[ https://issues.apache.org/jira/browse/MAPREDUCE-6454?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14705905#comment-14705905 ]
Allen Wittenauer commented on MAPREDUCE-6454: --------------------------------------------- bq. If the admin sets it in the daemons, and also configures yarn to white-list those envs, we are inheriting them into the task. Again similar to CLASSPATH. If the original problem is that the bin commands don't pick up this up, this fix makes the assumption that HADOOP_CLASSPATH allows inheritance from the parent shell. The default examples in trunk specifically don't do this to prevent the double settings problem that plagues prior releases. In fact, since HADOOP_CLASSPATH is really intended for users to use, why are we overloading it? For trunk at least, it would probably be better to have a different var that is handled via mapreduce's shellprofile.d bit. > MapReduce doesn't set the HADOOP_CLASSPATH for jar lib in distributed cache. > ---------------------------------------------------------------------------- > > Key: MAPREDUCE-6454 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-6454 > Project: Hadoop Map/Reduce > Issue Type: Bug > Reporter: Junping Du > Assignee: Junping Du > Priority: Critical > Attachments: MAPREDUCE-6454-v2.1.patch, MAPREDUCE-6454-v2.patch, > MAPREDUCE-6454-v3.1.patch, MAPREDUCE-6454-v3.patch, MAPREDUCE-6454.patch > > > We already set lib jars on distributed-cache to CLASSPATH. However, in some > corner cases (like: MR local mode, Hive Map side local join, etc.), we need > these jars on HADOOP_CLASSPATH so hadoop scripts can take it in launching > runjar process. -- This message was sent by Atlassian JIRA (v6.3.4#6332)