hive 3.1.2 在 hive-env.sh 配置 beeline 的选项不生效的解决方法

hive-env.sh 可以配置相关的 HADOOP 选项，里面有注掉的 cli 的示例如下。

# if [ "$SERVICE" = "cli" ]; then
#   if [ -z "$DEBUG" ]; then
#     export HADOOP_OPTS="$HADOOP_OPTS -XX:NewRatio=12 -Xms10m -XX:MaxHeapFreeRatio=40 -XX:MinHeapFreeRatio=15 -XX:+UseParNewGC -XX:-UseGCOverheadLimit"
#   else
#     export HADOOP_OPTS="$HADOOP_OPTS -XX:NewRatio=12 -Xms10m -XX:MaxHeapFreeRatio=40 -XX:MinHeapFreeRatio=15 -XX:-UseGCOverheadLimit"
#   fi
# fi

Hive server2 的配置可以生效

下面的配置说明，如果 HIVE_SERVER2_DEBUG 环境变量设置为 TRUE, 则启动 debug 模式，监听一个端口。

if [ "$SERVICE" = "hiveserver2" ]; then
   if [ "${HIVE_SERVER2_DEBUG}" = "TRUE" ] ; then
        HADOOP_OPTS="$HADOOP_OPTS -Xdebug -Xrunjdwp:transport=dt_socket,server=y,address=7902,suspend=y"
   fi
fi

beeline 的配置不生效生效

if [ "$SERVICE" = "beeline" ]; then
   if [ "${BEELINE_DEBUG}" = "TRUE" ] ; then
   	    echo "SERVICE" = "beeline"
        HADOOP_OPTS="$HADOOP_OPTS -Xdebug -Xrunjdwp:transport=dt_socket,server=y,address=7903,suspend=y"
        echo $HADOOP_OPTS
   fi
fi

分析

${HIVE_HOME}/bin/ext/hiveserver2.sh 里的方法有 export HADOOP_OPTS

hiveserver2() {
  >&2 echo "$(timestamp): Starting HiveServer2"
  CLASS=org.apache.hive.service.server.HiveServer2
  if $cygwin; then
    HIVE_LIB=`cygpath -w "$HIVE_LIB"`
  fi
  JAR=${HIVE_LIB}/hive-service-[0-9].*.jar

  export HADOOP_CLIENT_OPTS=" -Dproc_hiveserver2 $HADOOP_CLIENT_OPTS "
  export HADOOP_OPTS="$HIVESERVER2_HADOOP_OPTS $HADOOP_OPTS"
  exec $HADOOP jar $JAR $CLASS $HIVE_OPTS "$@"
}

${HIVE_HOME}/bin/ext/beeline.sh 里的方法没有 export HADOOP_OPTS

beeline () {
  CLASS=org.apache.hive.beeline.BeeLine;

  # include only the beeline client jar and its dependencies
  beelineJarPath=`ls ${HIVE_LIB}/hive-beeline-*.jar`
  superCsvJarPath=`ls ${HIVE_LIB}/super-csv-*.jar`
  jlineJarPath=`ls ${HIVE_LIB}/jline-*.jar`
  hadoopClasspath=""
  if [[ -n "${HADOOP_CLASSPATH}" ]]
  then
    hadoopClasspath="${HADOOP_CLASSPATH}:"
  fi
  export HADOOP_CLASSPATH="${hadoopClasspath}${HIVE_CONF_DIR}:${beelineJarPath}:${superCsvJarPath}:${jlineJarPath}"
  export HADOOP_CLIENT_OPTS="$HADOOP_CLIENT_OPTS -Dlog4j.configurationFile=beeline-log4j2.properties "
  
  exec $HADOOP jar ${beelineJarPath} $CLASS $HIVE_OPTS "$@"
}

解决方案

${HIVE_HOME}/bin/ext/beeline.sh 里 export HADOOP_OPTS

beeline () {
  CLASS=org.apache.hive.beeline.BeeLine;

  # include only the beeline client jar and its dependencies
  beelineJarPath=`ls ${HIVE_LIB}/hive-beeline-*.jar`
  superCsvJarPath=`ls ${HIVE_LIB}/super-csv-*.jar`
  jlineJarPath=`ls ${HIVE_LIB}/jline-*.jar`
  hadoopClasspath=""
  if [[ -n "${HADOOP_CLASSPATH}" ]]
  then
    hadoopClasspath="${HADOOP_CLASSPATH}:"
  fi
  export HADOOP_CLASSPATH="${hadoopClasspath}${HIVE_CONF_DIR}:${beelineJarPath}:${superCsvJarPath}:${jlineJarPath}"
  export HADOOP_CLIENT_OPTS="$HADOOP_CLIENT_OPTS -Dlog4j.configurationFile=beeline-log4j2.properties "
  export HADOOP_OPTS="$HADOOP_OPTS"

  exec $HADOOP jar ${beelineJarPath} $CLASS $HIVE_OPTS "$@"
}