环境：
CentOS release 6.2 (Final)
首先要下载ganglia
http://sourceforge.net/projects/ganglia/files/ganglia%20monitoring%20core/3.6.0/ganglia-3.6.0.tar.gz/download?use_mirror=nchc
目前最新的为ganglia-3.7.2.tar.gz
同时要下载ganglia-web-3.71.tar.gz
下载httd
http://archive.apache.org/dist/httpd/httpd-2.2.23.tar.gz
php-5.4.10.tar.gz http://cn2.php.net/distributions/php-5.4.10.tar.gz
目前最新版本为：php-5.6.13.tar.gz
confuse-2.7 http://www.nongnu.org/confuse/
安装路径规划：软件名称
路径
ganglia安装路径
/usr/local/ganglia
php安装路径
/usr/local/php
apache安装路径
/usr/local/apache2
ganglia-web安装路径
/opt/monitor/ganglia/ganglia-web-3.5.12
rrds数据路径
/opt/monitor/ganglia/rrds
保证yum源可用
yum -y install apr-devel apr-util check-devel cairo-devel pango-devel libxml2-devel rpm-build glib2-devel dbus-devel freetype-devel fontconfig-devel gcc-c++ expat-devel python-devel libXrender-devel pkgconfig
查询是否有包未安装
yum install gcc glibc glibc-common rrdtool rrdtool-devel apr apr-devel expat expat-devel pcre pcre-devel dejavu-lgc-sans-mono-fonts dejavu-sans-mono-fonts
并且安装
pcre：安装目录/usr/local/pcre
tar -xvf pcre-8.33.tar.gz
./configure --prefix=/usr/local/pcre --libdir=/usr/lib64
make
make install
confuse：
tar -xvf confuse-2.7.tar.gz
./configure CFLAGS=-fPIC --disable-nls
make
make install
rrdtool：安装目录/usr/local/rrdtool
tar -xvf rrdtool-1.4.7.tar.gz
./configure --prefix=/usr/local/rrdtool
make
make install
apache：安装目录/usr/local/apache2
tar -xvf httpd-2.2.23.tar.gz
cd srclib
cd apr
./configure --prefix=/usr/local/apr
make
make install
cd ../apr-util
./configure --prefix=/usr/local/apr-util --with-apr=/usr/local/apr
make
make install
cd ../../
./configure --prefix=/usr/local/apache2 --with-apr=/usr/local/apr --with-apr-util=/usr/local/apr-util
make
make install
php安装
tar -xvf php-5.5.30.tar.gz
./configure --prefix=/usr/local/php --with-apxs2=/usr/local/apache2/bin/apxs
编译PHP5.5 make 时出现错误
make: *** [ext/fileinfo/libmagic/apprentice.lo] Error 1
解决办法
这是由于内存小于1G所导致.
在./configure加上选项:
--disable-fileinfo
Disable fileinfo support 禁用 fileinfo
make
make test
make install
php配置
从源码包目录拷贝一份php.ini-development 到php的安装目录
cp php.ini-development /usr/local/php/lib/php.ini
配置apache
cd /usr/local/apache2/conf
vim httpd.conf
查找字符串AddType，并增加
AddType application/x-httpd-php .php .php3 .htm .phtml .php4
查找字符串DirectoryIndex，并修改为 DirectoryIndex index.html index.php
在apache根文件中（/usr/local/apache2/htdocs/）增加一个test.php文件，<?php phpinfo(); ?>,然后访问127.0.0.1/test.php,查看结果。
重启apache,访问，
python安装：tar -xvf Python-2.7.10.tar.xz
./configure --prefix=/usr/local/python --enable-shared
make
make install
下面开始进行ganglia的安装(所有节点都要安装)
tar -xvf ganglia-3.6.0.tar.gz
./configure --prefix=/usr/local/ganglia --with-librrd=/usr/local/rrdtool --with-python=/usr/local/python --with-libpcre=/usr/local/pcre --with-gmetad --enable-gexec --enable-status --sysconfdir=/etc/ganglia
make
make install
安装ganglia-web：仅主节点要装
解压，配置
cd ganglia-web-3.5.12/
cp conf_default.php conf.php
修改：$conf['rrdtool'] = "/usr/bin/rrdtool";-->$conf['rrdtool'] = "/usr/local/rrdtool/bin/rrdtool";
修改：$conf['gweb_confdir'] = "/var/lib/ganglia-web";-->$conf['gweb_confdir'] = "/opt/monitor/ganglia/ganglia-web-3.5.12";
修改：$conf['gmetad_root'] = "/var/lib/ganglia";-->$conf['gmetad_root'] = "/opt/monitor/ganglia";
修改header.php
<?php
session_start();
ini_set('date.timezone','PRC'); --修改时区为本地时区(新加这一行)
if (isset($_GET['date_only'])) {
$d = date("r");
echo $d;
exit(0);
}
配置临时目录
cd /opt/monitor/ganglia/ganglia-web-3.5.12/dwoo/
mkdir cache
chmod 777 cache
mkdir compiled
chmod 777 compiled
使用apache发布ganglia-web（主节点）
vi /usr/local/apache2/conf/httpd.conf
启动httpd服务
?
1
/usr/local/apache2/bin/apachectl restart
配置ganglia
配置gmetad(主节点配置)
cd /opt/monitor/ganglia-3.6.0
cp gmetad/gmetad.init /etc/init.d/gmetad
注：这里编译ganglia的时候加了参数sysconfdir，所以这里不需要进行如下步骤
cp gmetad/gmetad.conf /usr/local/ganglia/etc/
但是如果没有加上系统配置参数，则要执行上一步，然后
gmetad -t | tee /usr/loca/ganglia/etc/gmetad.conf这个是为了生成默认配置文件
vim /etc/init.d/gmetad
修改：GMETAD=/usr/sbin/gmetad-->GMETAD=/usr/local/ganglia/sbin/gmetad
vim etc/ganglia/gmetad.conf
修改：data_source "my cluster" localhost --> data_source "solrCloud" localhost 192.168.36.82:8649 192.168.36.81:8649 192.168.36.84:8649 192.168.36.80:8649
放开xml_port这里的端口不能和data_source的端口冲突
放开interactive_port
vim opt/monitor/ganglia/ganglia-web-3.5.12/conf.php
这个端口和interactive port端口要一致
修改rrd_rootdir的路径
# rrd_rootdir "/some/other/place"-->rrd_rootdir "/opt/monitor/ganglia/rrds"
查看case_sensitive_hostnames 0是否已经放开，并且值为0
修改# gridname "MyGrid"-->gridname "solrCloud"
修改rrds数据目录所有者
chown -R nobody:nobody /etc/monitor/ganglia/rrds
启动gmetad服务，并设为开机自动运行
service gmetad restart
Starting GANGLIA gmetad: [ OK ]
chkconfig --add gmetad
gmond配置（所有节点）
cd /opt/monitor/ganglia-3.6.0
cp gmond/gmond.init /etc/init.d/gmond
vim /etc/init.d/gmond
修噶:#GMOND=/usr/sbin/gmond-->GMOND=/usr/local/ganglia/sbin/gmond
./gmond/gmond -t > /etc/ganglia/gmond.conf（生成默认的配置文件）
或者gmond -t | tee /etc/ganglia/gmond.conf # generate initial gmond config（生成默认的配置文件）
vim /etc/ganglia/gmond.conf
参考如下进行修改
gmetad
注：这个数据源后面的ip的是配置的server的ip
data_source "Namenode" 192.168.1.28:8653
data_source "Datanode" 192.168.1.27:8649
data_source "Portal" 192.168.1.43:8650
data_source "Collector" 192.168.1.35:8651
data_source "DB" 192.168.1.51:8652
gridname "Hadoop"
rrd_rootdir "/opt/modules/ganglia/html/rrds"
#配置rrd数据保存文件的路径，给web界面用的，这个是固定的，最好放在web文件夹下，并赋予正确的权限
case_sensitive_hostnames 0
数据来源有5个，这5个分别是每个组的组长，相当于一道杠。但是组长是不需要配置gmetad的，除非你要做多级组播收集数据。每个组长只需要分配不同的端口号就可以了。你可能会问，IP不一样，端口一样不行吗？不行，因为这个IP是单播IP，相当于一个路由指向，而Ganglia实际的数据传输是在多播IP上进行的，而多播IP只有一个。在客户端配置，如果你需要多级gmetad，可以配多个多播IP。
客户端配置就比较复杂一些了。我只贴上需要修改的部分，其他都是默认就可以了
注：这里是多播的模式配置
gmond
cluster {
name = "Portal"
#对应gmetad中的Portal，名称一定要写对。
owner = "unspecified"---->nobody
latlong = "unspecified"
url = "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 {
#bind_hostname = yes # Highly recommended, soon to be default.
# This option tells gmond to use a source address
# that resolves to the machine's hostname. Without
# this, the metrics may appear to come from any
# interface and the DNS names associated with
# those IPs will be used to create the RRDs.
mcast_join = 239.2.11.71
port = 8650
#gmetad中的Portal所分配的端口号。
ttl = 1
}
/* You can specify as many udp_recv_channels as you like as well. */
udp_recv_channel {
mcast_join = 239.2.11.71
port = 8650
bind = 239.2.11.71
}
/* You can specify as many tcp_accept_channels as you like to share
an xml description of the state of the cluster */
tcp_accept_channel {
port = 8650
}
红色部分就是Portal小组的端口，从gmetad.conf中可以看到，Portal小组属于8650端口，那么相应的在gmond中，也要将udp和tcp端口写为8650。
如果是另外一个组的，就写上在gmetad中配置的那个端口。当然，你可以把这个端口号想像为小组的代号。可能更好理解一些。
再加上另外一个组的成员gmond就更容易理解了
cluster {
name = "DB"
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 {
#bind_hostname = yes # Highly recommended, soon to be default.
# This option tells gmond to use a source address
# that resolves to the machine's hostname. Without
# this, the metrics may appear to come from any
# interface and the DNS names associated with
# those IPs will be used to create the RRDs.
mcast_join = 239.2.11.71
port = 8652
ttl = 1
}
/* You can specify as many udp_recv_channels as you like as well. */
udp_recv_channel {
mcast_join = 239.2.11.71
port = 8652
bind = 239.2.11.71
}
/* You can specify as many tcp_accept_channels as you like to share
an xml description of the state of the cluster */
tcp_accept_channel {
port = 8652
}
红色对红色，蓝色对蓝色。一目了然。
复制python module到ganglia部署目录
mkdir /usr/local/ganglia/lib64/ganglia/python_modules
cp ./gmond/python_modules/*/*.py /usr/local/ganglia/lib64/ganglia/python_modules
安装程序ganglia-3.6.0默认提供了一些python module的配置文件,只需要部署到 /usr/local/ganglia/etc/conf.d目录下面即可生效
cp ./gmond/python_modules/conf.d/*.pyconf /etc/ganglia/conf.d
service gmond restart
chkconfig --add gmond
下面贴上单播的模式配置：
单播模式设置：
在每个节点上都要配置/etc/ganglia/gmond.conf
globals {
daemonize = yes
setuid = yes
user = root /*运行Ganglia的用户(一般可能是nobody)*/
debug_level = 0
max_udp_msg_len = 1472
mute = no
deaf = no
host_dmax = 120 /*secs */
cleanup_threshold = 300 /*secs */
gexec = no
send_metadata_interval = 15 /*发送数据的时间间隔*/
}
cluster {
name = "hadoop_hf" /*集群名称*/
owner = "root" /*运行Ganglia的用户*/
latlong = "unspecified"
url = "unspecified"
}
udp_send_channel {
# mcast_join = 239.2.11.71 /*注释掉组播*/
host = 192.168.228.172 /*发送给安装gmetad的机器*/
port = 8649
ttl = 1
}
/* You can specify as many udp_recv_channels as you like as well. */
udp_recv_channel {
# mcast_join = 239.2.11.71 /*注释掉组播*/
port = 8649
# bind = 239.2.11.71 /*注释掉组播*/
}
测试：
# 命令行打印当前活动client
gstat -a
常见错误：
err2:
There was an error collecting ganglia data (127.0.0.1:8652): fsockopen error: Connection refused
解决代码：
# mkdir -p /var/lib/ganglia/rrds
# chown nobody:nobody /var/lib/ganglia/rrds
注：注：当有信息出错时，可以查看log文件
tail –n 100 /var/log/message //查看对进的100行
echo /dev/null > /var/log/message //清空message
参考文献：
http://my.oschina.net/duangr/blog/181585
http://my.oschina.net/duangr/blog/181189
http://blog.csdn.net/kafeiflynn/article/details/6634380
http://www.cnblogs.com/Richardzhu/p/3481996.html
欢迎关注我的公众号
来源：oschina
链接：https://my.oschina.net/u/1262062/blog/3213294