网上视频和书本都说：启动之前要格式化，本节就来讲讲解格式化HDFS命令背后的原理！ 先来看看执行结果！ 下面我们来研究原理！ ===================== 首先需要研究下
网上视频和书本都说：启动之前要格式化，本节就来讲讲解格式化HDFS命令背后的原理！先来看看执行结果！
下面我们来研究原理！
=====================
首先需要研究下./bin/hadoop脚本怎么运行
THIS="$0"while [ -h "$THIS" ]; do ls=`ls -ld "$THIS"` link=`expr "$ls" : '.*-> /(.*/)$'` if expr "$link" : '.*/.*' > /dev/null; then THIS="$link" else THIS=`dirname "$THIS"`/"$link" fidone意思是判断./bin/hadoop是否是一个符号链接，这里不是符号链接，所以不用太关注这段代码！$0的值是 ./bin/hadoop------------
# if no args specified, show usageif [ $# = 0 ]; then echo "Usage: hadoop COMMAND" echo "where COMMAND is one of:" echo " namenode -format format the DFS filesystem" echo " namenode run the DFS namenode" echo " datanode run a DFS datanode" echo " dfs run a DFS admin client" echo " fsck run a DFS filesystem checking utility" echo " jobtracker run the MapReduce job Tracker node" echo " tasktracker run a MapReduce task Tracker node" echo " job manipulate MapReduce jobs" echo " jar <jar> run a jar file" echo " or" echo " CLASSNAME run the class named CLASSNAME" echo "Most commands print help when invoked w/o parameters." exit 1fi接下来是判断./bin/hadoop后面的参数个数，如果是0，表明用户还不太清楚这个命令的用法，则给出这个命令的用法！很简单！---------
# get argumentsCOMMAND=$1shift这里是获取参数，如果执行./bin/hadoop namenode -format则COMMAND的值就是namenode然后执行完shift后，$1就变成-format
----------
# some directoriesTHIS_DIR=`dirname "$THIS"`HADOOP_HOME=`cd "$THIS_DIR/.." ; pwd`接下来是获取2个变量值分别是./bin/usr/local/hadoop-0.1.0---
# Allow alternate conf dir location.HADOOP_CONF_DIR="${HADOOP_CONF_DIR:-$HADOOP_HOME/conf}"接下来就是获取HADOOP_CONF_DIR值为/usr/local/hadoop-0.1.0/conf=============
if [ -f "${HADOOP_CONF_DIR}/hadoop-env.sh" ]; then source "${HADOOP_CONF_DIR}/hadoop-env.sh"fi导入hadoop-env.sh里的环境变量，可见后续如果有需求的话，可以修改hadoop-env.sh这个文件就可以了而不是修改hadoop文件，很方便！备注：针对hadoop-env.sh这个文件，我只添加了export JAVA_HOME=/usr/lib/jvm/jdk1.7.0_21===
# some Java parametersif [ "$JAVA_HOME" != "" ]; then #echo "run java in $JAVA_HOME" JAVA_HOME=$JAVA_HOMEfi给JAVA_HOME赋值!结果自然就是 /usr/lib/jvm/jdk1.7.0_2==========
if [ "$JAVA_HOME" = "" ]; then echo "Error: JAVA_HOME is not set." exit 1fi这个就是校验JAVA_HOME是否配置了=======
JAVA=$JAVA_HOME/bin/javaJAVA_HEAP_MAX=-Xmx1000m 设置2个变量，值分别如下：/usr/lib/jvm/jdk1.7.0_21/bin/java-Xmx1000m---
# check envvars which might override default argsif [ "$HADOOP_HEAPSIZE" != "" ]; then #echo "run with heapsize $HADOOP_HEAPSIZE" JAVA_HEAP_MAX="-Xmx""$HADOOP_HEAPSIZE""m" #echo $JAVA_HEAP_MAXfi这里，通过echo得知 $HADOOP_HEAPSIZE为空，所以也不会执行！===
# CLASSPATH initially contains $HADOOP_CONF_DIRCLASSPATH="${HADOOP_CONF_DIR}"CLASSPATH=${CLASSPATH}:$JAVA_HOME/lib/tools.jar最终CLASSPATH的值是/usr/local/hadoop-0.1.0/conf:/usr/lib/jvm/jdk1.7.0_21/lib/tools.jar# for developers, add Hadoop classes to CLASSPATHif [ -d "$HADOOP_HOME/build/classes" ]; then CLASSPATH=${CLASSPATH}:$HADOOP_HOME/build/classesfiif [ -d "$HADOOP_HOME/build/webapps" ]; then CLASSPATH=${CLASSPATH}:$HADOOP_HOME/buildfiif [ -d "$HADOOP_HOME/build/test/classes" ]; then CLASSPATH=${CLASSPATH}:$HADOOP_HOME/build/test/classesfi经过修饰，CLASSPATH最终的值是/usr/local/hadoop-0.1.0/conf:/usr/lib/jvm/jdk1.7.0_21/lib/tools.jar:/usr/local/hadoop-0.1.0/build/classes:/usr/local/hadoop-0.1.0/build:/usr/local/hadoop-0.1.0/build/test/classes这样就把所有的相关类包含进来了！！！！！！！！！！！！！
===
# so that filenames w/ spaces are handled correctly in loops belowIFS=这个简单！---
# for releases, add hadoop jars & webapps to CLASSPATHif [ -d "$HADOOP_HOME/webapps" ]; then CLASSPATH=${CLASSPATH}:$HADOOP_HOMEfifor f in $HADOOP_HOME/hadoop-*.jar; do CLASSPATH=${CLASSPATH}:$f;done# add libs to CLASSPATHfor f in $HADOOP_HOME/lib/*.jar; do CLASSPATH=${CLASSPATH}:$f;donefor f in $HADOOP_HOME/lib/jetty-ext/*.jar; do CLASSPATH=${CLASSPATH}:$f;done最终CLASSPATH的值是/usr/local/hadoop-0.1.0/conf:/usr/lib/jvm/jdk1.7.0_21/lib/tools.jar:/usr/local/hadoop-0.1.0/build/classes:/usr/local/hadoop-0.1.0/build:/usr/local/hadoop-0.1.0/build/test/classes:/usr/local/hadoop-0.1.0:/usr/local/hadoop-0.1.0/hadoop-0.1.0-examples.jar:/usr/local/hadoop-0.1.0/hadoop-0.1.0.jar:/usr/local/hadoop-0.1.0/lib/commons-logging-api-1.0.4.jar:/usr/local/hadoop-0.1.0/lib/jetty-5.1.4.jar:/usr/local/hadoop-0.1.0/lib/junit-3.8.1.jar:/usr/local/hadoop-0.1.0/lib/lucene-core-1.9.1.jar:/usr/local/hadoop-0.1.0/lib/servlet-api.jar:/usr/local/hadoop-0.1.0/lib/jetty-ext/ant.jar:/usr/local/hadoop-0.1.0/lib/jetty-ext/commons-el.jar:/usr/local/hadoop-0.1.0/lib/jetty-ext/jasper-compiler.jar:/usr/local/hadoop-0.1.0/lib/jetty-ext/jasper-runtime.jar:/usr/local/hadoop-0.1.0/lib/jetty-ext/jsp-api.jar===
# figure out which class to runif [ "$COMMAND" = "namenode" ] ; then CLASS='org.apache.hadoop.dfs.NameNode'elif [ "$COMMAND" = "datanode" ] ; then CLASS='org.apache.hadoop.dfs.DataNode'elif [ "$COMMAND" = "dfs" ] ; then CLASS=org.apache.hadoop.dfs.DFSShellelif [ "$COMMAND" = "fsck" ] ; then CLASS=org.apache.hadoop.dfs.DFSckelif [ "$COMMAND" = "jobtracker" ] ; then CLASS=org.apache.hadoop.mapred.JobTrackerelif [ "$COMMAND" = "tasktracker" ] ; then CLASS=org.apache.hadoop.mapred.TaskTrackerelif [ "$COMMAND" = "job" ] ; then CLASS=org.apache.hadoop.mapred.JobClientelif [ "$COMMAND" = "jar" ] ; then JAR="$1" shift CLASS=`"$0" org.apache.hadoop.util.PrintJarMainClass "$JAR"` if [ $? != 0 ]; then echo "Error: Could not find main class in jar file $JAR" exit 1 fi CLASSPATH=${CLASSPATH}:${JAR}else CLASS=$COMMANDfi这里显然执行./bin/hadoop namenode -format的话CLASS就是 org.apache.hadoop.dfs.NameNode===
# cygwin path translationif expr `uname` : 'CYGWIN*' > /dev/null; then CLASSPATH=`cygpath -p -w "$CLASSPATH"`fi我的是在linux中运行的，所以不用管！---
最后就是执行了！
# run itexec "$JAVA" $JAVA_HEAP_MAX $HADOOP_OPTS -classpath "$CLASSPATH" $CLASS "$@"相关的值分别如下：/usr/lib/jvm/jdk1.7.0_21/bin/java-Xmx1000m-classpath/usr/local/hadoop-0.1.0/conf:/usr/lib/jvm/jdk1.7.0_21/lib/tools.jar:/usr/local/hadoop-0.1.0/build/classes:/usr/local/hadoop-0.1.0/build:/usr/local/hadoop-0.1.0/build/test/classes:/usr/local/hadoop-0.1.0:/usr/local/hadoop-0.1.0/hadoop-0.1.0-examples.jar:/usr/local/hadoop-0.1.0/hadoop-0.1.0.jar:/usr/local/hadoop-0.1.0/lib/commons-logging-api-1.0.4.jar:/usr/local/hadoop-0.1.0/lib/jetty-5.1.4.jar:/usr/local/hadoop-0.1.0/lib/junit-3.8.1.jar:/usr/local/hadoop-0.1.0/lib/lucene-core-1.9.1.jar:/usr/local/hadoop-0.1.0/lib/servlet-api.jar:/usr/local/hadoop-0.1.0/lib/jetty-ext/ant.jar:/usr/local/hadoop-0.1.0/lib/jetty-ext/commons-el.jar:/usr/local/hadoop-0.1.0/lib/jetty-ext/jasper-compiler.jar:/usr/local/hadoop-0.1.0/lib/jetty-ext/jasper-runtime.jar:/usr/local/hadoop-0.1.0/lib/jetty-ext/jsp-api.jarorg.apache.hadoop.dfs.NameNode-format这样就运行起来了！
很简单！
