Heartbeat+DRBD+NFS高可用实例配置

环境说明:   
操作系统: CentOS 6.5 x64,本文采用rpm方式安装heartbeat+drbd，本文只是试用heartbeat+drbd+nfs高可用基本功能。

app1: 192.168.0.24  
app1: 192.168.0.25   
VIP : 192.168.0.26

一、双机Heartbeat配置

1. app1,app2配置hosts文件

[root@app1 soft]# vi /etc/hosts  
127.0.0.1   localhost localhost.localdomain localhost4 localhost4.localdomain4   
::1         localhost localhost.localdomain localhost6 localhost6.localdomain6   
192.168.0.24         app1   
192.168.0.25         app2   
10.10.10.24          app1-priv   
10.10.10.25          app2-priv

说明：10段是心跳IP, 192.168段是业务IP, 采用VIP地址是192.168.0.26。

2. app1,app2配置安装epel源并安装heartbeat

# wget http://download.fedoraproject.org/pub/epel/6/x86_64/epel-release-6-8.noarch.rpm    
# rpm -ivh epel-release-6-8.noarch.rpm    
# yum install heartbeat

3. app1,app2配置安装Heartbeat   

说明：本文采用RPM方式包，提前下载到本地的，再安装heartbeat

[root@app1 ~]# cd soft/  
[root@app1 soft]# ll    
总用量 1924    
-rw-r--r-- 1 root root  72744 6月  25 2012 cluster-glue-1.0.5-6.el6.x86_64.rpm    
-rw-r--r-- 1 root root 119096 6月  25 2012 cluster-glue-libs-1.0.5-6.el6.x86_64.rpm    
-rw-r--r-- 1 root root 165292 12月  3 2013 heartbeat-3.0.4-2.el6.x86_64.rpm    
-rw-r--r-- 1 root root 269468 12月  3 2013 heartbeat-libs-3.0.4-2.el6.x86_64.rpm    
-rw-r--r-- 1 root root  38264 10月 18 2014 perl-TimeDate-1.16-13.el6.noarch.rpm    
-rw-r--r-- 1 root root 913840 7月   3 2011 PyXML-0.8.4-19.el6.x86_64.rpm    
-rw-r--r-- 1 root root 374068 11月 10 20:45 resource-agents-3.9.5-24.el6_7.1.x86_64.rpm    
[root@app1 soft]#    
[root@app1 soft]# rpm -ivh *.rpm    
warning: cluster-glue-1.0.5-6.el6.x86_64.rpm: Header V3 RSA/SHA1 Signature, key ID c105b9de: NOKEY    
warning: heartbeat-3.0.4-2.el6.x86_64.rpm: Header V3 RSA/SHA256 Signature, key ID 0608b895: NOKEY    
Preparing...                ########################################### [100%]    
   1:cluster-glue-libs      ########################################### [ 14%]    
   2:resource-agents        ########################################### [ 29%]    
   3:PyXML                  ########################################### [ 43%]    
   4:perl-TimeDate          ########################################### [ 57%]    
   5:cluster-glue           ########################################### [ 71%]    
   6:heartbeat-libs         ########################################### [ 86%]    
   7:heartbeat              ########################################### [100%]    
[root@app1 soft]#

(1) 设置授权KEY

# vi /etc/ha.d/authkeys  
auth 1   
1 sha1 47e9336850f1db6fa58bc470bc9b7810eb397f04    
# chmod 600 /etc/ha.d/authkeys

(2) 添加配置ha资源文件

# vi /etc/ha.d/haresources    
 
# 初始状态服务器绑定VIP的地址在哪个服务器、哪个网卡上，启动什么服务。    
app1  IPaddr::192.168.0.26/24/eth0:1    
#app1  IPaddr::192.168.0.26/24/eth0:1 drbddisk::data Filesystem::/dev/drbd0::/data::ext4

(3) 配置heartbeat主配置文件

APP1上配置文件:

# vi /etc/ha.d/ha.cf  
debugfile /var/log/ha-debug  
logfile /var/log/ha-log   
logfacility     local0   
keepalive 2   
deadtime 30   
warntime 10   
initdead 120   
udpport 694   
bcast eth1   
ucast eth1 10.10.10.25   
#mcast eth1 225.0.0.24 694 1 0   
auto_failback on   
node    app1   
node    app2  
crm     no    
#respawn hacluster /usr/lib64/heartbeat/ipfail  
#ping 192.168.0.253

APP2上配置文件,与主配置文件有些区别，需要修改。

# vi /etc/ha.d/ha.cf  
debugfile /var/log/ha-debug  
logfile /var/log/ha-log   
logfacility     local0   
keepalive 2   
deadtime 15   
warntime 10   
initdead 120   
udpport 694   
bcast eth1   
ucast eth1 10.10.10.24   
#mcast eth1 225.0.0.25 694 1 0   
auto_failback on   
node    app1   
node    app2  
crm     no    
#respawn hacluster /usr/lib64/heartbeat/ipfail  
#ping 192.168.0.253

4. 将刚才配置的三个文件同步至app2，同步过去后要修改ha.cf文件中的心跳IP

# scp authkeys ha.cf haresources root@app2:/etc/ha.d/  
root@app2's password:    
authkeys                                      100%   56     0.1KB/s   00:00   
ha.cf                                         100%  256     0.3KB/s   00:00   
haresources                                   100%   78     0.1KB/s   00:00

5. 启动heartbeat服务，测试能否正常提供服务

节点1:  
[root@app1 ha.d]# service heartbeat start   
Starting High-Availability services: INFO:  Resource is stopped   
Done.

节点2:  
[root@app2 ha.d]# service heartbeat start   
Starting High-Availability services: INFO:  Resource is stopped   
Done.

6. 手动测试VIP切换

(1) 手动切换成standby状态 
 
# /usr/share/heartbeat/hb_standby   
Going standby [all].

或者主服务器 service heartbeat stop  也可以切换VIP到备机上。

(2) 手动切换成主状态

# /usr/share/heartbeat/hb_takeover

主服务器 service heartbeat start  也可以将VIP切回来。

(3) 通过日志查看VIP接管过程

节点1：

# tail -f /var/log/message

Jan 12 12:46:30 app1 heartbeat: [4519]: info: app1 wants to go standby [all]  
Jan 12 12:46:30 app1 heartbeat: [4519]: info: standby: app2 can take our all resources   
Jan 12 12:46:30 app1 heartbeat: [6043]: info: give up all HA resources (standby).   
Jan 12 12:46:30 app1 ResourceManager(default)[6056]: info: Releasing resource group: app1 IPaddr::192.168.0.26/24/eth0   
Jan 12 12:46:30 app1 ResourceManager(default)[6056]: info: Running /etc/ha.d/resource.d/IPaddr 192.168.0.26/24/eth0 stop   
Jan 12 12:46:30 app1 IPaddr(IPaddr_192.168.0.26)[6119]: INFO: IP status = ok, IP_CIP=   
Jan 12 12:46:30 app1 /usr/lib/ocf/resource.d//heartbeat/IPaddr(IPaddr_192.168.0.26)[6093]: INFO:  Success   
Jan 12 12:46:30 app1 heartbeat: [6043]: info: all HA resource release completed (standby).   
Jan 12 12:46:30 app1 heartbeat: [4519]: info: Local standby process completed [all].   
Jan 12 12:46:31 app1 heartbeat: [4519]: WARN: 1 lost packet(s) for [app2] [1036:1038]   
Jan 12 12:46:31 app1 heartbeat: [4519]: info: remote resource transition completed.   
Jan 12 12:46:31 app1 heartbeat: [4519]: info: No pkts missing from app2!   
Jan 12 12:46:31 app1 heartbeat: [4519]: info: Other node completed standby takeover of all resources.

节点2：  
[root@app2 ha.d]# tail -f /var/log/messages   
Jan 12 12:46:30 app2 heartbeat: [4325]: info: app1 wants to go standby [all]   
Jan 12 12:46:30 app2 heartbeat: [4325]: info: standby: acquire [all] resources from app1   
Jan 12 12:46:30 app2 heartbeat: [5459]: info: acquire all HA resources (standby).   
Jan 12 12:46:30 app2 ResourceManager(default)[5472]: info: Acquiring resource group: app1 IPaddr::192.168.0.26/24/eth0   
Jan 12 12:46:30 app2 /usr/lib/ocf/resource.d//heartbeat/IPaddr(IPaddr_192.168.0.26)[5500]: INFO:  Resource is stopped   
Jan 12 12:46:30 app2 ResourceManager(default)[5472]: info: Running /etc/ha.d/resource.d/IPaddr 192.168.0.26/24/eth0 start   
Jan 12 12:46:31 app2 IPaddr(IPaddr_192.168.0.26)[5625]: INFO: Adding inet address 192.168.0.26/24 with broadcast address 192.168.0.255 to device eth0   
Jan 12 12:46:31 app2 IPaddr(IPaddr_192.168.0.26)[5625]: INFO: Bringing device eth0 up   
Jan 12 12:46:31 app2 IPaddr(IPaddr_192.168.0.26)[5625]: INFO: /usr/libexec/heartbeat/send_arp -i 200 -r 5 -p /var/run/resource-agents/send_arp-192.168.0.26 eth0 192.168.0.26 auto not_used not_used   
Jan 12 12:46:31 app2 /usr/lib/ocf/resource.d//heartbeat/IPaddr(IPaddr_192.168.0.26)[5599]: INFO:  Success   
Jan 12 12:46:31 app2 heartbeat: [5459]: info: all HA resource acquisition completed (standby).   
Jan 12 12:46:31 app2 heartbeat: [4325]: info: Standby resource acquisition done [all].   
Jan 12 12:46:31 app2 heartbeat: [4325]: info: remote resource transition completed.

手动添加VIP地址命令：  
/etc/ha.d/resource.d/IPaddr 192.168.0.27/24/eth0:2 start

(4)  查看VIP地址信息,VIP在主节点上。

[root@app1 ha.d]# ip a   
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:29:4c:39:43 brd ff:ff:ff:ff:ff:ff   
    inet 192.168.0.24/24 brd 192.168.0.255 scope global eth0   
    inet 192.168.0.26/24 brd 192.168.0.255 scope global secondary eth0:1    
    inet6 fe80::20c:29ff:fe4c:3943/64 scope link    
       valid_lft forever preferred_lft forever   
3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000   
    link/ether 00:0c:29:4c:39:4d brd ff:ff:ff:ff:ff:ff   
    inet 10.10.10.24/24 brd 10.10.10.255 scope global eth1   
    inet6 fe80::20c:29ff:fe4c:394d/64 scope link    
       valid_lft forever preferred_lft forever   

二、DRDB安装配置

1. app1,app2配置hosts文件以及准备磁盘分区

app1: /dev/sdb1  
app2: /dev/sdb1

2. 安装drbd并安装

(1) 下载drbd安装包,下载地址如下:

ftp://rpmfind.net/linux/atrpms/el6.5-x86_64/atrpms/stable/

# rpm -ivh drbd-8.4.3-33.el6.x86_64.rpm drbd-kmdl-2.6.32-431.el6-8.4.3-33.el6.x86_64.rpm    
warning: drbd-8.4.3-33.el6.x86_64.rpm: Header V4 DSA/SHA1 Signature, key ID 66534c2b: NOKEY    
Preparing...                ########################################### [100%]    
   1:drbd-kmdl-2.6.32-431.el########################################### [ 50%]    
   2:drbd                   ########################################### [100%]    
#

(2) 加载DRBD到内核模块

app1,app2分别操作，并加入到/etc/rc.local文件中，可以事先尝试有无自动加载。  
lsmode |grep drbd    
modprobe drbd

3. 创建修改配置文件。节点1，节点2一样配置。

[root@app1 ~]# vi /etc/drbd.d/global_common.conf  
global {    
        usage-count no;    
}    
common {    
        protocol C;    
        disk {    
                on-io-error detach;    
                no-disk-flushes;    
                no-md-flushes; 
        }    
        net {    
                sndbuf-size 512k;    
                max-buffers     8000;    
                unplug-watermark   1024;    
                max-epoch-size  8000;    
                cram-hmac-alg "sha1";    
                shared-secret "hdhwXes23sYEhart8t";    
                after-sb-0pri disconnect;    
                after-sb-1pri disconnect;    
                after-sb-2pri disconnect;    
                rr-conflict disconnect;    
        }    
        syncer {    
                rate 300M;    
                al-extents 517;    
        }    
}

resource data {  
      on app1 {    
               device    /dev/drbd0;    
               disk      /dev/sdb1;    
               address   192.168.1.120:7788;    
               meta-disk internal;    
      }    
      on app2 {    
               device     /dev/drbd0;    
               disk       /dev/sdb1;    
               address    192.168.1.121:7788;    
               meta-disk internal;    
      }    
}

下面是采用内部模式: 用于解决迁移问题，这个实验一直没有做成功，下次再试吧。    
resource data {    
      on app1 {    
               device    /dev/drbd0;    
               disk      /dev/sdb1;    
               address   192.168.0.24:7788;    
               meta-disk /dev/sdc1 [0];    
      }    
      on app2 {    
               device     /dev/drbd0;    
               disk       /dev/sdb1;    
               address    192.168.0.25:7788;    
               meta-disk  /dev/sdc1 [0];    
      }    
}

三、drbd启动和查看

1. 初始化资源

在app1和app2上分别执行：

# drbdadm create-md data

Writing meta data...  
initializing activity log    
NOT initializing bitmap    
New drbd meta data block successfully created.

说明: 这一部会出现的问题如下:    
# drbdadm create-md data    
Command 'drbdmeta 1 v08 /dev/sdb1 internal create-md' terminated with exit code 40

#解决如下，非要做如下DD操作，可能bug  
# dd if=/dev/zero of=/dev/sdb1 bs=1M count=10    
# sync    
# drbdadm create-md data

2. 启动服务

在app1和app2上分别执行：或采用 drbdadm up data

# service drbd start

Starting DRBD resources: [  
     create res: data    
   prepare disk: data    
    adjust disk: data    
     adjust net: data    
]    
..........    
#

3. 查看启动状态, 两节点应均处于Secondary状态。

cat /proc/drbd       #或者直接使用命令drbd-overview

节点1:    
[root@app1 drbd.d]# cat /proc/drbd 
version: 8.4.3 (api:1/proto:86-101)    
GIT-hash: 89a294209144b68adb3ee85a73221f964d3ee515 build by gardner@, 2013-11-29 12:28:00    
0: cs:Connected ro:Secondary/Secondary ds:Inconsistent/Inconsistent C r-----    
    ns:0 nr:0 dw:0 dr:0 al:0 bm:0 lo:0 pe:0 ua:0 ap:0 ep:1 wo:d oos:20970828    
[root@app1 drbd.d]#

节点2:    
[root@app2 drbd.d]# cat /proc/drbd 
version: 8.4.3 (api:1/proto:86-101)    
GIT-hash: 89a294209144b68adb3ee85a73221f964d3ee515 build by gardner@, 2013-11-29 12:28:00    
0: cs:Connected ro:Secondary/Secondary ds:Inconsistent/Inconsistent C r-----    
    ns:0 nr:0 dw:0 dr:0 al:0 bm:0 lo:0 pe:0 ua:0 ap:0 ep:1 wo:d oos:20970828    
[root@app2 drbd.d]#

4. 将其中一个节点配置为主节点

我们需要将其中一个节点设置为Primary，在要设置为Primary的节点上执行如下命令：  
drbdadm -- --overwrite-data-of-peer primary data  
drbdadm primary --force data

主节点查看同步状态：    
[root@app1 drbd.d]# cat /proc/drbd 
version: 8.4.3 (api:1/proto:86-101)    
GIT-hash: 89a294209144b68adb3ee85a73221f964d3ee515 build by gardner@, 2013-11-29 12:28:00    
0: cs:SyncSource ro:Primary/Secondary ds:UpToDate/Inconsistent C r---n-    
    ns:1440320 nr:0 dw:0 dr:1443488 al:0 bm:85 lo:0 pe:36 ua:3 ap:0 ep:1 wo:d oos:19566924    
        [>...................] sync'ed:  6.7% (19108/20476)M    
        finish: 0:24:03 speed: 13,536 (12,760) K/sec    
[root@app1 drbd.d]#

备节点查看同步状态：    
[root@app2 drbd.d]# cat /proc/drbd    
version: 8.4.3 (api:1/proto:86-101)    
GIT-hash: 89a294209144b68adb3ee85a73221f964d3ee515 build by gardner@, 2013-11-29 12:28:00    
0: cs:SyncTarget ro:Secondary/Primary ds:Inconsistent/UpToDate C r-----    
    ns:0 nr:2063360 dw:2030592 dr:0 al:0 bm:123 lo:33 pe:8 ua:32 ap:0 ep:1 wo:d oos:18940236    
        [>...................] sync'ed:  9.7% (18496/20476)M    
        finish: 0:23:54 speed: 13,196 (12,848) want: 3,240 K/sec    
[root@app2 drbd.d]#

查看同步状态:

[root@app1 ~]# cat /proc/drbd 
version: 8.4.3 (api:1/proto:86-101)    
GIT-hash: 89a294209144b68adb3ee85a73221f964d3ee515 build by gardner@, 2013-11-29 12:28:00    
0: cs:Connected ro:Primary/Secondary ds:UpToDate/UpToDate C r-----    
    ns:20970828 nr:0 dw:0 dr:20971500 al:0 bm:1280 lo:0 pe:0 ua:0 ap:0 ep:1 wo:d oos:0    
[root@app1 ~]#

[root@app2 drbd.d]# cat /proc/drbd    
version: 8.4.3 (api:1/proto:86-101)    
GIT-hash: 89a294209144b68adb3ee85a73221f964d3ee515 build by gardner@, 2013-11-29 12:28:00    
0: cs:Connected ro:Secondary/Primary ds:UpToDate/UpToDate C r-----    
    ns:0 nr:20970828 dw:20970828 dr:0 al:0 bm:1280 lo:0 pe:0 ua:0 ap:0 ep:1 wo:d oos:0

5. 创建文件系统

文件系统的挂载只能在Primary节点进行，只有在设置了主节点后才能对drbd设备进行格式化, 格式化与手动挂载测试。

[root@app1 ~]# mkfs.ext4 /dev/drbd0  
[root@app1 ~]# mount /dev/drbd0 /data

6. 手动切换Primary和Secondary

对主Primary/Secondary模型的drbd服务来讲，在某个时刻只能有一个节点为Primary，因此，要切换两个节点的角色，只能在先将原有的Primary节点设置为Secondary后，才能原来的Secondary节点设置为Primary:

手工切换DRBD的步骤:

(1) 主节点 umount /dev/drbd0 卸载挂载  
(2) 主节点 drbdadm secondary all  恢复从节点    
(3) 备节点 drbdadm primary all 配置主节点    
(4) 备节点 mount /dev/drbd0 /data  挂载

7. DRBD脑裂后的处理

当DRBD出现脑裂后，会导致drbd两边的磁盘不一致，处理方法如下：

在确定要作为从的节点上切换成secondary，并放弃该资源的数据：

drbdadm secondary all  
drbdadm -- --discard-my-data connect all

重新同步数据:    
drbdadm -- --overwrite-data-of-peer primary data 

四、安装配置NFS

1. app1,app2节点配置nfs

# vi /etc/exports  
/data 192.168.0.0/24(rw,no_root_squash)

2. app1,app2节点配置nfs

# service rpcbind start  
# service nfs start    
# chkconfig rpcbind on    
# chkconfig nfs on

3. app1,app2节点配置nfs

# vi haresources    
#app1  IPaddr::192.168.0.26/24/eth0:1    
app1   IPaddr::192.168.0.26/24/eth0:1 drbddisk::data Filesystem::/dev/drbd0::/data::ext4 nfs

参数说明:    
IPaddr::192.168.0.26/24/eth0:1        #虚拟IP地址    
drbddisk::data                        #管理drbd资源    
Filesystem::/dev/drbd0::/data::ext4   #挂载文件系统    
nfs                                   #nfs脚本

4. app1,app2配置nfs脚本

# vi /etc/ha.d/resource.d/nfs  
killall -9 nfsd    
/etc/init.d/nfs restart    
exit 0

# chmod +x /etc/ha.d/resource.d/nfs

五、测试切换

1. 通过一台客户机挂载

[root@vm15 ~]# mount -t nfs 192.168.0.26:/data/ /mnt  
[root@vm15 ~]# df -h    
Filesystem            Size  Used Avail Use% Mounted on    
/dev/sda3              21G  4.6G   15G  24% /    
/dev/sda1              99M   23M   72M  25% /boot    
tmpfs                 7.4G     0  7.4G   0% /dev/shm    
/dev/mapper/vg-data    79G   71G  4.2G  95% /data    
192.168.0.26:/data/   9.9G  151M  9.2G   2% /mnt    
[root@vm15 ~]#

2. 节点1上执行: service heartbeat stop 或  /usr/share/heartbeat/hb_standby

3. 观察日志情况

资源主节点：  
# tail -f /var/log/message

Jan 22 15:46:01 app2 heartbeat: [8050]: info: app2 wants to go standby [all]  
Jan 22 15:46:01 app2 heartbeat: [8050]: info: standby: app1 can take our all resources    
Jan 22 15:46:01 app2 heartbeat: [9310]: info: give up all HA resources (standby).    
Jan 22 15:46:01 app2 ResourceManager(default)[9323]: info: Releasing resource group: app1 IPaddr::192.168.0.26/24/eth0:1 drbddisk::data Filesystem::/dev/drbd0::/data::ext4 nfs    
Jan 22 15:46:01 app2 ResourceManager(default)[9323]: info: Running /etc/ha.d/resource.d/nfs  stop    
Jan 22 15:46:01 app2 kernel: nfsd: last server has exited, flushing export cache    
Jan 22 15:46:01 app2 rpc.mountd[9218]: Caught signal 15, un-registering and exiting.    
Jan 22 15:46:02 app2 rpc.mountd[9452]: Version 1.2.3 starting    
Jan 22 15:46:02 app2 kernel: NFSD: Using /var/lib/nfs/v4recovery as the NFSv4 state recovery directory    
Jan 22 15:46:02 app2 kernel: NFSD: starting 90-second grace period    
Jan 22 15:46:02 app2 ResourceManager(default)[9323]: info: Running /etc/ha.d/resource.d/Filesystem /dev/drbd0 /data ext4 stop    
Jan 22 15:46:02 app2 Filesystem(Filesystem_/dev/drbd0)[9519]: INFO: Running stop for /dev/drbd0 on /data    
Jan 22 15:46:02 app2 Filesystem(Filesystem_/dev/drbd0)[9519]: INFO: Trying to unmount /data    
Jan 22 15:46:02 app2 Filesystem(Filesystem_/dev/drbd0)[9519]: INFO: unmounted /data successfully    
Jan 22 15:46:02 app2 /usr/lib/ocf/resource.d//heartbeat/Filesystem(Filesystem_/dev/drbd0)[9511]: INFO:  Success    
Jan 22 15:46:02 app2 ResourceManager(default)[9323]: info: Running /etc/ha.d/resource.d/drbddisk data stop    
Jan 22 15:46:02 app2 kernel: block drbd0: role( Primary -> Secondary )    
Jan 22 15:46:02 app2 kernel: block drbd0: bitmap WRITE of 0 pages took 0 jiffies    
Jan 22 15:46:02 app2 kernel: block drbd0: 0 KB (0 bits) marked out-of-sync by on disk bit-map.    
Jan 22 15:46:02 app2 ResourceManager(default)[9323]: info: Running /etc/ha.d/resource.d/IPaddr 192.168.0.26/24/eth0:1 stop    
Jan 22 15:46:02 app2 IPaddr(IPaddr_192.168.0.26)[9679]: INFO: IP status = ok, IP_CIP=    
Jan 22 15:46:02 app2 /usr/lib/ocf/resource.d//heartbeat/IPaddr(IPaddr_192.168.0.26)[9653]: INFO:  Success    
Jan 22 15:46:02 app2 heartbeat: [9310]: info: all HA resource release completed (standby).    
Jan 22 15:46:02 app2 heartbeat: [8050]: info: Local standby process completed [all].    
Jan 22 15:46:03 app2 kernel: block drbd0: peer( Secondary -> Primary )    
Jan 22 15:46:04 app2 heartbeat: [8050]: WARN: 1 lost packet(s) for [app1] [5137:5139]    
Jan 22 15:46:04 app2 heartbeat: [8050]: info: remote resource transition completed.    
Jan 22 15:46:04 app2 heartbeat: [8050]: info: No pkts missing from app1!    
Jan 22 15:46:04 app2 heartbeat: [8050]: info: Other node completed standby takeover of all resources.

资源从节点：    
# tail -f /var/log/message

Jan 22 15:46:02 app1 heartbeat: [8622]: info: app2 wants to go standby [all]  
Jan 22 15:46:03 app1 kernel: block drbd0: peer( Primary -> Secondary )    
Jan 22 15:46:04 app1 heartbeat: [8622]: info: standby: acquire [all] resources from app2    
Jan 22 15:46:04 app1 heartbeat: [9131]: info: acquire all HA resources (standby).    
Jan 22 15:46:04 app1 ResourceManager(default)[9144]: info: Acquiring resource group: app1 IPaddr::192.168.0.26/24/eth0:1 drbddisk::data Filesystem::/dev/drbd0::/data::ext4 nfs    
Jan 22 15:46:04 app1 /usr/lib/ocf/resource.d//heartbeat/IPaddr(IPaddr_192.168.0.26)[9172]: INFO:  Resource is stopped    
Jan 22 15:46:04 app1 ResourceManager(default)[9144]: info: Running /etc/ha.d/resource.d/IPaddr 192.168.0.26/24/eth0:1 start    
Jan 22 15:46:04 app1 IPaddr(IPaddr_192.168.0.26)[9303]: INFO: Adding inet address 192.168.0.26/24 with broadcast address 192.168.0.255 to device eth0 (with label eth0:1)    
Jan 22 15:46:04 app1 IPaddr(IPaddr_192.168.0.26)[9303]: INFO: Bringing device eth0 up    
Jan 22 15:46:04 app1 IPaddr(IPaddr_192.168.0.26)[9303]: INFO: /usr/libexec/heartbeat/send_arp -i 200 -r 5 -p /var/run/resource-agents/send_arp-192.168.0.26 eth0 192.168.0.26 auto not_used not_used    
Jan 22 15:46:04 app1 /usr/lib/ocf/resource.d//heartbeat/IPaddr(IPaddr_192.168.0.26)[9277]: INFO:  Success    
Jan 22 15:46:04 app1 ResourceManager(default)[9144]: info: Running /etc/ha.d/resource.d/drbddisk data start    
Jan 22 15:46:04 app1 kernel: block drbd0: role( Secondary -> Primary )    
Jan 22 15:46:04 app1 /usr/lib/ocf/resource.d//heartbeat/Filesystem(Filesystem_/dev/drbd0)[9439]: INFO:  Resource is stopped    
Jan 22 15:46:04 app1 ResourceManager(default)[9144]: info: Running /etc/ha.d/resource.d/Filesystem /dev/drbd0 /data ext4 start    
Jan 22 15:46:04 app1 Filesystem(Filesystem_/dev/drbd0)[9529]: INFO: Running start for /dev/drbd0 on /data    
Jan 22 15:46:04 app1 kernel: EXT4-fs (drbd0): mounted filesystem with ordered data mode. Opts:    
Jan 22 15:46:04 app1 /usr/lib/ocf/resource.d//heartbeat/Filesystem(Filesystem_/dev/drbd0)[9518]: INFO:  Success    
Jan 22 15:46:04 app1 kernel: nfsd: last server has exited, flushing export cache    
Jan 22 15:46:05 app1 rpc.mountd[9050]: Caught signal 15, un-registering and exiting.    
Jan 22 15:46:05 app1 rpc.mountd[9698]: Version 1.2.3 starting    
Jan 22 15:46:05 app1 kernel: NFSD: Using /var/lib/nfs/v4recovery as the NFSv4 state recovery directory    
Jan 22 15:46:05 app1 kernel: NFSD: starting 90-second grace period    
Jan 22 15:46:05 app1 heartbeat: [9131]: info: all HA resource acquisition completed (standby).    
Jan 22 15:46:05 app1 heartbeat: [8622]: info: Standby resource acquisition done [all].    
Jan 22 15:46:05 app1 heartbeat: [8622]: info: remote resource transition completed.

[root@app1 resource.d]# ip a    
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 mq state UP qlen 1000    
    link/ether 00:0c:29:03:c8:10 brd ff:ff:ff:ff:ff:ff    
    inet 192.168.0.24/24 brd 192.168.0.255 scope global eth0    
    inet 192.168.0.26/24 brd 192.168.0.255 scope global secondary eth0:1    
    inet6 fe80::20c:29ff:fe03:c810/64 scope link    
       valid_lft forever preferred_lft forever    
3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP qlen 1000    
    link/ether 00:0c:29:03:c8:1a brd ff:ff:ff:ff:ff:ff    
    inet 10.10.10.24/24 brd 10.10.10.255 scope global eth1    
    inet6 fe80::20c:29ff:fe03:c81a/64 scope link    
       valid_lft forever preferred_lft forever

[root@app1 resource.d]# cat /proc/drbd    
version: 8.4.3 (api:1/proto:86-101)    
GIT-hash: 89a294209144b68adb3ee85a73221f964d3ee515 build by gardner@, 2013-11-29 12:28:00    
0: cs:Connected ro:Primary/Secondary ds:UpToDate/UpToDate C r-----    
    ns:628 nr:16 dw:644 dr:2968 al:5 bm:7 lo:0 pe:0 ua:0 ap:0 ep:1 wo:d oos:0

[root@app1 resource.d]# df -h  
Filesystem                   Size  Used Avail Use% Mounted on    
/dev/mapper/vg_app1-lv_root   36G  3.7G   30G  11% /    
tmpfs                       1004M   68K 1004M   1% /dev/shm    
/dev/sda1                    485M   39M  421M   9% /boot    
/dev/drbd0                   9.9G  151M  9.2G   2% /data    
[root@app1 resource.d]#

六，关于Heartbeat+DRBD+方案进一步思考

Heartbeat+DRBD+方案可以实现NFS,MySQL等比较精典的方案，实现方式类似，围绕着Heartbeat、DRBD还有很多的基于主备实用方案。  
如果用于生产环境确实还需要对DRBD进行很好的监控，以及加强对DRBD相对技术的进一步测试与实现，加深对DRBD的理解。    
下一步继续测试DRBD数据迁移、基于heartbeat+共享存储、双主等热备方案都有很大实用价值。