azure_hadoop_prep.sh

#!/bin/bash
NODE_TYPE=$1
HDD=()
SSD=()
i=0
j=0

etc_configs() {
    cp /etc/fstab ~/
    cp /etc/fstab /etc/fstab.original
    sed -i "s/%admin ALL=(ALL) ALL/%admin ALL=(ALL) NOPASSWD:ALL/g" /etc/sudoers
    sed -i "s/positive-time-to-live	hosts	60/positive-time-to-live	hosts   3600/" /etc/nscd.conf
    echo -e "root soft nofile 32768\nroot hard nofile 32768" >> /etc/security/limits.conf
    echo -e "session required pam_limits.so" >> /etc/pam.d/common-session
    echo -e "session required pam_limits.so" >> /etc/pam.d/common-session-noninteractive
    echo -e "vm.swappiness=10" >> /etc/sysctl.conf
    echo -e "vm.vfs_cache_pressure=50" >> /etc/sysctl.conf
}

install_pkgs() {
    apt-key adv --recv-keys --keyserver keyserver.ubuntu.com B9733A7A07513CAD
    cd /etc/apt/sources.list.d
    wget http://public-repo-1.hortonworks.com/ambari/ubuntu12/2.x/updates/2.2.2.0/ambari.list
    apt-get update
    DEBIAN_FRONTEND=noninteractive apt-get install -y openjdk-7-jdk ntp git ambari-agent htop sysstat lsscsi mdadm
    update-rc.d ntp defaults
}

swap_file() {
    fallocate -l 56G /mnt/swapfile
    chmod 600 /mnt/swapfile
    mkswap /mnt/swapfile
    swapon /mnt/swapfile
    echo "/mnt/swapfile none swap defaults 0 0" >> ~/fstab
}

check_scsi() {
    lsscsi > ~/scsi.txt
    mapfile -t SCSI < ~/scsi.txt
    OSDEV=$(df | awk 'FNR == 2 {print $1}' | sed -e 's/1//')
    for line in "${SCSI[@]}"; do
        TYPE=$(echo $line | awk '{print $2}')
        DEVICE=$(echo $line | awk '{print $7}')
        # tuple is scsi_host,channel,target_number,LUN
        TUPLE=$(echo $line | awk '{print $1}' | sed -e 's/^\[//;s/\]$//')
        TGT=$(echo $TUPLE | awk -F':' '{print $3}')
        LUN=$(echo $TUPLE | awk -F':' '{print $4}')
        # lun 0-12 are standard data disks, target_number 1 is local SSD, also skip cd/dvd + OS device 
        if [ $LUN -le 12 ] && [ $TGT -ne 1 ] && [ $TYPE != "cd/dvd" ] && [ $DEVICE != $OSDEV ];
        then
            HDD[$i]="${DEVICE}"
        # luns 13 and 14 are premium disks on storm/kafka nodes
        elif [ $LUN -gt 12 ];
        then
            case "$NODE_TYPE" in
                realtime)
                    SSD[$j]="${DEVICE}"
                    j=$[$j+1]
                    ;;
                worker)
                    HDD[$i]="${DEVICE}"
                    ;;
            esac
        fi
        i=$[$i+1]
    done
    case "$NODE_TYPE" in
        realtime)
            make_filesystems "${HDD[@]}" "1" "main" 
            make_filesystems "${SSD[@]}" "0" "kafka"
            ;;
        worker)
            make_filesystems "${HDD[@]}" "1" "main"
            ;;
    esac
}

make_filesystems() {
    disks=("$@")
    ((last_idx=${#disks[@]} - 1))
    mp=${disks[last_idx]}
    unset disks[last_idx]
    ((last_idx=${#disks[@]} - 1))
    b=${disks[last_idx]}
    unset disks[last_idx]
    
    i=0
    RAID_CMD="sudo mdadm --create /dev/md${#disks[@]} --level=stripe --raid-devices=${#disks[@]} "
    for d in "${disks[@]}"; do
        (echo n; echo p; echo 1; echo ; echo ; echo t; echo fd; echo p; echo w;) | fdisk ${d}
        RAID_CMD+="${d}1 "
        i=$[$i+1]
    done

    eval "$RAID_CMD"
    mkdir -p /media/${mp}
    mkfs.ext4 /dev/md${#disks[@]}
    uuid=$(blkid | grep "md${#disks[@]}" | grep -oP '[-a-z0-9]{36}')
    echo "UUID=${uuid} /media/${mp} ext4 defaults,noatime,barrier=${b} 0 0" >> ~/fstab
}

etc_configs
install_pkgs
check_scsi
mv ~/fstab /etc/fstab
mount -a