, YARN-10018-001.patch
>
>
> There are some places in the container-executor native, where the {{fork()}} 
> call is not handled properly. This operation can fail with -1, but sometimes 
> the necessary if branch is missing to validate that it's been successful.
> Also, at one location, the return value is defined as an {{int}}, not 
> {{pid_t}}. It's better to handle this transparently and change it.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

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

Reply via email to