drbd+nfs+keepalived

写的很详细 理论知识：

https://www.cnblogs.com/kevingrace/p/5740940.html

写的很详细 负载：

https://www.cnblogs.com/kevingrace/p/5740953.html

实际操作如下：

1、修改主机名

11.11.11.2  主服务器     主机名：Primary
11.11.11.3  备服务器     主机名：Secondary
11.11.11.4  VIP

******主要操作上边两台设备，下边两台用于辅助测试******

11.11.11.8  web服务器   IIS  (用于挂载测试)
11.11.11.9  nginx代理    设置缓存   主机名LB

2、两台机器的防火墙要相互允许访问。最好是关闭selinux和iptables防火墙（两台机器同样操作）

[root@Primary ~]# setenforce 0        //临时性关闭；永久关闭的话，需要修改/etc/sysconfig/selinux的SELINUX为disabled
[root@Primary ~]# /etc/init.d/iptables stop

3、设置hosts文件（两台机器同样操作）

[root@primary ~]# cat /etc/hosts
**********
11.11.11.2 primary
11.11.11.3 secondary

4、两台机器同步时间

[root@Primary ~]# yum install -y netpdate
[root@Primary ~]# ntpdate -u ntp1.aliyun.com

5、DRBD的安装（两台机器上同样操作）

rpm -ivh https://www.elrepo.org/elrepo-release-7.0-3.el7.elrepo.noarch.rpm
yum list drbd*
yum install -y drbd84-utils kmod-drbd84

加载模块：
[root@Primary ~]# modprobe drbd
查看模块是否已加上
[root@Primary ~]# lsmod |grep drbd
drbd                  332493  0

　6、DRBD配置（两台机器上同样操作）

[root@primary ~]# cat /etc/drbd.conf
# You can find an example in  /usr/share/doc/drbd.../drbd.conf.example

include "drbd.d/global_common.conf";      //这是主要的两个配置文件，这个定义软件策略
include "drbd.d/*.res";                   //定义使用磁盘

[root@primary ~]# cp /etc/drbd.d/global_common.conf /etc/drbd.d/global_common.conf.bak

[root@primary ~]# cat  /etc/drbd.d/global_common.conf
global {
    usage-count no;
    udev-always-use-vnr;
}

common {
    protocol C;
    handlers {

    }

    startup {
             wfc-timeout          120;
             degr-wfc-timeout     120;
             outdated-wfc-timeout 120;
    }

    disk {
        on-io-error detach;
    }

    net {
       cram-hmac-alg md5;
       sndbuf-size 512k;
       shared-secret "testdrbd";
    }
        syncer {
              rate 300M;
        }
}
[root@primary ~]# 

[root@primary ~]# cat /etc/drbd.d/r0.res 
resource r0 {
 on primary {
  device   /dev/drbd0;
  disk /dev/sdb1;     //这个磁盘需要分过区不需要格式化
  address 11.11.11.2:7789;
  meta-disk  internal;
 }
 on secondary {
  device /dev/drbd0;
  disk /dev/sdb1;
  address 11.11.11.3:7789;
  meta-disk internal;
 }
}
[root@primary ~]# 

7、在两台机器上添加DRBD磁盘

在Primary机器上添加一块30G的硬盘作为DRBD，分区为/dev/sdb1，不做格式化，并在本地系统创建/data目录，不做挂载操作。
[root@Primary ~]# fdisk -l
......
[root@Primary ~]# fdisk /dev/sdb      
依次输入"n->p->1->1->回车->w"                                    //分区创建后，再次使用"fdisk /dev/vdd"，输入p，即可查看到创建的分区，比如/dev/vdd1
   
在Secondary机器上添加一块30G的硬盘作为DRBD，分区为/dev/sdb1，不做格式化，并在本地系统创建/data目录，不做挂载操作。
[root@Secondary ~]# fdisk -l
......
[root@Secondary ~]# fdisk /dev/sdb      
依次输入"n->p->1->1->回车->w"

8、在两台机器上分别创建DRBD设备并激活r0资源（下面操作在两台机器上都要执行）

[root@Primary ~]# mknod /dev/drbd0 b 147 0
mknod: `/dev/drbd0': File exists

[root@Primary ~]# drbdadm create-md r0
Writing meta data...
initializing activity log
NOT initialized bitmap
New drbd meta data block successfully created.

[root@Primary ~]# drbdadm create-md r0
You want me to create a v08 style flexible-size internal meta data block.
There appears to be a v08 flexible-size internal meta data block
already in place on /dev/vdd1 at byte offset 10737340416
Do you really want to overwrite the existing v08 meta-data?
[need to type 'yes' to confirm] yes                                  //这里输入"yes"
Writing meta data...
initializing activity log
NOT initialized bitmap
New drbd meta data block successfully created.
   
启动drbd服务（注意：需要主从共同启动方能生效）
service drbd start

 查看状态（两台机器上都执行查看）

[root@primary ~]# cat /proc/drbd
version: 8.4.11-1 (api:1/proto:86-101)
GIT-hash: 66145a308421e9c124ec391a7848ac20203bb03c build by mockbuild@, 2018-11-03 01:26:55
0: cs:SyncSource ro:Secondary/Secondary ds:UpToDate/Inconsistent C r-----     //显示都为 备用；数据还在同步，同步完成后把第一个提升为主
ns:3472496 nr:0 dw:0 dr:3472496 al:0 bm:0 lo:0 pe:0 ua:0 ap:0 ep:1 wo:f oos:1594256
[============>.......] sync'ed: 68.6% (1556/4948)M
finish: 0:00:55 speed: 28,840 (27,776) K/sec
[root@primary ~]#

由上面两台主机的DRBD状态查看结果里的ro:Secondary/Secondary表示两台主机的状态都是备机状态，ds是磁盘状态，显示的状态内容为“不一致”，这是因为DRBD无法判断哪一方为主机，

应以哪一方的磁盘数据作为标准。

 9、接着将Primary主机配置为DRBD的主节点

[root@Primary ~]# drbdsetup /dev/drbd0 primary --force

分别查看主从DRBD状态：
[root@primary ~]# cat /proc/drbd
version: 8.4.11-1 (api:1/proto:86-101)
GIT-hash: 66145a308421e9c124ec391a7848ac20203bb03c build by mockbuild@, 2018-11-03 01:26:55
 0: cs:Connected ro:Primary/Secondary ds:UpToDate/UpToDate C r-----
    ns:5066752 nr:0 dw:0 dr:5068840 al:0 bm:0 lo:0 pe:0 ua:0 ap:0 ep:1 wo:f oos:0

[root@secondary ~]# cat /proc/drbd
version: 8.4.11-1 (api:1/proto:86-101)
GIT-hash: 66145a308421e9c124ec391a7848ac20203bb03c build by mockbuild@, 2018-11-03 01:26:55
 0: cs:Connected ro:Secondary/Primary ds:UpToDate/UpToDate C r-----
    ns:0 nr:5066752 dw:5066752 dr:0 al:0 bm:0 lo:0 pe:0 ua:0 ap:0 ep:1 wo:f oos:0
[root@secondary ~]#
 
ro在主从服务器上分别显示 Primary/Secondary和Secondary/Primary
ds显示UpToDate/UpToDate 表示主从配置成功

10、挂载DRBD  (Primary主节点机器上操作)

从上面Primary主节点的DRBD状态上看到mounted和fstype参数为空，所以这步开始挂载DRBD到系统目录

先格式化/dev/drbd0
[root@Primary ~]# mkfs.ext4 /dev/drbd0

创建挂载目录，然后执行DRBD挂载
[root@Primary ~]# mkdir /data
[root@Primary ~]# mount /dev/drbd0 /data
[root@primary ~]# df -h
文件系统                 容量  已用  可用 已用% 挂载点
/dev/mapper/centos-root   17G  1.4G   16G    8% /
devtmpfs                 899M     0  899M    0% /dev
/dev/sda1               1014M  145M  870M   15% /boot
/dev/drbd0                30G  1.7G   27G    6% /data

特别注意：
Secondary节点上不允许对DRBD设备进行任何操作，包括只读，所有的读写操作只能在Primary节点上进行。
只有当Primary节点挂掉时，Secondary节点才能提升为Primary节点

11、DRBD主备故障切换测试

模拟Primary节点发生故障，Secondary接管并提升为Primary

下面是在Primary主节点上操作记录：
[root@Primary ~]# cd /data
[root@Primary data]# touch wangshibo wangshibo1 wangshibo2 wangshibo3
[root@Primary data]# cd ../
[root@Primary /]# umount /data
   
[root@Primary /]# drbdsetup /dev/drbd0 secondary              //将Primary主机设置为DRBD的备节点。在实际生产环境中，直接在Secondary主机上提权（即设置为主节点）即可。
[root@Primary /]# /etc/init.d/drbd status
drbd driver loaded OK; device status:
version: 8.3.16 (api:88/proto:86-97)
GIT-hash: a798fa7e274428a357657fb52f0ecf40192c1985 build by phil@Build64R6, 2014-11-24 14:51:37
m:res  cs         ro                   ds                 p  mounted  fstype
0:r0   Connected  Secondary/Secondary  UpToDate/UpToDate  C
   
注意：这里实际生产环境若Primary主节点宕机，在Secondary状态信息中ro的值会显示为Secondary/Unknown,只需要进行DRBD提权操作即可。
   
下面是在Secondary 备份节点上操作记录：
先进行提权操作，即将Secondary手动升级为DRBD的主节点
[root@Secondary ~]# drbdsetup /dev/drbd0 primary
[root@Secondary ~]# /etc/init.d/drbd status
drbd driver loaded OK; device status:
version: 8.3.16 (api:88/proto:86-97)
GIT-hash: a798fa7e274428a357657fb52f0ecf40192c1985 build by phil@Build64R6, 2014-11-24 14:51:37
m:res  cs         ro                 ds                 p  mounted  fstype
0:r0   Connected  Primary/Secondary  UpToDate/UpToDate  C
   
然后挂载DRBD
[root@Secondary ~]# mkdir /data
[root@Secondary ~]# mount /dev/drbd0 /data
[root@Secondary ~]# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/VolGroup00-LogVol00
                      156G   13G  135G   9% /
tmpfs                 2.9G     0  2.9G   0% /dev/shm
/dev/vda1             190M   89M   92M  50% /boot
/dev/vdd              9.8G   23M  9.2G   1% /data2
/dev/drbd0            9.8G   23M  9.2G   1% /data
   
发现DRBD挂载目录下已经有了之前在远程Primary主机上写入的内容
[root@Secondary ~]# cd /data
[root@Secondary data]# ls
wangshibo  wangshibo1  wangshibo2  wangshibo3

在Secondary节点上继续写入数据
[root@Secondary data]# touch huanqiu huanqiu1 huanqiu2 huanqiu3
  
然后模拟Secondary节点故障，Primary节点再提权升级为DRBD主节点（操作同上，此处省略.......）

到此，DRBD的主从环境的部署工作已经完成。不过上面是记录的是主备手动切换，至于保证DRBD主从结构的智能切换，实现高可用，还需里用到Keepalived或Heartbeat来实现了（会在DRBD主端挂掉的情况下，自动切换从端为主端并自动挂载/data分区）

12、安装nfs

思路：
1）在两台机器上安装keepalived，VIP为192.168.1.200
2）将DRBD的挂载目录/data作为NFS的挂载目录。远程客户机使用vip地址挂载NFS
3）当Primary主机发生宕机或NFS挂了的故障时，Secondary主机提权升级为DRBD的主节点，并且VIP资源也会转移过来。
   当Primary主机的故障恢复时，会再次变为DRBD的主节点，并重新夺回VIP资源。从而实现故障转移

Primary主机（11.11.11.2）默认作为DRBD的主节点，DRBD挂载目录是/data

Secondary主机（11.11.11.3）是DRBD的备份节点

在Primary和Secondary两台主机上安装NFS
[root@Primary ~]# yum install rpcbind nfs-utils
[root@Primary ~]# vim /etc/exports
/data 11.11.11.0/24(rw,sync,no_root_squash)

service rpcbind start
service nfs start

关闭两台主机的iptables防火墙
防火墙最好关闭，否则可能导致客户机挂载nfs时会失败！
若开启防火墙，需要在iptables中开放nfs相关端口机以及VRRP组播地址
[root@Primary ~]# /etc/init.d/iptables stop
 
两台机器上的selinux一定要关闭！！！！！！！！！！
否则下面在keepalived.conf里配置的notify_master.sh等脚本执行失败！这是曾经踩过的坑！
[root@Primary ~]# setenforce 0     //临时关闭。永久关闭的话，还需要在/etc/sysconfig/selinux 文件里将SELINUX改为disabled
[root@Primary ~]# getenforce
Permissive

13、在两台主机上安装Keepalived，配合keepalived实现自动fail-over

13.1、Primary端配置

安装Keepalived
[root@Primary ~]# yum install -y openssl-devel popt-devel
[root@Primary ~]# cd /usr/local/src/
[root@Primary src]# wget http://www.keepalived.org/software/keepalived-1.3.5.tar.gz
[root@Primary src]# tar -zvxf keepalived-1.3.5.tar.gz
[root@Primary src]# cd keepalived-1.3.5
[root@Primary keepalived-1.3.5]# ./configure --prefix=/usr/local/keepalived
[root@Primary keepalived-1.3.5]# make && make install
        
[root@Primary keepalived-1.3.5]# cp /usr/local/src/keepalived-1.3.5/keepalived/etc/init.d/keepalived /etc/rc.d/init.d/
[root@Primary keepalived-1.3.5]# cp /usr/local/keepalived/etc/sysconfig/keepalived /etc/sysconfig/
[root@Primary keepalived-1.3.5]# mkdir /etc/keepalived/
[root@Primary keepalived-1.3.5]# cp /usr/local/keepalived/etc/keepalived/keepalived.conf /etc/keepalived/
[root@Primary keepalived-1.3.5]# cp /usr/local/keepalived/sbin/keepalived /usr/sbin/
[root@Primary keepalived-1.3.5]# echo "/etc/init.d/keepalived start" >> /etc/rc.local
         
[root@Primary keepalived-1.3.5]# chmod +x /etc/rc.d/init.d/keepalived      #添加执行权限
[root@Primary keepalived-1.3.5]# chkconfig keepalived on                   #设置开机启动
[root@Primary keepalived-1.3.5]# service keepalived start                   #启动
[root@Primary keepalived-1.3.5]# service keepalived stop                    #关闭
[root@Primary keepalived-1.3.5]# service keepalived restart                 #重启

yum install -y openssl-devel popt-devel
yum install keepalived

-----------Primary主机的keepalived.conf配置
[root@Primary ~]# cp /etc/keepalived/keepalived.conf /etc/keepalived/keepalived.conf-bak
[root@primary ~]# cat /etc/keepalived/keepalived.conf
global_defs {
   notification_email {
    root@localhost
   }
   notification_email_from keepalived@localhost
   smtp_server 127.0.0.1
   smtp_connect_timeout 30
   router_id DRBD_HA_MASTER
}

vrrp_script chk_nfs {
        script "/etc/keepalived/check_nfs.sh"
        interval 5
    }

vrrp_instance VI_1 {
    state MASTER
    interface ens33   //注意修改成自己的网卡名称
    virtual_router_id 51
    priority 100
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    track_script {
       chk_nfs
    }
    virtual_ipaddress {
       11.11.11.4
    }
    nopreempt
    notify_stop "/etc/keepalived/notify_stop.sh"
    notify_master "/etc/keepalived/notify_master.sh"
}

启动keepalived服务

[root@primary ~]# service keepalived start
Redirecting to /bin/systemctl start keepalived.service
[root@primary ~]# ps -ef|grep keepalived
root 7314 1 0 14:22 ? 00:00:00 /usr/sbin/keepalived -D
root 7315 7314 0 14:22 ? 00:00:00 /usr/sbin/keepalived -D
root 7316 7314 0 14:22 ? 00:00:00 /usr/sbin/keepalived -D
root 7431 7119 0 14:22 pts/1 00:00:00 grep --color=auto keepalived
[root@primary ~]#

查看VIP

1、primary的脚本内容

1）此脚本只在Primary机器上配置
[root@primary ~]# cat /etc/keepalived/check_nfs.sh 
/sbin/service nfs status &>/dev/null
if [ $? -ne 0 ];then
    ###如果服务状态不正常，先尝试重启服务
    /sbin/service nfs restart
    /sbin/service nfs status &>/dev/null
    if [ $? -ne 0 ];then
        ###若重启nfs服务后，仍不正常
        ###卸载drbd设备
        umount /dev/drbd0
        ###将drbd主降级为备
        drbdadm secondary r0
        #关闭keepalived
        /sbin/service keepalived stop
    fi
fi
[root@Primary ~]# chmod 755 /etc/keepalived/check_nfs.sh 

2）此脚本只在Primary机器上配置
[root@Primary ~]# mkdir /etc/keepalived/logs
[root@primary ~]# cat /etc/keepalived/notify_stop.sh
#!/bin/bash
  
time=`date "+%F  %H:%M:%S"`
echo -e "$time  ------notify_stop------
" >> /etc/keepalived/logs/notify_stop.log
/sbin/service nfs stop &>> /etc/keepalived/logs/notify_stop.log
/bin/umount /dev/drbd0 &>> /etc/keepalived/logs/notify_stop.log
/sbin/drbdadm secondary r0 &>> /etc/keepalived/logs/notify_stop.log
echo -e "
" >> /etc/keepalived/logs/notify_stop.log
[root@primary ~]# 
[root@Primary ~]# chmod 755 /etc/keepalived/notify_stop.sh

2、此脚本在两台机器上都要配置

[root@primary ~]# cat /etc/keepalived/notify_master.sh
#!/bin/bash
  
time=`date "+%F  %H:%M:%S"`
echo -e "$time    ------notify_master------
" >> /etc/keepalived/logs/notify_master.log
/sbin/drbdadm primary data1 &>> /etc/keepalived/logs/notify_master.log
/bin/mount /dev/drbd0 /data &>> /etc/keepalived/logs/notify_master.log
/sbin/service nfs restart &>> /etc/keepalived/logs/notify_master.log
echo -e "
" >> /etc/keepalived/logs/notify_master.log
[root@primary ~]# 

[root@Primary ~]# chmod 755   /etc/keepalived/notify_master.sh

 13.2、Secondary端配置

-----------Secondary主机的keepalived.conf配置
[root@Secondary ~]# cp /etc/keepalived/keepalived.conf /etc/keepalived/keepalived.conf-bak
[root@secondary ~]# cat /etc/keepalived/keepalived.conf
global_defs {
   notification_email {
     root@localhost
   }
   notification_email_from keepalived@localhost
   smtp_server 127.0.0.1
   smtp_connect_timeout 30
   router_id DRBD_HA_BACKUP
}

vrrp_instance VI_1 {
    state BACKUP
    interface ens33  //注意修改成自己的网卡名称
    virtual_router_id 51
    priority 90
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }

   nopreempt
   notify_master "/etc/keepalived/notify_master.sh"    //当此机器为keepalived的master角色时执行这个脚本
   notify_backup "/etc/keepalived/notify_backup.sh"     //当此机器为keepalived的backup角色时执行这个脚本
    virtual_ipaddress {
        11.11.11.4
    }
}

启动keepalived服务

[root@Primary ~]# chmod 755  /etc/keepalived/notify_master.sh
[root@Primary ~]# chmod 755  /etc/keepalived/notify_backup.sh

 [root@primary ~]# service keepalived start
  Redirecting to /bin/systemctl start keepalived.service

1、此脚本只在Secondary机器上配置

[root@Secondary ~]# mkdir /etc/keepalived/logs
[root@Secondary ~]# vim /etc/keepalived/notify_backup.sh
#!/bin/bash
  
time=`date "+%F  %H:%M:%S"`
echo -e "$time    ------notify_backup------
" >> /etc/keepalived/logs/notify_backup.log
/sbin/service nfs stop &>> /etc/keepalived/logs/notify_backup.log
/bin/umount /dev/drbd0 &>> /etc/keepalived/logs/notify_backup.log
/sbin/drbdadm secondary data1 &>> /etc/keepalived/logs/notify_backup.log
echo -e "
" >> /etc/keepalived/logs/notify_backup.log
[root@Secondary ~]# chmod 755 /etc/keepalived/notify_backup.sh

14、远程客户挂载NFS

客户端只需要安装rpcbind程序，并确认服务正常
[root@huanqiu ~]# yum install rpcbind nfs-utils
[root@huanqiu ~]# /etc/init.d/rpcbind start

linux端：
umount /web/
mount -t nfs 11.11.11.4:/data /web

windos2012端：
mount \11.11.11.4data -o nolock,rsize=1024,wsize=1024,timeo=15 z:     优，切换中几乎不断

主从出现UpToDate/DUnknown 故障恢复
https://blog.csdn.net/kjsayn/article/details/52958282


windos修改注册表（挂载上去之后没有写权限）
https://jingyan.baidu.com/article/c910274bfd6800cd361d2df3.html

 15、前端nginx缓存

创建缓存空间
mkdir /cache 
nginx -t 
nginx -s reload 

[root@localhost conf.d]# pwd
/etc/nginx/conf.d
[root@localhost conf.d]# cat web.conf 
upstream node {
    server 11.11.11.9:80;
}
proxy_cache_path /cache levels=1:2 keys_zone=cache:10m max_size=10g inactive=60m use_temp_path=off;
server {
    listen 80;
    server_name www.test.com;
    index index.html;
    location / {
    proxy_pass http://node;
    proxy_cache cache;
    proxy_cache_valid   200 304 12h;
    proxy_cache_valid   any 10m;
    add_header  Nginx-Cache "$upstream_cache_status";
    proxy_next_upstream error timeout invalid_header http_500 http_502 http_503 http_504;
        }
    }
[root@localhost conf.d]# 

 关于nginx缓存问题，详情请查看另一个博文

15.2、清除缓存

rm删除已缓存的数据
rm -rf /cache/*

16、测试

1）
先关闭Primary主机上的keepalived服务。就会发现VIP资源已经转移到Secondary主机上了。
同时，Primary主机的nfs也会主动关闭，同时Secondary会升级为DRBD的主节点
[root@Primary ~]# /etc/init.d/keepalived stop
Stopping keepalived:                                       [  OK  ]
[root@Primary ~]# ip addr
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 fa:16:3e:35:d1:d6 brd ff:ff:ff:ff:ff:ff
    inet 192.168.1.151/24 brd 192.168.1.255 scope global eth0
    inet6 fe80::f816:3eff:fe35:d1d6/64 scope link
       valid_lft forever preferred_lft forever
  
查看系统日志，也能看到VIP资源转移信息
[root@Primary ~]# tail -1000 /var/log/messages
........
May 25 11:50:03 localhost Keepalived_vrrp[30940]: Sending gratuitous ARP on eth0 for 192.168.1.200
May 25 11:50:03 localhost Keepalived_vrrp[30940]: Sending gratuitous ARP on eth0 for 192.168.1.200
May 25 11:50:03 localhost Keepalived_vrrp[30940]: Sending gratuitous ARP on eth0 for 192.168.1.200
May 25 11:50:03 localhost Keepalived_vrrp[30940]: Sending gratuitous ARP on eth0 for 192.168.1.200
May 25 11:58:51 localhost Keepalived[30937]: Stopping
May 25 11:58:51 localhost Keepalived_vrrp[30940]: VRRP_Instance(VI_1) sent 0 priority
May 25 11:58:51 localhost Keepalived_vrrp[30940]: VRRP_Instance(VI_1) removing protocol VIPs.
  
[root@Primary ~]# ps -ef|grep nfs
root       588 10364  0 12:13 pts/1    00:00:00 grep --color nfs
[root@Primary ~]# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/VolGroup00-LogVol00
                      156G   36G  112G  25% /
tmpfs                 2.9G     0  2.9G   0% /dev/shm
/dev/vda1             190M   98M   83M  55% /boot
[root@Primary ~]# /etc/init.d/drbd status
drbd driver loaded OK; device status:
version: 8.3.16 (api:88/proto:86-97)
GIT-hash: a798fa7e274428a357657fb52f0ecf40192c1985 build by phil@Build64R6, 2014-11-24 14:51:37
m:res  cs         ro                   ds                 p  mounted  fstype
0:r0   Connected  Secondary/Secondary  UpToDate/UpToDate  C
  
登录到Secondary备份机器上，发现VIP资源已经转移过来
[root@Secondary ~]# ip addr
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 fa:16:3e:4c:7e:88 brd ff:ff:ff:ff:ff:ff
    inet 192.168.1.152/24 brd 192.168.1.255 scope global eth0
    inet 192.168.1.200/32 scope global eth0
    inet6 fe80::f816:3eff:fe4c:7e88/64 scope link
       valid_lft forever preferred_lft forever
  
[root@Secondary ~]# tail -1000 /var/log/messages
........
May 25 11:58:53 localhost Keepalived_vrrp[17131]: Sending gratuitous ARP on eth0 for 192.168.1.200
May 25 11:58:53 localhost Keepalived_vrrp[17131]: Sending gratuitous ARP on eth0 for 192.168.1.200
May 25 11:58:53 localhost Keepalived_vrrp[17131]: Sending gratuitous ARP on eth0 for 192.168.1.200
May 25 11:58:53 localhost Keepalived_vrrp[17131]: Sending gratuitous ARP on eth0 for 192.168.1.200
May 25 11:58:58 localhost Keepalived_vrrp[17131]: Sending gratuitous ARP on eth0 for 192.168.1.200
May 25 11:58:58 localhost Keepalived_vrrp[17131]: VRRP_Instance(VI_1) Sending/queueing gratuitous ARPs on eth0 for 192.168.1.200
  
[root@Secondary ~]# ip addr
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 fa:16:3e:4c:7e:88 brd ff:ff:ff:ff:ff:ff
    inet 192.168.1.152/24 brd 192.168.1.255 scope global eth0
    inet 192.168.1.200/32 scope global eth0
    inet6 fe80::f816:3eff:fe4c:7e88/64 scope link
       valid_lft forever preferred_lft forever
[root@Secondary ~]# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/VolGroup00-LogVol00
                      156G   13G  135G   9% /
tmpfs                 2.9G     0  2.9G   0% /dev/shm
/dev/vda1             190M   89M   92M  50% /boot
/dev/drbd0            9.8G   23M  9.2G   1% /data
  
当Primary机器的keepalived服务恢复启动后，VIP资源又会强制夺回来（可以查看/var/log/message系统日志）
并且Primary还会再次变为DRBD的主节点
  
2）
关闭Primary主机的nfs服务。根据监控脚本，会主动去启动nfs，只要当启动失败时，才会强制由DRBD的主节点降为备份节点，并关闭keepalived。
从而跟上面流程一样实现故障转移
  
结论：
在上面的主从故障切换过程中，对于客户端来说，挂载NFS不影响使用，只是会有一点的延迟。
这也验证了drbd提供的数据一致性功能（包括文件的打开和修改状态等），在客户端看来，真个切换过程就是"一次nfs重启"（主nfs停，备nfs启）。

Top：

1、说一下"Split-Brain"（脑裂）的情况：

假设把Primary主机的的eth0设备宕掉，然后直接在Secondary主机上进行提权升级为DRBD的主节点，并且mount挂载DRBD，这时会发现之前在Primary主机上写入的数据文件确实同步过来了。
接着再把Primary主机的eth0设备恢复，看看有没有自动恢复 主从关系。经过查看，发现DRBD检测出了Split-Brain的状况，也就是两个节点都处于standalone状态，
故障描述如下：Split-Brain detected,dropping connection! 这就是传说中的“脑裂”。
  
DRBD官方推荐的手动恢复方案：
  
1）Secondary主机上的操作
# drbdadm secondary r0
# drbdadm disconnect all
# drbdadm --discard-my-data connect r0            //或者"drbdadm  --  --discard-my-data  connect r0"
  
2）Primary主机上的操作
# drbdadm disconnect all
# drbdadm connect r0
# drbdsetup /dev/drbd0 primary