Heartbeat+DRBD+NFS高可用实例配置

opijoiu

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[iyunv@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.or ... ease-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
[iyunv@app1 ~]# cd soft/   
[iyunv@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   
[iyunv@app1 soft]#     
[iyunv@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%]   
[iyunv@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:   
[iyunv@app1 ha.d]# service heartbeat start   
Starting High-Availability services: INFO:  Resource is stopped   
Done.
节点2:   
[iyunv@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：   
[iyunv@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。
[iyunv@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。[iyunv@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:   
[iyunv@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   
[iyunv@app1 drbd.d]#
节点2:   
[iyunv@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   
[iyunv@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：   
[iyunv@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   
[iyunv@app1 drbd.d]#

备节点查看同步状态：   
[iyunv@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   
[iyunv@app2 drbd.d]#

查看同步状态:
[iyunv@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   
[iyunv@app1 ~]#
[iyunv@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。
[iyunv@app1 ~]# mkfs.ext4 /dev/drbd0   
[iyunv@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NFS1. 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
[iyunv@vm15 ~]# mount -t nfs 192.168.0.26:/data/ /mnt   
[iyunv@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   
[iyunv@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.

[iyunv@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
[iyunv@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
[iyunv@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   
[iyunv@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。