晚上花了两个小时看了一下，先配置起来，貌似很大型的一个东西
先配置了 Pseudo-Distributed Operation 模式，就是伪分布式，所有都是在本机执行
一步步按照quickstart来配置，但是启动的时候，懒得配置sshd，所以要修改启动脚本才能实现：
./bin/slaves.sh
# ssh $HADOOP_SSH_OPTS $slave $”${@// /\\ }” \
# 2>&1 | sed “s/^/$slave: /” &
SLAVE_SCRIPT=${@// /\\ }
bash -c “$SLAVE_SCRIPT”
starting namenode, logging to /home/ssmax/hadoop-0.17.1/bin/../logs/hadoop-ssmax-namenode-gz63-238.i.netease.com.out
starting datanode, logging to /home/ssmax/hadoop-0.17.1/bin/../logs/hadoop-ssmax-datanode-gz63-238.i.netease.com.out
starting secondarynamenode, logging to /home/ssmax/hadoop-0.17.1/bin/../logs/hadoop-ssmax-secondarynamenode-gz63-238.i.netease.com.out
starting jobtracker, logging to /home/ssmax/hadoop-0.17.1/bin/../logs/hadoop-ssmax-jobtracker-gz63-238.i.netease.com.out
starting tasktracker, logging to /home/ssmax/hadoop-0.17.1/bin/../logs/hadoop-ssmax-tasktracker-gz63-238.i.netease.com.out
namenode/datanode/secondarynamenode/jobtracker/tasktracker 全部启动了，才算成功，一开始datanode没有起来，搞了很久，都是could only be replicated to 0 nodes, instead of 1错误。。。
起来了以后就可以用./bin/hadoop dfs xxx来搞嘢。。。明天继续。。。
Leave a Reply