[SQOOP-3102] There are two questions in the sqoop_server_classpath_set function of the sqoop.sh. The sqoop2 start failure when using the same user with hadoop.

• Type: Bug
• Status: Patch Available
• Priority: Minor
• Resolution: Unresolved
• Component/s: sqoop2-dist
• Labels: patch
• Assignee: Unassigned
• Reporter: peng.jianhua
• Created: 11/Jan/17 02:50
• Updated: 21/Mar/17 11:30

Description

In start-all.sh, we have already default $HADOOP_HDFS_HOME and $HADOOP_YARN_HOME as $HADOOP_HOME and it works well.

1. start hdfs daemons if hdfs is present
   if [ -f "${HADOOP_HDFS_HOME}"/sbin/start-dfs.sh ]; then
   "${HADOOP_HDFS_HOME}"/sbin/start-dfs.sh --config $HADOOP_CONF_DIR
   fi
2. start yarn daemons if yarn is present
   if [ -f "${HADOOP_YARN_HOME}"/sbin/start-yarn.sh ]; then
   "${HADOOP_YARN_HOME}"/sbin/start-yarn.sh --config $HADOOP_CONF_DIR
   fi

When we execute sqoop.sh, we need to use these two settings: $HADOOP_HDFS_HOME and $HADOOP_YARN_HOME. If we set them in our environment, the path should be ${HADOOP_HOME}/share/hadoop/hdfs and ${HADOOP_HOME}/share/hadoop/yarn, then it will cause start-all.sh faulure, if we set them as $HADOOP_HOME, sqoop2 start failure.
HADOOP_COMMON_HOME=${HADOOP_COMMON_HOME:-${HADOOP_HOME}/share/hadoop/common}
HADOOP_HDFS_HOME=${HADOOP_HDFS_HOME:-${HADOOP_HOME}/share/hadoop/hdfs}
HADOOP_MAPRED_HOME=${HADOOP_MAPRED_HOME:-${HADOOP_HOME}/share/hadoop/mapreduce}
HADOOP_YARN_HOME=${HADOOP_YARN_HOME:-${HADOOP_HOME}/share/hadoop/yarn}

IMO, we can do some improvement in sqoop.sh as my attached patch. Just remove the validate to check if HADOOP_COMMON_HOME and HADOOP_YARN_HOME is null.

Issue Links

• links to: https://reviews.apache.org/r/55401/