最近在看Linux I/O子系统相关的资料，需要监控每台服务器的磁盘I/O，特别是MongoDB数据库服务器的磁盘I/O，由于每台服务器的磁盘名称可能不一样，所以需要用到Zabbix的LLD功能来发现磁盘并监控I/O
参考文章
http://www.programering.com/a/MDOwYDNwATc.html
1.编写发现磁盘的脚步
shell版本disk_discovery.sh
#!/bin/bash
diskarray=(`cat /proc/diskstats |grep -E "\bsd[a-z]\b|\bxvd[a-z]\b|\bvd[a-z]\b"|awk '{print $3}'|sort|uniq 2>/dev/null`)
length=${#diskarray[@]}
printf "{\n"
printf '\t'"\"data\":["
for ((i=0;i<$length;i++))
do
printf '\n\t\t{'
printf "\"{#DISK_NAME}\":\"${diskarray[$i]}\"}"
if [ $i -lt $[$length-1] ];then
printf ','
fi
done
printf "\n\t]\n"
printf "}\n"
$ sh disk_discovery.sh
{
"data":[
{"{#DISK_NAME}":"xvda"},
{"{#DISK_NAME}":"xvdb"}
]
}
Python版本disk_discovery.py
#/usr/bin/python
#This script is used to discovery disk on the server
import subprocess
import json
args="cat /proc/diskstats |grep -E '\ssd[a-z]\s|\sxvd[a-z]\s|\svd[a-z]\s'|awk '{print $3}'|sort|uniq 2>/dev/null"
t=subprocess.Popen(args,shell=True,stdout=subprocess.PIPE).communicate()[0]
disks=[]
for disk in t.split('\n'):
if len(disk) != 0:
disks.append({'{#DISK_NAME}':disk})
print json.dumps({'data':disks},indent=4,separators=(',',':'))
$ python disk_discovery.py
{
"data":[
{
"{#DISK_NAME}":"xvda"
},
{
"{#DISK_NAME}":"xvdb"
}
]
}
2.编写获取磁盘I/O信息的脚本
使用iostat收集磁盘I/O信息
#/bin/sh
device=$1
item=$2
/usr/bin/iostat -dxkt 1 5 > /tmp/iostat_output 2>/dev/null
case $item in
rrqm)
/usr/bin/tail -n20 /tmp/iostat_output |grep "\b$device\b"|tail -1|awk '{print $2}'
;;
wrqm)
/usr/bin/tail -n20 /tmp/iostat_output |grep "\b$device\b"|tail -1|awk '{print $3}'
;;
rps)
/usr/bin/tail -n20 /tmp/iostat_output |grep "\b$device\b"|tail -1|awk '{print $4}'
;;
wps)
/usr/bin/tail -n20 /tmp/iostat_output |grep "\b$device\b" |tail -1|awk '{print $5}'
;;
rKBps)
/usr/bin/tail -n20 /tmp/iostat_output |grep "\b$device\b" |tail -1|awk '{print $6}'
;;
wKBps)
/usr/bin/tail -n20 /tmp/iostat_output |grep "\b$device\b" |tail -1|awk '{print $7}'
;;
avgrq-sz)
/usr/bin/tail -n20 /tmp/iostat_output |grep "\b$device\b" |tail -1|awk '{print $8}'
;;
avgqu-sz)
/usr/bin/tail -n20 /tmp/iostat_output |grep "\b$device\b" |tail -1|awk '{print $9}'
;;
await)
/usr/bin/tail -n20 /tmp/iostat_output |grep "\b$device\b" |tail -1|awk '{print $10}'
;;
svctm)
/usr/bin/tail -n20 /tmp/iostat_output |grep "\b$device\b" |tail -1|awk '{print $11}'
;;
util)
/usr/bin/tail -n20 /tmp/iostat_output |grep "\b$device\b" |tail -1|awk '{print $12}'
;;
esac
3.修改Zabbix agent配置文件
添加disk_status.conf
Timeout=10
### Option: UserParameter
# User-defined parameter to monitor. There can be several user-defined parameters.
# Format: UserParameter=<key>,<shell command>
# See 'zabbix_agentd' directory for examples.
#
# Mandatory: no
# Default:
# UserParameter=
UserParameter=disk.discovery,/usr/bin/python /usr/local/zabbix/bin/disk_discovery.py
UserParameter=disk.status[*],/usr/local/zabbix/bin/disk_status.sh $1 $2
这里需要注意，Zabbix agent默认的Timeout值为3秒，由于这里使用iostat -ydxkt 1 3，每隔1秒刷新一次，刷新3次，所以如果获取磁盘信息Timeout设置时间短了的话会出现ZBX_NOTSUPPORTED这样的错误
然后重新加载zabbix agent
4.通过在zabbix server或zabbix proxy端使用zabbix_get获取磁盘信息
$ /usr/local/zabbix/bin/zabbix_get -s 192.168.1.190 -p 10055 -k "disk.discovery"
{
"data":[
{
"{#DISK_NAME}":"xvda"
},
{
"{#DISK_NAME}":"xvdb"
},
{
"{#DISK_NAME}":"xvdc"
}
]
}
$ /usr/local/zabbix/bin/zabbix_get -s 192.168.1.190 -p 10055 -k "disk.status[xvda,wps]"10.00
一些Zabbix相关教程集合：
安装部署分布式监控系统Zabbix 2.06 http://www.linuxidc.com/Linux/2013-07/86942.htm
《安装部署分布式监控系统Zabbix 2.06》 http://www.linuxidc.com/Linux/2013-07/86942.htm
CentOS 6.3下Zabbix安装部署 http://www.linuxidc.com/Linux/2013-05/83786.htm
Zabbix分布式监控系统实践 http://www.linuxidc.com/Linux/2013-06/85758.htm
CentOS 6.3下Zabbix监控apache server-status http://www.linuxidc.com/Linux/2013-05/84740.htm
CentOS 6.3下Zabbix监控MySQL数据库参数 http://www.linuxidc.com/Linux/2013-05/84800.htm
