[ https://issues.apache.org/jira/browse/YARN-3685?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14559464#comment-14559464 ]
Chris Nauroth commented on YARN-3685: ------------------------------------- bq. Which ones are these? I was thinking of stuff like {{yarn.application.classpath}}, where values are defined in terms of things like the {{HADOOP_YARN_HOME}} and {{HADOOP_COMMON_HOME}} environment variables, and those values might not match the file system layout at the client side. bq. Not clear this is true or not. Have to see the final solution/patch to realistically reason about this. Let's say hypothetically a change for this goes into 2.8.0. I was thinking that would make it impossible for a 2.7.0 client to work correctly with a 2.8.0 NodeManager, because that client wouldn't take care of classpath bundling and instead expect the NodeManager to do it. Brainstorming a bit, maybe we can figure out a way for a 2.8.0 NodeManager to detect if the client hasn't already taken care of classpath bundling, and if not, stick to the current logic. Backwards-compatibility logic like this would go into branch-2, but could be dropped from trunk. > NodeManager unnecessarily knows about classpath-jars due to Windows > limitations > ------------------------------------------------------------------------------- > > Key: YARN-3685 > URL: https://issues.apache.org/jira/browse/YARN-3685 > Project: Hadoop YARN > Issue Type: Sub-task > Components: nodemanager > Reporter: Vinod Kumar Vavilapalli > > Found this while looking at cleaning up ContainerExecutor via YARN-3648, > making it a sub-task. > YARN *should not* know about classpaths. Our original design modeled around > this. But when we added windows suppport, due to classpath issues, we ended > up breaking this abstraction via YARN-316. We should clean this up. -- This message was sent by Atlassian JIRA (v6.3.4#6332)