getdata

#!/bin/bash

HADOOP_HOME='/home/excelle08/hadoop'
NAMENODE='tmp/namenode-metrics.out'
DATANODE='tmp/datanode-metrics.out'

task_name=$1
start_time=$2
end_time=$3

if [ -z "$1" -o -z "$2" -o -z "$3" ];then
    echo "Usage: $0 <Task name> <Start time> <End time>";
    exit 0;
fi

if ! [ -f "namenode" -a -f "datanode" ];then
    echo "Namenode or datanode file does not exist.";
    exit 1;
fi

if [ -d $task_name ];then
    i=1
    while [ -d "$task_name.$i" ];do
        i=`expr $i + 1`;
    done
    task_name=$task_name.$i;
fi
mkdir -p $task_name;
echo "Get namenode:"
cat namenode | while read -r node;
do
    ssh -n $node "cat $HADOOP_HOME/$NAMENODE | grep dfs.namenode | while read -r line;do time=\`echo \$line | cut -d ' ' -f 1\`;if [ \$time -ge $start_time -a \$time -le $end_time ];then echo \$line;fi done" > $task_name/$node.namenode.out;
    printf "$node\r"
done

echo "Get datanode in parallel:"
cat datanode | while read -r node;
do
    (ssh -n $node "cat $HADOOP_HOME/$DATANODE | grep dfs.datanode | while read -r line;do time=\`echo \$line | cut -d ' ' -f 1\`;if [ \$time -ge $start_time -a \$time -le $end_time ];then echo \$line;fi done" > $task_name/$node.datanode.out;printf "$node\r") &
done