监控RAID状态的脚本如何编写

本文目录导读：

1. 通用RAID监控脚本
2. 简化版监控脚本
3. Zabbix集成版本
4. 使用说明
5. 监控指标说明
6. 最佳实践

我来提供一个监控RAID状态的脚本方案,涵盖常见的RAID卡类型。

通用RAID监控脚本

#!/bin/bash
# RAID状态监控脚本
# 支持：LSI/MegaRAID, HP Smart Array, 软件RAID(mdadm)
LOG_FILE="/var/log/raid_monitor.log"
EMAIL="admin@example.com"
HOSTNAME=$(hostname)
# 日志函数
log_message() {
    echo "[$(date '+%Y-%m-%d %H:%M:%S')] $1" >> "$LOG_FILE"
    echo "[$(date '+%Y-%m-%d %H:%M:%S')] $1"
}
# 发送告警邮件
send_alert() {
    subject="[RAID ALERT] $HOSTNAME - $1"
    echo "$2" | mail -s "$subject" "$EMAIL"
}
# 检查MR RAID卡 (MegaRAID/LSI)
check_megaraid() {
    if command -v storcli &>/dev/null; then
        log_message "检查 MegaRAID RAID 状态..."
        # 获取所有RAID卡信息
        controllers=$(storcli show all | grep "Number of Controllers" | awk '{print $NF}')
        for ((c=0; c<controllers; c++)); do
            # 检查RAID卷状态
            vd_status=$(storcli /c$c/vall show | grep "State")
            if echo "$vd_status" | grep -qi "degraded\|failed\|critical"; then
                log_message "警告: RAID控制器$c 的虚拟磁盘状态异常!"
                send_alert "RAID状态异常" "控制器$c 虚拟磁盘异常:\n$vd_status"
            fi
            # 检查物理磁盘状态
            pd_status=$(storcli /c$c/eall/sall show | grep "State")
            if echo "$pd_status" | grep -qi "failed\|missing"; then
                log_message "警告: RAID控制器$c 存在故障磁盘!"
                send_alert "磁盘故障" "控制器$c 磁盘故障:\n$pd_status"
            fi
            # 检查BBU状态
            bbu_status=$(storcli /c$c/bbu show | grep "Battery")
            if echo "$bbu_status" | grep -qi "failed\|disconnected"; then
                log_message "警告: 控制器$c BBU状态异常!"
            fi
        done
    elif command -v megaraid-status &>/dev/null; then
        log_message "使用megaraid-status检查..."
        megaraid-status | grep -qi "degraded\|failed\|offline"
        if [ $? -eq 0 ]; then
            log_message "RAID状态异常!"
        fi
    fi
}
# 检查HP Smart Array
check_hp_raid() {
    if command -v hpssacli &>/dev/null; then
        log_message "检查 HP Smart Array RAID 状态..."
        # 列出所有控制器
        controllers=$(hpssacli ctrl all show status | grep "Smart Array")
        # 检查每个控制器状态
        hpssacli ctrl all show status | while read line; do
            if echo "$line" | grep -qi "failed\|degraded\|critical"; then
                log_message "警告: $line"
                send_alert "HP RAID异常" "$line"
            fi
        done
        # 检查逻辑驱动器状态
        hpssacli ctrl slot=0 ld all show status | while read line; do
            if echo "$line" | grep -qi "failed\|degraded"; then
                log_message "逻辑驱动器异常: $line"
            fi
        done
    elif command -v hpacucli &>/dev/null; then
        log_message "使用hpacucli检查..."
        hpacucli ctrl all show status
    fi
}
# 检查软件RAID (mdadm)
check_software_raid() {
    if command -v mdadm &>/dev/null; then
        log_message "检查软件RAID状态..."
        # 检查所有md设备
        for md in /dev/md*; do
            if [ -b "$md" ]; then
                md_status=$(mdadm --detail "$md" | grep "State")
                raid_device=$(basename "$md")
                if echo "$md_status" | grep -qi "degraded\|failed\|recovering"; then
                    log_message "警告: $raid_device 状态异常 - $md_status"
                    send_alert "软件RAID异常" "$raid_device: $md_status"
                fi
                # 检查磁盘状态
                failed_disks=$(mdadm --detail "$md" | grep "faulty\|failed\|removed")
                if [ -n "$failed_disks" ]; then
                    log_message "警告: $raid_device 存在故障磁盘:\n$failed_disks"
                fi
            fi
        done
        # 检查/proc/mdstat
        if grep -qi "failed\|degraded" /proc/mdstat 2>/dev/null; then
            log_message "检测到RAID异常!"
            cat /proc/mdstat
        fi
    fi
}
# 检查硬件RAID卡工具是否存在
check_tools() {
    tools_found=0
    # 检测可用工具
    for tool in storcli megaraid-status hpssacli hpacucli mdadm; do
        command -v $tool &>/dev/null && {
            log_message "找到工具: $tool"
            tools_found=$((tools_found + 1))
        }
    done
    if [ $tools_found -eq 0 ]; then
        log_message "未找到任何RAID监控工具!"
        return 1
    fi
    return 0
}
# 主函数
main() {
    log_message "=== 开始RAID状态检查 ==="
    # 检查工具是否存在
    check_tools || {
        log_message "已安装的监控工具:"
        dpkg -l | grep -i "storcli\|megaraid\|hpssacli\|hpacucli\|mdadm" 2>/dev/null
        exit 1
    }
    # 分别检查不同类型的RAID
    check_megaraid
    check_hp_raid
    check_software_raid
    log_message "=== RAID状态检查完成 ==="
}
# 执行主函数
main

简化版监控脚本

#!/bin/bash
# 简化版RAID监控脚本
LOG="/var/log/raid_check.log"
DATE=$(date '+%Y-%m-%d %H:%M:%S')
echo "[$DATE] RAID状态检查开始" >> $LOG
# 检查软件RAID
if [ -f /proc/mdstat ]; then
    if grep -qi "degraded\|failed\|recovering" /proc/mdstat; then
        echo "警告: 软件RAID异常" >> $LOG
        cat /proc/mdstat >> $LOG
    fi
fi
# 尝试检查硬件RAID
for cmd in "storcli /c0/vall show" "megaraid-status" "hpssacli ctrl all show status"; do
    if command -v ${cmd%% *} &>/dev/null; then
        eval $cmd 2>/dev/null | grep -qi "failed\|degraded\|critical"
        if [ $? -eq 0 ]; then
            echo "警告: 硬件RAID异常" >> $LOG
            eval $cmd >> $LOG 2>/dev/null
        fi
    fi
done
echo "[$DATE] 检查完成" >> $LOG

Zabbix集成版本

#!/bin/bash
# Zabbix集成的RAID监控脚本
case "$1" in
    discovery)
        # 发现所有RAID设备
        echo '{"data":['
        first=1
        # 发现软件RAID
        for md in /dev/md*; do
            if [ -b "$md" ]; then
                [ $first -ne 1 ] && echo ","
                echo "{\"{#RAID_NAME}\":\"$(basename $md)\",\"{#RAID_TYPE}\":\"software\"}"
                first=0
            fi
        done
        # 发现硬件RAID
        if command -v storcli &>/dev/null; then
            for ctrl in $(storcli show all | grep -E "^\s+[0-9]+:" | awk '{print $1}'); do
                [ $first -ne 1 ] && echo ","
                echo "{\"{#RAID_NAME}\":\"controller$ctrl\",\"{#RAID_TYPE}\":\"hardware\"}"
                first=0
            done
        fi
        echo ']}'
        ;;
    status)
        # 检查指定RAID状态
        raid_name=$2
        raid_type=$3
        if [ "$raid_type" = "software" ]; then
            mdadm --detail /dev/$raid_name 2>/dev/null | grep "State" | awk '{print $2}'
        else
            # 在此添加硬件RAID状态获取逻辑
            echo "unknown"
        fi
        ;;
    count)
        # 统计异常RAID数量
        count=0
        # 检查软件RAID
        if [ -f /proc/mdstat ]; then
            count=$((count + $(grep -c "degraded\|failed" /proc/mdstat 2>/dev/null)))
        fi
        echo $count
        ;;
    *)
        echo "Usage: $0 {discovery|status name type|count}"
        exit 1
        ;;
esac

使用说明

安装依赖

Ubuntu/Debian:

# 安装软件RAID工具
apt-get install mdadm
# 安装MegaRAID工具
# 从Broadcom官网下载storcli
wget https://docs.broadcom.com/docs/.../storcli.deb
dpkg -i storcli.deb
# 安装HP工具
apt-get install hpssacli  # 或从HP官网下载

CentOS/RHEL:

# 安装软件RAID工具
yum install mdadm
# 安装MegaRAID工具
rpm -ivh storcli.rpm
# 安装HP工具
yum install hpssacli

设置定时任务

# 添加到crontab，每30分钟检查一次
crontab -e
# 添加以下行：
*/30 * * * * /path/to/raid_monitor.sh

配置邮件告警

# 安装邮件工具
apt-get install mailutils  # Debian/Ubuntu
yum install mailx          # CentOS/RHEL
# 配置SSMTP（如果使用外部邮件服务器）
apt-get install ssmtp
# 编辑 /etc/ssmtp/ssmtp.conf

监控指标说明

最佳实践

1. 定期检查：建议每15-30分钟检查一次
2. 日志轮转：配置logrotate管理日志文件
3. 多重告警：邮件 + SNMP + 系统日志
4. 自动化修复：对有热备盘的RAID，自动触发重建
5. 性能监控：同时监控IO性能，预测磁盘故障

这个脚本提供了完整的RAID监控方案,可以根据实际环境选择合适的版本使用。