Hadoop脚本分析

start-all.sh

libexec/hadoop-config.sh        --  设置变量
sbin/start-dfs.sh   --config $HADOOP_CONF_DIR  --启动hdfs
sbin/start-yarn.sh  --config $HADOOP_CONF_DIR  --启动yarn

libexec/hadoop-config.sh –设置变量

COMMON_DIR
...
HADOOP_CONF_DIR=...
HEAF_SIZE=1000M,
CLASSPATH=...

sbin/start-dfs.sh

–config $HADOOP_CONF_DIR –启动hdfs

1.libexec/hdfs-cofig.sh

2.#获得名称节点主机名
NAMENAODES = hdfs getconf -namenodes

3.启动名称节点
"$HADOOP_PREFIX/bsin/hadoop-daemnos.sh" \ 
--config "$HADOOP_CONF_DIR" \
--hostnames "$NAMENODES"
--script "$bin/hdfs" start namenode $nameStartOpt

4.启动datanode
"HADOOP——PREFIX/SBIN/hadoop-daemods.sh" \ 
    --config "$HADOOP_CONF_DIR" \
    --script "$bin/hdfs" start datanode $dataStartOpt

5.启动2nn
"$HADOOP_prefix/sbin/hadoop-daemons.sh"    \
    --config "$HADOOP_CONF_DIR" \
    --hostnames "$SECONDARY_NAMENODES" \
    --script "$bin/hdfs" start secondarynameneode

libexec/hdfs-config.sh

libexec/hdfs-config.sh

sbin/hadoop-daemons.sh

–启动守护进程脚本

1.libexec/hdfs -config.sh   --启动守护进程脚本
2.slave.sh "--config $HADOOP CONF DIR cd "$HADOOP_PREFIX"\;"$bin/hadoop-daemon.sh" --config $HADOOP_CONF_DIR"$@"

#循环slaves文件，通过ssh方式远程登录主机，执行相应命令

[bin/hadoop-daemon.sh]
hadoop-config.sh

bin/hads