ganglia监控部署

时间 2019-11-14

标签 ganglia 监控部署繁體版

原文原文链接

1.ganglia组件

ganglia 相比于falcon和zabbix主要在于集群的状态集中显示，能够很便捷的对比各主机的性能状态。php

gmond：至关因而agent端，主要用于收集各node的性能状态
gmetad：至关因而server端，从gmond以poll的方式收集和存储原数据
ganglia-web：至关于一个web页面，读取gmetad存储于rrd中的数据进行web展现

2.ganglia工做模式

多播模式：适用于同一网段，各gmond节点相互发送数据，而后gmetad从自身配置文件中的data_source设置值中依次读取，若是第一个数据源读取失败，则会读取下一处数据源数据。
单播模式：能跨网段工做，各gmond节点将数据统一发送于同一节点，而后gmetad再从中读取数据。

多播模式配置比较简单，gmond节点无需配置直接启动便可运行。html

本文采用单播模式安装node

3.ganglia部署

[root@jump-halo hosts]# ansible kylin --list-host
  hosts (32):
    10.239.1.35
    10.239.1.36
    10.110.80.21
    10.110.80.23
    10.110.80.36
如下略

进程	ip
gmetad	10.239.1.35
ganglia-web	10.239.1.35
gmond	以上全部主机

3.1 gmetad端安装

epel源中均存在rpm包，能够直接yum安装nginx

[root@10-239-1-35 ganglia]# yum install -y ganglia-web.x86_64
[root@10-239-1-35 ganglia]# yum install -y ganglia-gmetad.x86_64 
[root@10-239-1-35 ganglia]# yum install -y nginx  php-fpm
[root@10-239-1-35 ganglia]# pwd
/etc/ganglia
[root@10-239-1-35 ganglia]# vim gmetad.conf
[root@10-239-1-35 nginx]# vim nginx.conf
[root@10-239-1-35 nginx]# vim conf.d/default.conf 
[root@10-239-1-35 nginx]# service php-fpm start
Starting php-fpm:                                          [  OK  ]
[root@10-239-1-35 nginx]# service nginx start
Starting nginx:                                            [  OK  ]
[root@10-239-1-35 nginx]# service gmetad start
Starting GANGLIA gmetad:                                   [  OK  ]
[root@10-239-1-35 nginx]# service gmond start
Starting GANGLIA gmond:                                    [  OK  ]
[root@10-239-1-35 nginx]# chkconfig --add nginx
[root@10-239-1-35 nginx]# chkconfig --add php-fpm
[root@10-239-1-35 nginx]# chkconfig nginx on
[root@10-239-1-35 nginx]# chkconfig php-fpm on
[root@10-239-1-35 nginx]# chkconfig --add gmetad
[root@10-239-1-35 nginx]# chkconfig --add gmond
[root@10-239-1-35 nginx]# chkconfig gmetad on
[root@10-239-1-35 nginx]# chkconfig gmond on
[root@10-239-1-35 nginx]# chkconfig --list | egrep "nginx|php-fpm|gmond|gmetad"
gmetad          0:off   1:off   2:on    3:on    4:on    5:on    6:off
gmond           0:off   1:off   2:on    3:on    4:on    5:on    6:off
nginx           0:off   1:off   2:on    3:on    4:on    5:on    6:off
php-fpm         0:off   1:off   2:on    3:on    4:on    5:on    6:off

3.2 gmetad的配置文件更改

[root@10-239-1-35 ganglia]# grep ^[^#] /etc/ganglia/gmetad.conf 
data_source "bigdata-kylin" localhost    
gridname "kylin"    
setuid_username ganglia
case_sensitive_hostnames 0

data_source "bigdata-kylin" localhost此处设置cluster名称，gmond端务必一致，特别是在多播模式中，依此进行识别，最后的localhost是gmond传输的节点，由于设置gmond端将全部数据传输于gmetad处的的gmond节点，因此能够设置为localhost
gridname "kylin"此处为web端cluster处显示名称，无太多做用，就是为了去掉web中最上端显示的unspecial

3.3 gmetad端web配置文件设置

ganglia-web并不须要设置，此处主要为nginx与php-fpm的配置文件更改，php-fpm配置无需更改web

[root@10-239-1-35 ganglia]# cat /etc/nginx/conf.d/default.conf 
server {
    listen       80 default_server;
    server_name  _;
    root         /usr/share/ganglia;
    index   index.php;
    include /etc/nginx/default.d/*.conf;

    location / {
    }

    location ~ \.php$ {
        fastcgi_pass   127.0.0.1:9000;
        fastcgi_index  index.php;
        fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
        include        fastcgi_params;
        }

    error_page 404 /404.html;
        location = /40x.html {
    }

    error_page 500 502 503 504 /50x.html;
        location = /50x.html {
    }
}

3.4 gmond端安装

[root@jump-halo hosts]# ansible kylin -f 10 -m shell -a "yum install ganglia-gmond.x86_64 -y"
[root@jump-halo hosts]# ansible kylin -f 10 -m copy -a "dest=/etc/ganglia src=/tmp/gmond.conf"
[root@jump-halo hosts]# ansible kylin -f 10 -m shell -a "service gmond start"

须要注意的是，gmetad端的gmond须要优先于其它节点的gmond启动shell

3.5 gmond配置文件

主要更改之处vim

[root@10-239-1-35 ganglia]# grep ^[^#] /etc/ganglia/gmond.conf
以上略
cluster {
  name = "bigdata-kylin"  ##须要与gmetad中定义的datasource名称一致
  owner = "unspecified"
  latlong = "unspecified"
  url = "unspecified" 
}
/* The host section describes attributes of the host, like the location */
host {
  location = "unspecified"
}
/* Feel free to specify as many udp_send_channels as you like.  Gmond
   used to only support having a single channel */
udp_send_channel {
  #mcast_join = 239.27.1.1 ## 此处为多播设置，单播模式要去掉
  host = 10.239.1.35   ##向gmond传送数据的ip，设置
  port = 8649          ## 接收gmond节点的端口
  ttl = 1
}
/* You can specify as many udp_recv_channels as you like as well. */
udp_recv_channel {
  port = 8649       ##gmond自身节点的端口
  bind = 0.0.0.0    ##gmond监听地址，去掉也能够
  retry_bind = true
}
如下略

4.ganglia web页面显示

能够看出，全部节点的状态图均在同一web页面进行展现，在Metric处能够进行监控项的选择对比php-fpm

5.ganglia节点删除

[root@10-239-1-35 bigdata-kylin]# ll
total 224
drwxr-xr-x 2 ganglia ganglia 12288 Aug  8 16:25 10-110-101-105
drwxr-xr-x 2 ganglia ganglia 12288 Aug  8 16:25 10-110-101-106
drwxr-xr-x 2 ganglia ganglia  4096 Aug  8 16:22 10-110-144-168
drwxr-xr-x 2 ganglia ganglia  4096 Aug  8 16:22 10-110-144-169
drwxr-xr-x 2 ganglia ganglia  4096 Aug  8 16:22 10-110-144-170
drwxr-xr-x 2 ganglia ganglia  4096 Aug  8 16:22 10-110-144-171
drwxr-xr-x 2 ganglia ganglia  4096 Aug  8 16:22 10-110-144-172
drwxr-xr-x 2 ganglia ganglia  4096 Aug  8 16:22 10-110-144-173
drwxr-xr-x 2 ganglia ganglia  4096 Aug  8 16:22 10-110-144-174
drwxr-xr-x 2 ganglia ganglia  4096 Aug  8 16:22 10-110-144-175
drwxr-xr-x 2 ganglia ganglia 12288 Aug  8 16:25 10-110-150-199
drwxr-xr-x 2 ganglia ganglia 12288 Aug  8 16:25 10-110-150-234
drwxr-xr-x 2 ganglia ganglia 12288 Aug  8 16:31 10-110-152-21
drwxr-xr-x 2 ganglia ganglia 12288 Aug  8 16:25 10-110-152-25
drwxr-xr-x 2 ganglia ganglia  4096 Aug  8 16:22 10-110-156-191
drwxr-xr-x 2 ganglia ganglia  4096 Aug  8 16:22 10-110-80-21
drwxr-xr-x 2 ganglia ganglia  4096 Aug  8 16:22 10-110-80-23
drwxr-xr-x 2 ganglia ganglia  4096 Aug  8 16:22 10-110-80-36
drwxr-xr-x 2 ganglia ganglia  4096 Aug  8 16:22 10-110-80-56
drwxr-xr-x 2 ganglia ganglia 12288 Aug  8 16:49 10-239-1-104
drwxr-xr-x 2 ganglia ganglia 12288 Aug  8 16:37 10-239-1-105
drwxr-xr-x 2 ganglia ganglia 20480 Aug  8 16:24 10-239-1-35
drwxr-xr-x 2 ganglia ganglia  4096 Aug  8 16:23 10-239-1-50
drwxr-xr-x 2 ganglia ganglia 12288 Aug  8 16:23 10-239-1-51
drwxr-xr-x 2 ganglia ganglia  4096 Aug  8 16:22 10-239-1-64
drwxr-xr-x 2 ganglia ganglia  4096 Aug  8 16:25 10-239-1-75
drwxr-xr-x 2 ganglia ganglia 12288 Aug  8 16:25 10-239-1-76
drwxr-xr-x 2 ganglia ganglia 20480 Aug  8 16:23 __SummaryInfo__
[root@10-239-1-35 bigdata-kylin]# pwd
/var/lib/ganglia/rrds/bigdata-kylin

down掉对应节点的gmond服务，同时须要删除gmetad节点中/var/lib/ganglia/rrds/bigdata-kylin中对应主机的数据目录便可oop

6.其它

在ganglia-web中的grid以后，有source的选择，便可以选择分组，此处的分组，能够在gmetad中进行设置性能

在/etc/ganglia/gmetad.conf中，能够进行以下设置

data_source "bigdata-kylin" localhost
data_souce "hadoop-kylin" 10.239.1.76
data_souce "letv-kylin" localhost:7387
gridname "kylin"