目標
配置分布式的icinga2監控系統。分布式監控適用於服務器遍布在多個區域,且需要一個master做統一管理。
搭建環境
服務器
系統: ubuntu 15.04/ubuntu 14.04
icinga2主節點: 192.168.19.77 負責分發配置,統一展示監控結果。
icinga2子節點1: 192.168.19.45 負責監控openstack RegionOne區域的所有服務器
icinga2子節點2: 192.168.19.30 負責將空openstack RegionTwo區域的所有服務器
為了充分利用現有的nagios插件,使用nrpe監控服務器。
拓撲圖
安裝配置icinga2
注意: 如果特別說明,所有操作均在192.168.19.77,45,30執行。
設置包源
# add-apt-repository ppa:formorer/icinga
# apt-get update
安裝icinga2
# apt-get install icinga2
安裝classicui界面
在192.168.77上執行
apt-get install icinga2-classicui -y
配置分布式監控
設置ssl
在192.168.77上執行
生成ca證書
#icinga2 pki new-ca
生成各個節點需要用的key,crt
設置節點名
key與crt的名字需要與節點名吻合,默認使用hostname。如果要自定義節點名需要編輯/etc/icinga2/constants.conf
,修改如下配置
const NodeName = "node-master"
node-master是新的節點名
生成key,crt
# cd /tmp
##192.168.19.77
# icinga2 pki new-cert --cn node-master --key node-master.key --csr node-master.csr
# icinga2 pki sign-csr --csr node-master.csr --cert node-master.crt
##192.168.19.45
# icinga2 pki new-cert --cn node-45 --key node-45.key --csr node-45.csr
# icinga2 pki sign-csr --csr node-45.csr --cert node-45.crt
##192.168.19.30
# icinga2 pki new-cert --cn node-30 --key node-30.key --csr node-30.csr
# icinga2 pki sign-csr --csr node-30.csr --cert node-30.crt
分布復制ca.crt,
/etc/icinga2/pki/
目錄。下面顯示的是192.168.19.77的pki目錄
# ll /etc/icinga2/pki/
total 20
drwxr-xr-x 2 root root 4096 May 11 17:19 ./
drwxr-x--- 9 nagios nagios 4096 May 13 12:09 ../
-rw-rw-rw- 1 root root 1688 May 11 15:25 ca.crt
-rw-rw-rw- 1 root root 1663 May 11 15:28 node-master.crt
-rw-rw-rw- 1 root root 3243 May 11 15:26 node-master.key
啟用api功能
# icinga2 feature enable api
添加accept_config = true
, accept_commands = true
參數
# vim /etc/icinga2/features-enabled/api.conf
/**
* The API listener is used for distributed monitoring setups.
*/
object ApiListener "api" {
cert_path = SysconfDir + "/icinga2/pki/" + NodeName + ".crt"
key_path = SysconfDir + "/icinga2/pki/" + NodeName + ".key"
ca_path = SysconfDir + "/icinga2/pki/ca.crt"
ticket_salt = TicketSalt
accept_config = true
accept_commands = true
}
# service icinga2 restart
配置endopoint,zone
EndPoint 需要與NodeName一致。
# vim /etc/icinga2/zones.conf
object Endpoint "node-master" {
host = "192.168.19.77"
}
object Endpoint "node-45" {
host = "192.168.19.45"
}
object Endpoint "node-30" {
host = "192.168.19.30"
}
object Zone "zone-master" {
endpoints = [ "node-master" ]
}
object Zone "zone-45" {
parent = "zone-master"
endpoints = [ "node-45" ]
}
object Zone "zone-30" {
parent = "zone-master"
endpoints = [ "node-30" ]
}
object Zone "global-templates" {
global = true
}
這里配置了4個區域,分別是zone-master,zone-45,zone-30,global-templates。global-templates區域中的配置會分發到所有的區域。
配置文件管理
創建配置目錄,目錄名要與區域名一致。
在192.168.77上執行
# mkdir /etc/icinga2/zones.d/global-templates/
# mkdir /etc/icinga2/zones.d/zone-30/
# mkdir /etc/icinga2/zones.d/zone-45/
# mkdir /etc/icinga2/zones.d/zone-master/
將/etc/icinga2/conf.d中的文件復制到/etc/icinga2/zones.d/global-templates/
# cp -rf /etc/icinga2/conf.d/* /etc/icinga2/zones.d/global-templates/
注釋掉3台服務器上conf.d目錄
# vim /etc/icinga2/icinga2.conf
#include_recursive "conf.d"
配置master監控兩個子節點
object Host NodeName {
import "generic-host"
address = "127.0.0.1"
vars.os = "Linux"
vars.disks["disk"] = {
}
vars.notification["mail"] = {
groups = [ "icingaadmins" ]
}
/*NIC dell2950 */
vars.interfaces["eth0"]={
interface_speed = 100
}
}
object Host "node-45" {
import "generic-host"
address = "192.168.19.45"
vars.os = "Linux"
vars.disks["disk"] = {
}
vars.notification["mail"] = {
groups = [ "icingaadmins" ]
}
/*NIC*/
vars.interfaces["em1"]={
interface_speed = 1000
}
/* openstack */
vars.openstack_controller_listen="192.168.19.45"
vars.openstack["keystone"]="controller"
vars.openstack["cinder"]="controller"
vars.openstack["glance"]="controller"
vars.openstack["heat"]="controller"
vars.openstack["nova"]="controller"
}
object Host "node-30" {
import "generic-host"
address = "192.168.19.30"
vars.os = "Linux"
vars.disks["disk"] = {
}
vars.notification["mail"] = {
groups = [ "icingaadmins" ]
}
/*NIC*/
vars.interfaces["em1"]={
interface_speed = 1000
}
}
清空緩存
# rm -rf /var/lib/icinga2/api/zones/*
重啟icinga2,查看結果
# service icinga2 restart
打開http://192.168.19.77/icinga2-classicui/
配置正確的話,應該可以看到77,45,39服務器的監控信息
使用SNMP監控網絡
在icinga2服務端安裝監控插件
-
安裝依賴
apt-get install snmpd sudo perl libconfig-general-perl libdata-dumper-simple-perl libsort-naturally-perl libexception-class-perl libencode-perl
文檔里漏掉下面幾個包:
apt-get install libdata-dumper-simple-perl libsys-statistics-linux-perl libnet-snmp-perl
-
編譯安裝snmp網絡檢測程序
從
http://www.tontonitch.com/tiki/tiki-index.php?page=Nagios+plugins+-+interfacetable_v3t
下載最新的包,解壓進入目錄后:./configure make fullinstall
-
測試
/usr/local/nagios/libexec/check_interface_table_v3t.pl -V
-
啟用icinga2監控腳本
cp /usr/share/icinga2/include/plugins-contrib.d/network-components.conf /etc/icinga2/zones.d/global-templates/
注意:
修改其中的
command
為正確腳本路徑.修改
-C
為正確的值,默認是public
. -
應用檢測命令到主機上
# vim services.conf apply Service "network"{ import "generic-service" display_name = "check network" check_command = "interfacetable" assign where host.vars.os == "Linux" }
被監控端安裝配置SNMP服務
-
apt-get install snmpd snmp
-
修改監聽地址,連接串,且只允許icinga2服務端連接
# vim /etc/snmp/snmpd.conf agentAddress udp:10.33.10.66:161 rocommunity NmQwZmVmZT 10.33.10.84 #注釋掉這行 #rocommunity public default -V systemonly #注釋這行,它會監聽一個隨機端口 #trapsink localhost public
-
重啟snmpd服務
service snmpd restart
使用nrpe監控服務器
源碼安裝
不用apt的原因是apt安裝的nrpe不接受監控命令傳參。
創建nagios用戶
useradd nagios -M -s /bin/false
編譯安裝
apt-get install build-essential libssl-dev libssl0.9.8 unzip make -y
cd /usr/local/src
wget http://sourceforge.net/projects/nagios/files/nrpe-2.x/nrpe-2.15/nrpe-2.15.tar.gz/download -O nrpe-2.15.tar.gz
tar -xf nrpe-2.15.tar.gz
cd nrpe-2.15
./configure --enable-command-args --with-ssl=/usr/bin/openssl --with-ssl-lib=/usr/lib/x86_64-linux-gnu
make all
make install-daemon
安裝通用插件
apt-get install nagios-plugins -y
添加配置文件
mkdir /etc/nagios
vim /etc/nagios/nrpe.cfg
log_facility=daemon
pid_file=/var/run/nrpe.pid
server_address=172.16.240.30
server_port=5666
nrpe_user=nagios
nrpe_group=nagios
allowed_hosts=127.0.0.1,172.16.240.30
dont_blame_nrpe=1
debug=0
command_timeout=60
connection_timeout=300
command[check_load]=/usr/lib/nagios/libexec/check_load -w 20,15,10 -c 50,40,30
command[check_disk]=/usr/lib/nagios/libexec/check_disk -w 10% -c 5% -W 10% -K 5% -A
command[check_mem]=/usr/lib/nagios/libexec/check_mem -u -C -w 85 -c 90
command[check_proc_num]=/usr/lib/nagios/libexec/check_procs -m PROCS -w 1500:1000 -c 300000:1500
command[check_zombie_procs]=/usr/lib/nagios/libexec/check_procs -w 5 -c 10 -s Z
command[check_swap]=/usr/lib/nagios/libexec/check_swap -a -w 30% -c 15%
command[check_local_port]=/usr/lib/nagios/libexec/check_tcp -H localhost -p $ARG1$ -w 2 -c 5
command[check_linux_raid]=/usr/lib/nagios/libexec/check_linux_raid
command[check_md_raid]=/usr/lib/nagios/libexec/check_md_raid
command[check_icmp]=/usr/lib/nagios/libexec/check_icmp $ARG1$
command[check_lvs]=/usr/lib/nagios/libexec/check_ipvsadm
command[check_backup]=/usr/lib/nagios/libexec/check_backup
command[check_dns]=/usr/lib/nagios/libexec/check_dns -H www.baidu.com
command[check_http]=/usr/lib/nagios/libexec/check_http
command[check_rsyncd]=/usr/lib/nagios/libexec/check_rsyncd
command[check_monitorwebbackup]=/usr/lib/nagios/libexec/check_monitorwebbackup
command[check_monitormysqlbackup]=/usr/lib/nagios/libexec/check_monitormysqlbackup
command[check_diskhealth]=/usr/lib/nagios/libexec/check_openmanage --check storage -b ctrl_fw=all/ctrl_driver=all/ctrl_stdr=all/bat_charge=all/encl=all/ps=all -t 30
command[check_safe]=/usr/lib/nagios/libexec/check_safe $ARG1$
command[check_tcptraffic]=/usr/lib/nagios/plugins/contrib/check_tcptraffic -w $ARG1$ -c $ARG2$ -s $ARG3$ -i $ARG4$
command[check_iostat_io]=/usr/lib/nagios/libexec/check_iostat -d $ARG1$ -i -p
command[check_iostat_waittime]=/usr/lib/nagios/libexec/check_iostat -d $ARG1$ -W -p
command[check_keystone_api]=/usr/lib/nagios/libexec/check_keystone --auth_url "http://$ARG1$:35357/v2.0" --username $ARG2$ --tenant $ARG3$ --password $ARG4$
command[check_keystone_proc]=/usr/lib/nagios/plugins/check_procs -c 1: -u keystone
command[check_cinder_api_proc]=/usr/lib/nagios/libexec/check_service.sh -o linux -s cinder-api
command[check_cinder_scheduler_proc]=/usr/lib/nagios/libexec/check_service.sh -o linux -s cinder-scheduler
command[check_cinder_api]=/usr/lib/nagios/libexec/check_cinder_api --auth_url "http://$ARG1$:35357/v2.0" --username $ARG2$ --tenant $ARG3$ --password $ARG4$
command[check_cinder_scheduler_connectivity]=/usr/lib/nagios/libexec/check_cinder-scheduler.sh
在rc.local加入啟動命令
# vim /etc/rc.local
/usr/local/nagios/bin/nrpe -c /etc/nagios/nrpe.cfg -d
從其他服務器拷貝libexec文件夾
例如: 221.228.84.84(無錫三線)
文件夾位置:/usr/lib/nagios/libexec
注意: 需要將check_icmp
插件設置為root用戶,然后設置setuid,不然沒權限執行命令:
chown root.root /usr/lib/nagios/libexec/check_icmp
chmod u+s /usr/lib/nagios/libexec/check_icmp
chmod g+s /usr/lib/nagios/libexec/check_icmp
安裝check_tcptraffic插件
apt-get install libnagios-plugin-perl libreadonly-xs-perl -y
-
在mac上解壓后用zip重新打包上傳到ubuntu,不然解壓會出錯
-
進入源碼目錄執行下面命令
perl Makefile.PL make make install
-
測試
/usr/lib/nagios/plugins/contrib/check_tcptraffic -w 10485760 -c 10485760 -s 1000 -i em1 /usr/lib/nagios/plugins/check_nrpe -H 127.0.0.1 -c check_tcptraffic -p 5666 -a 10485760 20971520 1000 em1
安裝dell磁盤監控
-
設置軟件源
echo 'deb http://linux.dell.com/repo/community/ubuntu trusty openmanage' | sudo tee -a /etc/apt/sources.list.d/linux.dell.com.sources.list gpg --keyserver pool.sks-keyservers.net --recv-key 1285491434D8786F gpg -a --export 1285491434D8786F | sudo apt-key add -
-
安裝
apt-get update apt-get install srvadmin-all
-
啟動服務
service dataeng start
安裝iostat監控
-
安裝iostat
apt-get install sysstat
-
(如果已有
check_iostat
腳本則忽略) 下載監控腳本 到/usr/lib/nagios/libexec/check_iostat
,添加執行權限,所屬用戶和組改為nagios
注意:要下載評論里的腳本,原始腳本有問題。
-
記得在icinga2的host里配置里加入:
#check disk io and waittime vars.disk_device = "sdb"
修改
sdb
為需要監控的硬盤
啟動nrpe
# /usr/local/nagios/bin/nrpe -c /etc/nagios/nrpe.cfg -d
在icinga2中啟用nrpe
配置LibxecDir常量
根據系統取消選擇相應的注釋:
# vim /etc/icinga2/constants.conf
//ubuntu
//const LibexecDir = "/usr/lib/nagios/libexec"
//youfu centos
//const LibexecDir = "/usr/local/nagios/libexec"
配置checkcommand模板
# vim /etc/icinga2/zones.d/global-templates/nrpe_base.conf
template CheckCommand "nrpe-common" {
import "plugin-check-command"
command = [ LibexecDir + "/check_nrpe" ]
arguments = {
"-H" = "$nrpe_address$"
"-t" = "$nrpe_timeout$"
"-p" = "$nrpe_port$"
"-c" = "$nrpe_command$"
"-a" = {
value = "$nrpe_args$"
repeat_key = false
order = 1
}
}
vars.nrpe_address = "$address$"
vars.nrpe_port = 5666
vars.nrpe_timeout = 60
}
配置常用linux檢測命令
# vim /etc/icinga2/zones.d/global-templates/nrpe_linux.conf
object CheckCommand "nrpe-disk" {
import "nrpe-common"
#vars.nrpe_args = [ "$disk_wfree$%", "$disk_cfree$%"]
vars.nrpe_command = "check_disk"
#vars.disk_wfree = 20
#vars.disk_cfree = 10
}
object CheckCommand "nrpe-diskhealth" {
import "nrpe-common"
vars.nrpe_command = "check_diskhealth"
}
object CheckCommand "nrpe-tcptraffic" {
import "nrpe-common"
vars.nrpe_args = [ "$tcptraffic_wbytes$", "$tcptraffic_cbytes$","$interface_speed$","$interface_name$"]
vars.nrpe_command = "check_tcptraffic"
vars.tcptraffic_wbytes = 10485760 /*10M=10*1024*1024*/
vars.tcptraffic_cbytes = 20971520 /*20M=20*1024*1024*/
#vars.tcptraffic_wbytes = 1 /*10M=10*1024*1024*/
#vars.tcptraffic_cbytes = 2 /*20M=20*1024*1024*/
}
object CheckCommand "nrpe-load" {
import "nrpe-common"
vars.nrpe_command = "check_load"
}
object CheckCommand "nrpe-mem" {
import "nrpe-common"
vars.nrpe_command = "check_mem"
}
object CheckCommand "nrpe-proc_num" {
import "nrpe-common"
vars.nrpe_command = "check_proc_num"
}
object CheckCommand "nrpe-zombie_procs" {
import "nrpe-common"
vars.nrpe_command = "check_zombie_procs"
}
object CheckCommand "nrpe-swap" {
import "nrpe-common"
vars.nrpe_command = "check_swap"
}
object CheckCommand "nrpe-dns" {
import "nrpe-common"
vars.nrpe_command = "check_dns"
}
object CheckCommand "nrpe-safe" {
import "nrpe-common"
vars.nrpe_command = "check_safe"
}
object CheckCommand "nrpe-iostat_io" {
import "nrpe-common"
vars.nrpe_command = "check_iostat_io"
vars.nrpe_args = [ "$disk_device$" ]
}
object CheckCommand "nrpe-iostat_waittime" {
import "nrpe-common"
vars.nrpe_command = "check_iostat_waittime"
vars.nrpe_args = [ "$disk_device$" ]
}
apply Service "check_nrpe:" for (disk_nrpe_linux => config in host.vars.disks) {
import "generic-service"
display_name = "Check Nrpe:" + disk_nrpe_linux
check_command = "nrpe-disk"
vars += config
assign where host.vars.os == "Linux"
}
#Storage Error! No controllers found on ubuntu15.04 dell r720
apply Service "check_nrpe:diskhealth"{
import "generic-service"
display_name = "Check Nrpe: Diskhealth"
check_command = "nrpe-diskhealth"
assign where host.vars.os == "Linux"
}
apply Service "check_nrpe:tcptraffic" for (interface_name =>interface_info in host.vars.interfaces){
import "generic-service"
display_name = "Check Nrpe: Tcptraffic "+ interface_name
check_command = "nrpe-tcptraffic"
vars.interface_name = interface_name
vars+=interface_info
assign where host.vars.interfaces && host.vars.os == "Linux"
}
apply Service "check_nrpe:load"{
import "generic-service"
display_name = "Check Nrpe: Load"
check_command = "nrpe-load"
assign where host.vars.os == "Linux"
}
apply Service "check_nrpe:mem"{
import "generic-service"
display_name = "Check Nrpe: Mem"
check_command = "nrpe-mem"
assign where host.vars.os == "Linux"
}
apply Service "check_nrpe:proc_num"{
import "generic-service"
display_name = "Check Nrpe: proc_num"
check_command = "nrpe-proc_num"
assign where host.vars.os == "Linux"
}
apply Service "check_nrpe:zombie_procs"{
import "generic-service"
display_name = "Check Nrpe: zombie_procs"
check_command = "nrpe-zombie_procs"
assign where host.vars.os == "Linux"
}
apply Service "check_nrpe:swap"{
import "generic-service"
display_name = "Check Nrpe: swap"
check_command = "nrpe-swap"
assign where host.vars.os == "Linux"
}
apply Service "check_nrpe:dns"{
import "generic-service"
display_name = "Check Nrpe: dns"
check_command = "nrpe-dns"
assign where host.vars.os == "Linux"
}
apply Service "check_nrpe:safe"{
import "generic-service"
display_name = "Check Nrpe: safe"
check_command = "nrpe-safe"
assign where host.vars.os == "Linux"
}
apply Service "check_nrpe:iostat_io"{
import "generic-service"
display_name = "Check Nrpe: iostat_io"
check_command = "nrpe-iostat_io"
vars.disk_device = host.vars.disk_device
assign where host.vars.os == "Linux"
}
apply Service "check_nrpe:iostat_waittime"{
import "generic-service"
display_name = "Check Nrpe: iostat_waittime"
check_command = "nrpe-iostat_waittime"
vars.disk_device = host.vars.disk_device
assign where host.vars.os == "Linux"
}
配置常用window檢測命令
# vim /etc/icinga2/zones.d/global-templates/nrpe_windows.conf
object CheckCommand "windows-nrpe-cpu" {
import "nrpe-common"
vars.nrpe_args = []
vars.nrpe_command = "alias_cpu"
}
object CheckCommand "windows-nrpe-disk" {
import "nrpe-common"
vars.nrpe_args = []
vars.nrpe_command = "alias_disk"
}
object CheckCommand "windows-nrpe-uptime" {
import "nrpe-common"
vars.nrpe_args = []
vars.nrpe_command = "uptime"
}
object CheckCommand "windows-nrpe-mem" {
import "nrpe-common"
vars.nrpe_args = []
vars.nrpe_command = "alias_mem"
}
apply Service "windows_check_nrpe:cpu"{
import "generic-service"
display_name = "Windows-Check Nrpe: CPU"
check_command = "windows-nrpe-cpu"
assign where host.vars.os == "Windows"
}
apply Service "windows_check_nrpe:disk"{
import "generic-service"
display_name = "Windows-Check Nrpe: Disk"
check_command = "windows-nrpe-disk"
assign where host.vars.os == "Windows"
}
apply Service "windows_check_nrpe:uptime"{
import "generic-service"
display_name = "Windows-Check Nrpe: Uptime"
check_command = "windows-nrpe-uptime"
assign where host.vars.os == "Windows"
}
apply Service "windows_check_nrpe:Mem"{
import "generic-service"
display_name = "Windows-Check Nrpe: Mem"
check_command = "windows-nrpe-mem"
assign where host.vars.os == "Windows"
}
重載icinga2
在reload后,master會自動分發配置給兩個子節點,從而做到統一配置,統一分發。
# service icinga2 reload
url: /icinga2-classicui/
默認用戶: icingaadmin
排錯
-
如果配置分布式節點后發現無法監控,或者global-templates配置沒有生效,可以刪除
/var/lib/icinga2/api/zones
下的文件,然后重啟icinga2. -
CHECK_NRPE: Error - Could not complete SSL handshake.: 查看·nrpe.cfg·的
allowed_hosts
配置項是否正確 -
check_iostat_*
沒有數據:OK - I/O stats: Transfers/Sec= Read Requests/Sec= Write Requests/Sec= KBytes Read/Sec= KBytes_Written/Sec=
,sysstat
包沒有安裝。