| #!/bin/bash |
| |
| # Start all hadoop daemons. Run this on master node. |
| |
| bin=`dirname "$0"` |
| bin=`cd "$bin"; pwd` |
| |
| # start dfs daemons |
| # start namenode after datanodes, to minimize time namenode is up w/o data |
| # note: datanodes will log connection errors until namenode starts |
| "$bin"/hadoop-daemons.sh start datanode |
| "$bin"/hadoop-daemon.sh start namenode |
| |
| # start mapred daemons |
| # start jobtracker first to minimize connection errors at startup |
| "$bin"/hadoop-daemon.sh start jobtracker |
| "$bin"/hadoop-daemons.sh start tasktracker |
