Centos 6.5 Zabbix HA

Network Topology:

zabbix mysql server

1. Make sure you have successfully set up DNS resolution and NTP time synchronization for both your Linux Cluster nodes.

vi /etc/sysconfig/network
设定主机名为short name, not FQDN

here, I use /etc/hosts to simplify

vi /etc/hosts
# eth0 network for production
192.168.1.13    mysql01
192.168.1.23    mysql02

# for dedicated heartbeat network to avoid problems
# eth1 network
10.10.12.10    mysql01p
10.10.12.20    mysql02p

2. wget http://download.opensuse.org/repositories/network:/ha-clustering:/Stable/CentOS_CentOS-6/network:ha-clustering:Stable.repo -O /etc/yum.repos.d/ha.repo

3. on two nodes
yum -y install pacemaker cman crmsh

4. on two nodes
vi /etc/sysconfig/cman
CMAN_QUORUM_TIMEOUT=0

5. on two nodes
vi /etc/cluster/cluster.conf

<?xml version="1.0"?>
<cluster config_version="1" name="mycluster">
<logging debug="off"/>
<clusternodes>
  <clusternode name="mysql01p" nodeid="1">
    <fence>
      <method name="pcmk-redirect">
         <device name="pcmk" port="mysql01p"/>
      </method>
    </fence>
</clusternode>
<clusternode name="mysql02p" nodeid="2">
   <fence>
      <method name="pcmk-redirect">
         <device name="pcmk" port="mysql02p"/>
       </method>
     </fence>
 </clusternode>
</clusternodes>
<fencedevices>
  <fencedevice name="pcmk" agent="fence_pcmk"/>
</fencedevices>
</cluster>

6. on two nodes
ccs_config_validate

7. on node1 then node2
service cman start

check nodes status with cman_tool nodes

8. on node1 then node2
service pacemaker start
chkconfig cman on
chkconfig pacemaker on

9. for a two nodes cluster, run from one node
crm configure property stonith-enabled="false"
crm configure property no-quorum-policy="ignore"
#crm configure rsc_defaults resource-stickiness="100"

10. crm_verify -LV
crm status

11. on two nodes
yum -y install mysql mysql-server mysql-devel
chkconfig mysqld off

12. on two nodes
mkdir -p /new/mysql; chown -R mysql.mysql /new/mysql

vi /etc/my.cnf
[mysqld]
datadir=/new/mysql
socket=/new/mysql/mysql.sock
default-storage-engine=INNODB

[client]
socket=/new/mysql/mysql.sock

13. on two nodes
Install DRBD 8.3
rpm -ivh http://www.elrepo.org/elrepo-release-6-5.el6.elrepo.noarch.rpm
yum -y install drbd83-utils kmod-drbd83
chkconfig drbd off

14. on two nodes
create a new 100M disk just for testing purpose

vi /etc/drbd.d/zabbixdata.res
resource zabbixdata {
device    /dev/drbd0;
disk      /dev/sdb;
meta-disk internal;
syncer {
  verify-alg sha1;
}  
on mysql01.test.com {
  address   192.168.1.13:7789;
}
on mysql02.test.com {
  address   192.168.1.23:7789;
}
}

15. on mysql01
drbdadm create-md zabbixdata
modprobe drbd
drbdadm up zabbixdata

16. on mysql02
drbdadm create-md zabbixdata
modprobe drbd
drbdadm up zabbixdata

17. make mysql01 primary
drbdadm -- --overwrite-data-of-peer primary zabbixdata

cat /proc/drbd
Note: you should wait until sync finished

18. on mysql01
mkfs.ext4 /dev/drbd0

19. #crm(live)configure
primitive clusterip ocf:heartbeat:IPaddr2 params ip="192.168.1.52" cidr_netmask="24" nic="eth0"
commit
run crm status to check

primitive zabbixdata ocf:linbit:drbd params drbd_resource="zabbixdata"
ms zabbixdataclone zabbixdata meta master-max="1" master-node-max="1" clone-max="2" clone-node-max="1" notify="true"
primitive zabbixfs ocf:heartbeat:Filesystem params device="/dev/drbd/by-res/zabbixdata" directory="/new/mysql" fstype="ext4"

group myzabbix zabbixfs clusterip
colocation myzabbix_on_drbd inf: myzabbix zabbixdataclone:Master
order myzabbix_after_drbd inf: zabbixdataclone:promote myzabbix:start

op_defaults timeout=240s

commit

20. ensure resource is on mysql01

on mysql01
run mount to see if /dev/drbd0 --> /new/mysql

service mysqld start

Create zabbix database and user on MySQL.

mysql -u root
mysql> create database zabbix character set utf8;
mysql> grant all privileges on zabbix.* to zabbix@'%' identified by 'zabbix';
mysql> flush privileges;
mysql> exit

Import initial schema and data.
copy zabbix server /usr/share/doc/zabbix-server-mysql-2.2.1/create/* to local
mysql -uroot zabbix < schema.sql
mysql -uroot zabbix < images.sql
mysql -uroot zabbix < data.sql

service mysqld stop

21. #crm(live)configure
primitive mysqld lsb:mysqld
edit myzabbix
like this -- group myzabbix zabbixfs clusterip mysqld
commit

22. put mysql01 to standby
crm node standby mysql01p

23. on mysql02 to check mount and mysql -uroot

24. put mysql01 to online
crm node online mysql01p

zabbix server

1. Make sure you have successfully set up DNS resolution and NTP time synchronization for both your Linux Cluster nodes.

vi /etc/sysconfig/network
设定主机名为short name, not FQDN

here, I use /etc/hosts to simplify

vi /etc/hosts
# eth0 network for production
192.168.1.14    zabbixserver01
192.168.1.24    zabbixserver02

# for dedicated heartbeat network to avoid problems
# eth1 network
10.10.11.10    zabbixserver01p
10.10.11.20   zabbixserver02p

2. wget http://download.opensuse.org/repositories/network:/ha-clustering:/Stable/CentOS_CentOS-6/network:ha-clustering:Stable.repo -O /etc/yum.repos.d/ha.repo

3. on two nodes
yum -y install pacemaker cman crmsh

4. on two nodes
vi /etc/sysconfig/cman
CMAN_QUORUM_TIMEOUT=0

5. on two nodes
vi /etc/cluster/cluster.conf

<?xml version="1.0"?>
<cluster config_version="1" name="mycluster">
<logging debug="off"/>
<clusternodes>
  <clusternode name="zabbixserver01p" nodeid="1">
    <fence>
      <method name="pcmk-redirect">
         <device name="pcmk" port="zabbixserver01p"/>
      </method>
    </fence>
</clusternode>
<clusternode name="zabbixserver02p" nodeid="2">
   <fence>
      <method name="pcmk-redirect">
         <device name="pcmk" port="zabbixserver02p"/>
       </method>
     </fence>
 </clusternode>
</clusternodes>
<fencedevices>
  <fencedevice name="pcmk" agent="fence_pcmk"/>
</fencedevices>
</cluster>

6. on two nodes
ccs_config_validate

7. on two nodes
service cman start

check nodes status with cman_tool nodes

8. on two nodes
service pacemaker start
chkconfig cman on
chkconfig pacemaker on

9. for a two nodes cluster, run from one node
crm configure property stonith-enabled="false"
crm configure property no-quorum-policy="ignore"
#crm configure rsc_defaults resource-stickiness="100"

10. crm_verify -LV
crm status

11. on two nodes
yum -y install zlib-devel glibc-devel libcurl-devel OpenIPMI-devel libssh2-devel net-snmp-devel openldap-devel
rpm -ivh http://mirrors.sohu.com/fedora-epel/6Server/x86_64/epel-release-6-8.noarch.rpm
yum -y install fping iksemel-devel

12. on two nodes
rpm -ivh http://repo.zabbix.com/zabbix/2.2/rhel/6/x86_64/zabbix-release-2.2-1.el6.noarch.rpm
yum -y install zabbix-server-mysql
chkconfig zabbix-server off

vi /etc/zabbix/zabbix_server.conf
DBHost=192.168.1.52 #mysql server cluster ip
DBName=zabbix
DBUser=zabbix
DBPassword=zabbix
DBSocket=/new/mysql/mysql.sock
SourceIP=192.168.1.51
ListenIP=192.168.1.51

13.#crm(live)configure
primitive clusterip ocf:heartbeat:IPaddr2 params ip="192.168.1.51" cidr_netmask="24" nic="eth0"
commit
run crm status to check

14. ensure resource is on zabbixserver01
on zabbixserver01

service zabbix-server start
check /var/log/zabbix/zabbix_server.log
service zabbix-server stop

15. put zabbixserver01 to standby
crm node standby zabbixserver01p

16. on zabbixserver02
service zabbix-server start
check /var/log/zabbix/zabbix_server.log
service zabbix-server stop

17. put zabbixserver01 to online
crm node online zabbixserver01p

18. #crm(live)configure
primitive myzabbixserver lsb:zabbix-server
group myzabbix clusterip myzabbixserver
commit

19. for Zabbix agent -- on zabbix server

yum -y install zabbix-agent
chkconfig zabbix-agent off

20. for zabbix agent configuration

vi /etc/zabbix/zabbix_agentd.conf

#Server=[zabbix server ip] --This is for zabbix agent Passive check, port 10050
#ServerActive=[zabbix server ip] --This is for zabbix agent(active) Active check, port 10051
#Hostname=[ Hostname of client system ]
#hostname value set on the agent side should exactly match the ”Host name” configured for the host in the web frontend

Server=192.168.1.51
ServerActive=192.168.1.51
Hostname=Zabbix server

21. ensure resource is on zabbixserver01
on zabbixserver01

service zabbix-agent start
check /var/log/zabbix/zabbix_agent.log
service zabbix-agent stop

22. put zabbixserver01 to standby
crm node standby zabbixserver01p

23. on zabbixserver02
service zabbix-agent start
check /var/log/zabbix/zabbix_agent.log
service zabbix-agent stop

24. put zabbixserver01 to online
crm node online zabbixserver01p

25. #crm(live)configure
primitive zabbixagent lsb:zabbix-agent
edit myzabbix
like this -- group myzabbix clusterip myzabbixserver zabbixagent
commit

zabbix web

1. Make sure you have successfully set up DNS resolution and NTP time synchronization for both your Linux Cluster nodes.

vi /etc/sysconfig/network
设定主机名为short name, not FQDN

here, I use /etc/hosts to simplify

vi /etc/hosts
# eth0 network for production
192.168.1.15    zabbixweb01
192.168.1.25    zabbixweb02

# for dedicated heartbeat network to avoid problems
# eth1 network
10.10.10.10    zabbixweb01p
10.10.10.20    zabbixweb02p

2. wget http://download.opensuse.org/repositories/network:/ha-clustering:/Stable/CentOS_CentOS-6/network:ha-clustering:Stable.repo -O /etc/yum.repos.d/ha.repo

3. on two nodes
yum -y install pacemaker cman crmsh

4. on two nodes
vi /etc/sysconfig/cman
CMAN_QUORUM_TIMEOUT=0

5. on two nodes
vi /etc/cluster/cluster.conf

<?xml version="1.0"?>
<cluster config_version="1" name="mycluster">
<logging debug="off"/>
<clusternodes>
  <clusternode name="zabbixweb01p" nodeid="1">
    <fence>
      <method name="pcmk-redirect">
         <device name="pcmk" port="zabbixweb01p"/>
      </method>
    </fence>
</clusternode>
<clusternode name="zabbixweb02p" nodeid="2">
   <fence>
      <method name="pcmk-redirect">
         <device name="pcmk" port="zabbixweb02p"/>
       </method>
     </fence>
 </clusternode>
</clusternodes>
<fencedevices>
  <fencedevice name="pcmk" agent="fence_pcmk"/>
</fencedevices>
</cluster>

6. on two nodes
ccs_config_validate

7. on two nodes
service cman start

check nodes status with cman_tool nodes

8. on two nodes
service pacemaker start
chkconfig cman on
chkconfig pacemaker on

9. for a two nodes cluster, run from one node
crm configure property stonith-enabled="false"
crm configure property no-quorum-policy="ignore"
#crm configure rsc_defaults resource-stickiness="100"

10. crm_verify -LV
crm status

11. on two nodes
yum -y install httpd httpd-devel php php-cli php-common php-devel php-pear php-gd php-bcmath php-mbstring php-mysql php-xml wget
chkconfig httpd off

12. on two nodes
rpm -ivh http://repo.zabbix.com/zabbix/2.2/rhel/6/x86_64/zabbix-release-2.2-1.el6.noarch.rpm
yum -y install zabbix-web-mysql

13. on two nodes
Apache configuration file for Zabbix frontend is located in /etc/httpd/conf.d/zabbix.conf. Some PHP settings are already configured.

php_value max_execution_time 300
php_value memory_limit 128M
php_value post_max_size 16M
php_value upload_max_filesize 2M
php_value max_input_time 300
php_value date.timezone Asia/Shanghai

14. on two nodes
vi /etc/httpd/conf/httpd.conf

<Location /server-status>
SetHandler server-status
Order deny,allow
Deny from all
Allow from 127.0.0.1 192.168.1.0/24
</Location>

15. on zabbixweb01
service httpd start
run http://192.168.1.15/zabbix to finish zabbix setup, put 192.168.1.52 as mysql database address
service httpd stop

16. on zabbixweb02
service httpd start
run http://192.168.1.25/zabbix to finish zabbix setup, put 192.168.1.52 as mysql database address
service httpd stop

17. #crm(live)configure
op_defaults timeout=240s
primitive clusterip ocf:heartbeat:IPaddr2 params ip="192.168.1.50" cidr_netmask="24" nic="eth0"
commit
run crm status to check

primitive website ocf:heartbeat:apache params configfile="/etc/httpd/conf/httpd.conf"
group myweb clusterip website
commit

Note:

1. 在使用专属心跳网卡eth1建立群集后，把某个节点置于standby/online时，必须附加节点名字：

crm node standby node_name

crm node online node_name

2. 如果机器down了，重装系统，然后按步骤1-8即可重新加入群集。