一、高可用集群之heartbeat基于crm进行资源管理
1、集群的工作模型：
A/P：两个节点，工作与主备模型
N-M N>M，N个节点，M个服务
N-N：N个节点，N个服务
A/A：双主模型：
2、资源转移的方式
rgmanager：failover domain priority
pacemaker：
资源黏性：
资源约束（三种类型）：
位置约束：资源更倾向于那个节点上
inf：无穷大
n:
-n:
-inf:负无穷
排列约束：资源运行在同一节点的倾向性
inf：
-inf：
顺序约束：资源的启动次序及关闭次序
3、如何让web service中的三个资源：VIP、httpd和filesystem运行于同一节点上
1.排列约束
2.资源组（resource group）
4、如果节点不在是集群节点成员时，如何处理运行于当前节点的资源
stopped：停止
ignore：忽略
freeze：不连接新的请求
suicide：将服务器kill
5、一个资源刚配置完成时，是否启动
target-role？
6、RA类型
heartbeat legacy
LSB
OCF
STONITH
7、资源类型
primitive，native：主资源，只能运行于一个节点
group：组资源
clone：克隆资源
总克隆数，每个节点最多可运行的克隆数
stonith cluster filesystem
master/salve：主从资源
8、分布式锁：
/usr/lib64/heartbeat
hearsources2cib.py
9、图形化配置
ha.cf
crm on
/usr/lib64/heartbeat/ha_propagate 将配置文件传送到别的节点
10、安装gui
heartbeat v2使用crm作为ijiqun资源管理器：需要在ha.cf中添加
crm on
crm通过mgmtd集成监听5560/tcp
需要启动hb_gui的主机为hacluster用户添加密码，使用hb_gui启动
with quorum：拥有法定票数
without quorum ：不拥有法定票数
11、定义高可用的web service
VIP
httpd
from
to：以它为基础
web service
VIP
httpd
NFS
注意haresources与crm不兼容，不被crm所读取
二、配置
1、ha.cf
[root@snn heartbeat]# vim /etc/ha.d/ha.cf
mcast eth0 225.0.100.19 694 1 0
crm on
[root@snn heartbeat]# /usr/lib64/heartbeat/ha_propagate
Propagating HA configuration files to node datanode4.abc.com.
ha.cf 100% 10KB 10.4KB/s 00:00
authkeys 100% 694 0.7KB/s 00:00
Setting HA startup configuration on node datanode4.abc.com.
2、注意haresources与crm不兼容，不被crm所读取
[root@snn heartbeat]# mv /etc/ha.d/haresources /root
底下mv是datanode4的主机
[root@datanode4 ha.d]# mv haresources /root/
[root@snn heartbeat]# service heartbeat start
logd is already running
Starting High-Availability services:
Done.
[root@snn heartbeat]# ssh datanode4 'service heartbeat start'
logd is already running
Starting High-Availability services:
Done.
3、查看日志
[root@snn heartbeat]# tail -f /var/log/messages
Jun 19 16:00:29 snn crmd: [2223]: notice: populate_cib_nodes: Node: datanode4.abc.com (uuid: 0862d824-047e-4826-9e26-21a7603f53c8)
Jun 19 16:00:30 snn crmd: [2223]: notice: populate_cib_nodes: Node: snn.abc.com (uuid: 6009ca6a-56eb-4d35-872e-3b8dc0fc9851)
Jun 19 16:00:30 snn crmd: [2223]: info: do_ha_control: Connected to Heartbeat
Jun 19 16:00:30 snn crmd: [2223]: info: do_ccm_control: CCM connection established... waiting for first callback
Jun 19 16:00:30 snn crmd: [2223]: info: do_started: Delaying start, CCM (0000000000100000) not connected
Jun 19 16:00:30 snn crmd: [2223]: info: crmd_init: Starting crmd's mainloop
Jun 19 16:00:30 snn crmd: [2223]: notice: crmd_client_status_callback: Status update: Client snn.abc.com/crmd now has status [online]
Jun 19 16:00:30 snn crmd: [2223]: notice: crmd_client_status_callback: Status update: Client snn.abc.com/crmd now has status [online]
Jun 19 16:00:30 snn crmd: [2223]: notice: crmd_client_status_callback: Status update: Client datanode4.abc.com/crmd now has status [online]
Jun 19 16:00:30 snn cib: [2219]: info: mem_handle_event: Got an event OC_EV_MS_NEW_MEMBERSHIP from ccm
Jun 19 16:00:30 snn cib: [2219]: info: mem_handle_event: instance=5, nodes=2, new=2, lost=0, n_idx=0, new_idx=0, old_idx=4
Jun 19 16:00:30 snn cib: [2219]: info: cib_ccm_msg_callback: PEER: datanode4.abc.com
Jun 19 16:00:30 snn cib: [2219]: info: cib_ccm_msg_callback: PEER: snn.abc.com
Jun 19 16:00:31 snn crmd: [2223]: info: do_started: Delaying start, CCM (0000000000100000) not connected
Jun 19 16:00:31 snn crmd: [2223]: info: mem_handle_event: Got an event OC_EV_MS_NEW_MEMBERSHIP from ccm
Jun 19 16:00:31 snn crmd: [2223]: info: mem_handle_event: instance=5, nodes=2, new=2, lost=0, n_idx=0, new_idx=0, old_idx=4
Jun 19 16:00:31 snn crmd: [2223]: info: crmd_ccm_msg_callback: Quorum (re)attained after event=NEW MEMBERSHIP (id=5)
Jun 19 16:00:31 snn crmd: [2223]: info: ccm_event_detail: NEW MEMBERSHIP: trans=5, nodes=2, new=2, lost=0 n_idx=0, new_idx=0, old_idx=4
Jun 19 16:00:31 snn crmd: [2223]: info: ccm_event_detail: #011CURRENT: datanode4.abc.com [nodeid=0, born=3]
Jun 19 16:00:31 snn crmd: [2223]: info: ccm_event_detail: #011CURRENT: snn.abc.com [nodeid=1, born=5]
Jun 19 16:00:31 snn crmd: [2223]: info: ccm_event_detail: #011NEW: datanode4.abc.com [nodeid=0, born=3]
Jun 19 16:00:31 snn crmd: [2223]: info: ccm_event_detail: #011NEW: snn.abc.com [nodeid=1, born=5]
Jun 19 16:00:31 snn crmd: [2223]: info: do_started: The local CRM is operational
Jun 19 16:00:31 snn crmd: [2223]: info: do_state_transition: State transition S_STARTING -> S_PENDING [ input=I_PENDING cause=C_CCM_CALLBACK origin=do_started ]
4、查看集群监控状态
//如果想它只显示一次使用crm_mon --one-shot
[root@snn heartbeat]# crm_mon
Refresh in 6s...
============
Last updated: Fri Jun 19 16:11:34 2015
Current DC: snn.abc.com (6009ca6a-56eb-4d35-872e-3b8dc0fc9851)
2 Nodes configured.
0 Resources configured.
============
Node: datanode4.abc.com (0862d824-047e-4826-9e26-21a7603f53c8): online
Node: snn.abc.com (6009ca6a-56eb-4d35-872e-3b8dc0fc9851): online
4、crm的命令工具
[root@snn heartbeat]# crm_sh
/usr/sbin/crm_sh:31: DeprecationWarning: The popen2 module is deprecated. Use the subprocess module.
from popen2 import Popen3
crm # help
Usage: crm (nodes|config|resources)
crm # nodes
crm nodes # help
Usage: nodes (status|list)
crm nodes # list
<node id="0862d824-047e-4826-9e26-21a7603f53c8" uname="datanode4.abc.com" type="normal"/>
<node id="6009ca6a-56eb-4d35-872e-3b8dc0fc9851" uname="snn.abc.com" type="normal"/>
crm nodes #
5、安装heartbeat的时候自动创建一个用户hacluster，但没有密码，需要创建
[root@snn heartbeat]# cat /etc/passwd ｜grep hacluster
hacluster:x:498:498:heartbeat user:/var/lib/heartbeat/cores/hacluster:/sbin/nologin
[root@snn heartbeat]# passwd hacluster
更改用户 hacluster 的密码 。
新的 密码：
无效的密码： WAY 过短
无效的密码： 过于简单
重新输入新的 密码：
passwd： 所有的身份验证令牌已经成功更新。
6、直接运行hb_gui
[root@snn ~]# hb_gui
Traceback (most recent call last):
File "/usr/bin/hb_gui", line 41, in <module>
import gtk, gtk.glade, gobject
File "/usr/lib64/python2.6/site-packages/gtk-2.0/gtk/__init__.py", line 64, in <module>
_init()
File "/usr/lib64/python2.6/site-packages/gtk-2.0/gtk/__init__.py", line 52, in _init
_gtk.init_check()
RuntimeError: could not open display
以上有错误提示
在客户端下载安装Xmanager即可
在重执行命令
三、ha_gui定义
1、定义主资源名称
2、继继定义主资源
3、让两个资源运行同一个节点，方法有两种：（1）定义排列约束，（2）定义资源组
（1）定义排列约束
4、让snn节点成为备的
四、定义组的方式
web server:
vip:192.168.1.8
httpd
nfs:/192.168.1.4:/web/htdocs挂在到/var/www/html
1、删除原来主资源
2、定义群主源
3、httpd无法启动，查看日志如下
从日志 来看，nfs正常挂在到4这主机上，但httpd先启动后又关闭，奇怪了
4、来到datanode4这台机子，单独启动httpd看看，没有成功
[root@datanode4 ~]# /etc/init.d/httpd restart
停止 httpd： [失败]
正在启动 httpd：Syntax error on line 292 of /etc/httpd/conf/httpd.conf:
5、查看SElinux状态，吓了一跳，问题出现在这里
[root@datanode4 conf]# getenforce
Enforcing
[root@datanode4 conf]# setenforce 0
[root@datanode4 conf]# getenforce
Permissive
把配置文件改成disabled
[root@datanode4 conf]# vim /etc/selinux/config
SELINUX=disabled
6、单独在启动httpd看看
[root@datanode4 conf]# /etc/init.d/httpd start
正在启动 httpd： [确定]
[root@datanode4 conf]# /etc/init.d/httpd stop
停止 httpd：
7、再回到snn输入hb_gui看看，之前是webserivce，这不影响名称是可以随便定义，我之前删了，就重新建一资源为了好识别，就定义了httpd
五、验证
1、nfs的index.html内容
[root@datanode ~]# cat /web/htdocs/index.html
<h1>datanode.abc.com</h1>
[root@datanode ~]# ifconfig eth0
eth0 Link encap:Ethernet HWaddr 00:0C:29:50:AC:6E
inet addr:192.168.1.4 Bcast:192.168.1.255 Mask:255.255.255.0
inet6 addr: fe80::20c:29ff:fe50:ac6e/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:83505 errors:0 dropped:0 overruns:0 frame:0
TX packets:2037 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:7403212 (7.0 MiB) TX bytes:228350 (222.9 KiB)
2、datanode4的主机的vip地址，如果单纯输入ifocnfig，不能显示出来的，它没有利用别名来定义，所以要用的ip addr show
[root@datanode4 html]# ifconfig eth0
eth0 Link encap:Ethernet HWaddr 00:0C:29:E1:2F:66
inet addr:192.168.1.6 Bcast:192.168.1.255 Mask:255.255.255.0
inet6 addr: fe80::20c:29ff:fee1:2f66/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:147365 errors:0 dropped:0 overruns:0 frame:0
TX packets:66651 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:20284443 (19.3 MiB) TX bytes:14571080 (13.8 MiB)
[root@datanode4 html]# ip addr show
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:e1:2f:66 brd ff:ff:ff:ff:ff:ff
inet 192.168.1.6/24 brd 192.168.1.255 scope global eth0
inet 192.168.1.8/24 brd 192.168.1.255 scope global secondary eth0 //显示vip地址
inet6 fe80::20c:29ff:fee1:2f66/64 scope link
valid_lft forever preferred_lft forever
3、在浏览器输入，注意这里输入的是vip地址
4、如果datanode4成为备用
到snn主机上看，转移成功
[root@snn ~]# ip addr show
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:b1:89:48 brd ff:ff:ff:ff:ff:ff
inet 192.168.1.5/24 brd 192.168.1.255 scope global eth0
inet 192.168.1.8/24 brd 192.168.1.255 scope global secondary eth0
inet6 fe80::20c:29ff:feb1:8948/64 scope link
valid_lft forever preferred_lft forever
涮新浏览器，还是原来的内容
本文出自 “散人” 博客，请务必保留此出处http://zouqingyun.blog.51cto.com/782246/1663710