Hadoop的hadoop-config.sh脚本详解

先简要说明下我的读该脚本的环境，集群是CDH5.3，在研究命令行中hadoop -jar ***.jar命令的相关脚本时，在hadoop文件中追到了hadoop-config.sh这个文件，所以下边的注解添加了我当前情况的处理。

this="${BASH_SOURCE-$0}" #显示本脚本文件全路径
common_bin=$(cd -P -- "$(dirname -- "$this")" && pwd -P) #获取本脚本所在目录路径
script="$(basename -- "$this")" #获取脚本名称
this="$common_bin/$script"
[ -f "$common_bin/hadoop-layout.sh" ] && . "$common_bin/hadoop-layout.sh"
#在hadoop-config.sh中会执行hadoop-layout.sh（/opt/cloudera/parcels/CDH-5.3.0-1.cdh5.3.0.p0.30/lib/hadoop/libexec/hadoop-layout.sh）初始化些基本设置：

#设置hadoop common、 hdfs、yarn、mapred目录的一下变量

HADOOP_COMMON_DIR=${HADOOP_COMMON_DIR:-"share/hadoop/common"}

HADOOP_COMMON_LIB_JARS_DIR=${HADOOP_COMMON_LIB_JARS_DIR:-"share/hadoop/common/lib"}
HADOOP_COMMON_LIB_NATIVE_DIR=${HADOOP_COMMON_LIB_NATIVE_DIR:-"lib/native"}
HDFS_DIR=${HDFS_DIR:-"share/hadoop/hdfs"}
HDFS_LIB_JARS_DIR=${HDFS_LIB_JARS_DIR:-"share/hadoop/hdfs/lib"}
YARN_DIR=${YARN_DIR:-"share/hadoop/yarn"}
YARN_LIB_JARS_DIR=${YARN_LIB_JARS_DIR:-"share/hadoop/yarn/lib"}
MAPRED_DIR=${MAPRED_DIR:-"share/hadoop/mapreduce"}
MAPRED_LIB_JARS_DIR=${MAPRED_LIB_JARS_DIR:-"share/hadoop/mapreduce/lib"}
！！设置的值：
HADOOP_COMMON_DIR:./
HADOOP_COMMON_LIB_JARS_DIR:lib
HDFS_DIR:./
HDFS_LIB_JARS_DIR:lib
YARN_DIR:./
YARN_LIB_JARS_DIR:lib
MAPRED_DIR:./
MAPRED_LIB_JARS_DIR:lib

 #hadoop-layout.sh执行结束，继续执行hadoop-config.sh

# the root of the Hadoop installation，设定hadoop目录

# See HADOOP-6255 for directory structure layout
HADOOP_DEFAULT_PREFIX=$(cd -P -- "$common_bin"/.. && pwd -P)
HADOOP_PREFIX=${HADOOP_PREFIX:-$HADOOP_DEFAULT_PREFIX}
！！设置的值：
HADOOP_DEFAULT_PREFIX:/opt/cloudera/parcels/CDH-5.3.0-1.cdh5.3.0.p0.30/lib/hadoop
HADOOP_PREFIX:/opt/cloudera/parcels/CDH-5.3.0-1.cdh5.3.0.p0.30/lib/hadoop

#check to see if the conf dir is given as an optional argument #校验是否设置配置文件目录，调用该config.sh,没有参数，直接过
if [ $# -gt 1 ]
then
    if [ "--config" = "$1" ]
          then
              shift
              confdir=$1
              if [ ! -d "$confdir" ]; then
                echo "Error: Cannot find configuration directory: $confdir"
                exit 1
             fi
              shift
              HADOOP_CONF_DIR=$confdir
    fi
fi

# Allow alternate conf dir location.（设定配置文件目录，环境中没有${HADOOP_PREFIX}/conf/hadoop-env.sh，所以使用etc/hadoop/的目录，环境中hadoop -> /etc/hadoop/conf有这么个链接，所以读的是/etc/hadoop/conf/下的配置文件）
if [ -e "${HADOOP_PREFIX}/conf/hadoop-env.sh" ]; then
  DEFAULT_CONF_DIR="conf"
else
  DEFAULT_CONF_DIR="etc/hadoop"
fi
export HADOOP_CONF_DIR="${HADOOP_CONF_DIR:-$HADOOP_PREFIX/$DEFAULT_CONF_DIR}"

# User can specify hostnames or a file where the hostnames are (not both)（用户可指定slave的hostnames，不能两个变量都指定值，本环境中两个变量都是空的）
if [[ ( "$HADOOP_SLAVES" != '' ) && ( "$HADOOP_SLAVE_NAMES" != '' ) ]] ; then
  echo \
    "Error: Please specify one variable HADOOP_SLAVES or " \
    "HADOOP_SLAVE_NAME and not both."
  exit 1
fi

# Process command line options that specify hosts or file with host（可在脚本参数中指定hostnames。没有参数，所以没有指定HADOOP_SLAVES和HADOOP_SLAVE_NAMES）
# list
if [ $# -gt 1 ]
then
    if [ "--hosts" = "$1" ]
    then
        shift
        export HADOOP_SLAVES="${HADOOP_CONF_DIR}/$1"
        shift
    elif [ "--hostnames" = "$1" ]
    then
        shift
        export HADOOP_SLAVE_NAMES=$1
        shift
    fi
fi

# User can specify hostnames or a file where the hostnames are (not both)
# (same check as above but now we know it's command line options that cause
# the problem)（校验参数中是否指定该两个参数，如果都指定也是错误的）
if [[ ( "$HADOOP_SLAVES" != '' ) && ( "$HADOOP_SLAVE_NAMES" != '' ) ]] ; then
  echo \
    "Error: Please specify one of --hosts or --hostnames options and not both."
  exit 1
fi

#执行hadoop-env.sh(/etc/hadoop/conf/hadoop-env.sh).设置些环境变量
if [ -f "${HADOOP_CONF_DIR}/hadoop-env.sh" ]; then
  . "${HADOOP_CONF_DIR}/hadoop-env.sh"
fi

# check if net.ipv6.bindv6only is set to 1 （检验net.ipv6.bindv6only是否为1，这个参数的作用没有深究）
bindv6only=$(/sbin/sysctl -n net.ipv6.bindv6only 2> /dev/null)
if [ -n "$bindv6only" ] && [ "$bindv6only" -eq "1" ] && [ "$HADOOP_ALLOW_IPV6" != "yes" ]
then
  echo "Error: \"net.ipv6.bindv6only\" is set to 1 - Java networking could be broken"
  echo "For more info: http://wiki.apache.org/hadoop/HadoopIPv6"
  exit 1
fi

# Newer versions of glibc use an arena memory allocator that causes virtual
# memory usage to explode. This interacts badly with the many threads that
# we use in Hadoop. Tune the variable down to prevent vmem explosion.（因为版本问题，设置参数防止内存崩溃）
export MALLOC_ARENA_MAX=${MALLOC_ARENA_MAX:-4}

# Attempt to set JAVA_HOME if it is not set（如果没有JAVA_HOME，设定JAVA_HOME，JAVA_HOME这个环境变量很重要啊，，，）
if [[ -z $JAVA_HOME ]]; then
  # On OSX use java_home (or /Library for older versions)
  if [ "Darwin" == "$(uname -s)" ]; then
    if [ -x /usr/libexec/java_home ]; then
      export JAVA_HOME=($(/usr/libexec/java_home))
    else
      export JAVA_HOME=(/Library/Java/Home)
    fi
  fi
  # Bail if we did not detect it
  if [[ -z $JAVA_HOME ]]; then
    echo "Error: JAVA_HOME is not set and could not be found." 1>&2
    exit 1
  fi
fi

JAVA=$JAVA_HOME/bin/java
# some Java parameters  （设置java堆栈最大值，如果hadoop程序中需要很大的内存就可以修改这个参数）
JAVA_HEAP_MAX=-Xmx1000m 

# check envvars which might override default args（如果系统设置了HADOOP_HEAPSIZE，就覆盖JAVA_HEAP_MAX）
if [ "$HADOOP_HEAPSIZE" != "" ]; then
  #echo "run with heapsize $HADOOP_HEAPSIZE"
  JAVA_HEAP_MAX="-Xmx""$HADOOP_HEAPSIZE""m"
  #echo $JAVA_HEAP_MAX
fi

# CLASSPATH initially contains $HADOOP_CONF_DIR （设置CLASSPATH）
CLASSPATH="${HADOOP_CONF_DIR}"


#设定HADOOP_COMMON_HOM，上边是设置过了的
if [ "$HADOOP_COMMON_HOME" = "" ]; then
  if [ -d "${HADOOP_PREFIX}/$HADOOP_COMMON_DIR" ]; then
    export HADOOP_COMMON_HOME=$HADOOP_PREFIX
  fi
fi

# for releases, add core hadoop jar & webapps to CLASSPATH（将webapps，lib中的东西都加载到CLASSPATH）
if [ -d "$HADOOP_COMMON_HOME/$HADOOP_COMMON_DIR/webapps" ]; then
  CLASSPATH=${CLASSPATH}:$HADOOP_COMMON_HOME/$HADOOP_COMMON_DIR
fi
if [ -d "$HADOOP_COMMON_HOME/$HADOOP_COMMON_LIB_JARS_DIR" ]; then
  CLASSPATH=${CLASSPATH}:$HADOOP_COMMON_HOME/$HADOOP_COMMON_LIB_JARS_DIR'/*'
fi
CLASSPATH=${CLASSPATH}:$HADOOP_COMMON_HOME/$HADOOP_COMMON_DIR'/*'

# default log directory & file (设定日志文件目录和文件名称)
if [ "$HADOOP_LOG_DIR" = "" ]; then
  HADOOP_LOG_DIR="$HADOOP_PREFIX/logs"
fi
if [ "$HADOOP_LOGFILE" = "" ]; then
  HADOOP_LOGFILE='hadoop.log'
fi

# default policy file for service-level authorization（指定hadoop-policy.xml文件，该文件用于hadoop服务级别的授权认证）
if [ "$HADOOP_POLICYFILE" = "" ]; then
  HADOOP_POLICYFILE="hadoop-policy.xml"
fi

# setup 'java.library.path' for native-hadoop code if necessary （设定java.library.path，将native-hadoop的一些东西加进去）
if [ -d "${HADOOP_PREFIX}/build/native" -o -d "${HADOOP_PREFIX}/$HADOOP_COMMON_LIB_NATIVE_DIR" ]; then
  if [ -d "${HADOOP_PREFIX}/$HADOOP_COMMON_LIB_NATIVE_DIR" ]; then
    if [ "x$JAVA_LIBRARY_PATH" != "x" ]; then
      JAVA_LIBRARY_PATH=${JAVA_LIBRARY_PATH}:${HADOOP_PREFIX}/$HADOOP_COMMON_LIB_NATIVE_DIR
    else
      JAVA_LIBRARY_PATH=${HADOOP_PREFIX}/$HADOOP_COMMON_LIB_NATIVE_DIR
    fi
  fi
fi

# setup a default TOOL_PATH（设置hadoop的工具包路径）
TOOL_PATH="${TOOL_PATH:-$HADOOP_PREFIX/share/hadoop/tools/lib/*}"


#设定hadoop的HADOOP_OPTS，将这些参数都设置进去
HADOOP_OPTS="$HADOOP_OPTS -Dhadoop.log.dir=$HADOOP_LOG_DIR"
HADOOP_OPTS="$HADOOP_OPTS -Dhadoop.log.file=$HADOOP_LOGFILE"
HADOOP_OPTS="$HADOOP_OPTS -Dhadoop.home.dir=$HADOOP_PREFIX"
HADOOP_OPTS="$HADOOP_OPTS -Dhadoop.id.str=$HADOOP_IDENT_STRING"
HADOOP_OPTS="$HADOOP_OPTS -Dhadoop.root.logger=${HADOOP_ROOT_LOGGER:-INFO,console}"
if [ "x$JAVA_LIBRARY_PATH" != "x" ]; then
  HADOOP_OPTS="$HADOOP_OPTS -Djava.library.path=$JAVA_LIBRARY_PATH"
  export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$JAVA_LIBRARY_PATH
fi
HADOOP_OPTS="$HADOOP_OPTS -Dhadoop.policy.file=$HADOOP_POLICYFILE"

# Disable ipv6 as it can cause issues（使ipv6不可用）
HADOOP_OPTS="$HADOOP_OPTS -Djava.net.preferIPv4Stack=true"


# put hdfs in classpath if present（将hdfs的相关东西设置到CLASSPATH）
if [ "$HADOOP_HDFS_HOME" = "" ]; then
  if [ -d "${HADOOP_PREFIX}/$HDFS_DIR" ]; then
    export HADOOP_HDFS_HOME=$HADOOP_PREFIX
  fi
fi
if [ -d "$HADOOP_HDFS_HOME/$HDFS_DIR/webapps" ]; then
  CLASSPATH=${CLASSPATH}:$HADOOP_HDFS_HOME/$HDFS_DIR
fi
if [ -d "$HADOOP_HDFS_HOME/$HDFS_LIB_JARS_DIR" ]; then
  CLASSPATH=${CLASSPATH}:$HADOOP_HDFS_HOME/$HDFS_LIB_JARS_DIR'/*'
fi
CLASSPATH=${CLASSPATH}:$HADOOP_HDFS_HOME/$HDFS_DIR'/*'

# put yarn in classpath if present（将yarn的相关东西设置到CLASSPATH）
if [ "$HADOOP_YARN_HOME" = "" ]; then
  if [ -d "${HADOOP_PREFIX}/$YARN_DIR" ]; then
    export HADOOP_YARN_HOME=$HADOOP_PREFIX
  fi
fi
if [ -d "$HADOOP_YARN_HOME/$YARN_DIR/webapps" ]; then
  CLASSPATH=${CLASSPATH}:$HADOOP_YARN_HOME/$YARN_DIR
fi
if [ -d "$HADOOP_YARN_HOME/$YARN_LIB_JARS_DIR" ]; then
  CLASSPATH=${CLASSPATH}:$HADOOP_YARN_HOME/$YARN_LIB_JARS_DIR'/*'
fi
CLASSPATH=${CLASSPATH}:$HADOOP_YARN_HOME/$YARN_DIR'/*'

# put mapred in classpath if present AND different from YARN（将mapred的相关东西设置到CLASSPATH）
if [ "$HADOOP_MAPRED_HOME" = "" ]; then
  if [ -d "${HADOOP_PREFIX}/$MAPRED_DIR" ]; then
    export HADOOP_MAPRED_HOME=$HADOOP_PREFIX
  fi
fi
if [ "$HADOOP_MAPRED_HOME/$MAPRED_DIR" != "$HADOOP_YARN_HOME/$YARN_DIR" ] ; then
  if [ -d "$HADOOP_MAPRED_HOME/$MAPRED_DIR/webapps" ]; then
    CLASSPATH=${CLASSPATH}:$HADOOP_MAPRED_HOME/$MAPRED_DIR
  fi
  if [ -d "$HADOOP_MAPRED_HOME/$MAPRED_LIB_JARS_DIR" ]; then
    CLASSPATH=${CLASSPATH}:$HADOOP_MAPRED_HOME/$MAPRED_LIB_JARS_DIR'/*'
  fi
  CLASSPATH=${CLASSPATH}:$HADOOP_MAPRED_HOME/$MAPRED_DIR'/*'
fi

# Add the user-specified CLASSPATH via HADOOP_CLASSPATH
# Add it first or last depending on if user has
# set env-var HADOOP_USER_CLASSPATH_FIRST(将HADOOP_CLASSPATH设置到CLASSPATH中，并且设置到CLASSPATH前边)
if [ "$HADOOP_CLASSPATH" != "" ]; then
  # Prefix it if its to be preceded
  if [ "$HADOOP_USER_CLASSPATH_FIRST" != "" ]; then
    CLASSPATH=${HADOOP_CLASSPATH}:${CLASSPATH}
  else
    CLASSPATH=${CLASSPATH}:${HADOOP_CLASSPATH}
  fi
fi