MFS+Keepalived双机高可用热备方案操作记录

 基于MFS的单点及手动备份的缺陷,考虑将其与Keepalived相结合以提高可用性。在Centos下MooseFS(MFS)分布式存储共享环境部署记录这篇文档部署环境的基础上,只需要做如下改动:

1)将master-server作为Keepalived_MASTER(启动mfsmaster、mfscgiserv)2)将matelogger作为Keepalived_BACKUP(启动mfsmaster、mfscgiserv)3)将ChunkServer服务器里配置的MASTER_HOST参数值改为VIP地址4)clinet挂载的master的ip地址改为VIP地址按照这样调整后,需要将Keepalived_MASTER和Keepalived_BACKUP里面的hosts绑定信息也修改下。

方案实施原理及思路

1)mfsmaster的故障恢复在1.6.5版本后可以由mfsmetalogger产生的日志文件changelog_ml.*.mfs和metadata.mfs.back文件通过命令mfsmetarestore恢复2)定时从mfsmaster获取metadata.mfs.back 文件用于master恢复3)Keepalived MASTER检测到mfsmaster进程宕停时会执行监控脚本,即自动启动mfsmaster进程,如果启动失败,则会强制kill掉keepalived和mfscgiserv进程,由此转移VIP到BACKUP上面。4)若是Keepalived MASTER故障恢复,则会将VIP资源从BACKUP一方强制抢夺回来,继而由它提供服务5)整个切换在2~5秒内完成 根据检测时间间隔。

架构拓扑图如下:

1)Keepalived_MASTER(mfs master)机器上的操作

mfs的master日志服务器的安装配置已经在另一篇文档中详细记录了,这里就不做赘述了。下面直接记录下keepalived的安装配置:-----------------------------------------------------------------------------------------------------------------------安装Keepalived[root@Keepalived_MASTER ~]# yum install -y openssl-devel popt-devel[root@Keepalived_MASTER ~]# cd /usr/local/src/[root@Keepalived_MASTER src]# wget http://www.keepalived.org/software/keepalived-1.3.5.tar.gz[root@Keepalived_MASTER src]# tar -zvxf keepalived-1.3.5.tar.gz[root@Keepalived_MASTER src]# cd keepalived-1.3.5[root@Keepalived_MASTER keepalived-1.3.5]# ./configure --prefix=/usr/local/keepalived[root@Keepalived_MASTER keepalived-1.3.5]# make && make install     [root@Keepalived_MASTER keepalived-1.3.5]# cp /usr/local/src/keepalived-1.3.5/keepalived/etc/init.d/keepalived /etc/rc.d/init.d/[root@Keepalived_MASTER keepalived-1.3.5]# cp /usr/local/keepalived/etc/sysconfig/keepalived /etc/sysconfig/[root@Keepalived_MASTER keepalived-1.3.5]# mkdir /etc/keepalived/[root@Keepalived_MASTER keepalived-1.3.5]# cp /usr/local/keepalived/etc/keepalived/keepalived.conf /etc/keepalived/[root@Keepalived_MASTER keepalived-1.3.5]# cp /usr/local/keepalived/sbin/keepalived /usr/sbin/[root@Keepalived_MASTER keepalived-1.3.5]# echo "/etc/init.d/keepalived start" >> /etc/rc.local      [root@Keepalived_MASTER keepalived-1.3.5]# chmod +x /etc/rc.d/init.d/keepalived      #添加执行权限[root@Keepalived_MASTER keepalived-1.3.5]# chkconfig keepalived on                   #设置开机启动[root@Keepalived_MASTER keepalived-1.3.5]# service keepalived start                   #启动[root@Keepalived_MASTER keepalived-1.3.5]# service keepalived stop                    #关闭[root@Keepalived_MASTER keepalived-1.3.5]# service keepalived restart                 #重启    配置Keepalived[root@Keepalived_MASTER ~]# cp /etc/keepalived/keepalived.conf /etc/keepalived/keepalived.conf-bak[root@Keepalived_MASTER ~]# vim /etc/keepalived/keepalived.conf! Configuration File for keepalivedglobal_defs {  notification_email {    root@localhost    }  notification_email_from keepalived@localhostsmtp_server 127.0.0.1smtp_connect_timeout 30router_id MFS_HA_MASTER}  vrrp_script chk_mfs {                             script "/usr/local/mfs/keepalived_check_mfsmaster.sh"  interval 2  weight 2}  vrrp_instance VI_1 {  state MASTER  interface eth0  virtual_router_id 51  priority 100  advert_int 1  authentication {    auth_type PASS    auth_pass 1111}  track_script {    chk_mfs}virtual_ipaddress {    182.48.115.239}notify_master "/etc/keepalived/clean_arp.sh 182.48.115.239"}        接着编写监控脚本[root@Keepalived_MASTER ~]# vim /usr/local/mfs/keepalived_check_mfsmaster.sh#!/bin/bashA=`ps -C mfsmaster --no-header | wc -l`if [ $A -eq 0 ];then/etc/init.d/mfsmaster startsleep 3   if [ `ps -C mfsmaster --no-header | wc -l ` -eq 0 ];then      /usr/bin/killall -9 mfscgiserv      /usr/bin/killall -9 keepalived   fifi[root@Keepalived_MASTER ~]# chmod 755 /usr/local/mfs/keepalived_check_mfsmaster.sh设置更新虚拟服务器(VIP)地址的arp记录到网关脚本[root@Keepalived_MASTER ~]# vim /etc/keepalived/clean_arp.sh #!/bin/shVIP=$1GATEWAY=182.48.115.254                                        //这个是网关地址                       /sbin/arping -I eth0 -c 5 -s $VIP $GATEWAY &>/dev/null[root@Keepalived_MASTER ~]# chmod 755 /etc/keepalived/clean_arp.sh    启动keepalived(确保Keepalived_MASTER机器的mfs master服务和Keepalived服务都要启动)[root@Keepalived_MASTER ~]# /etc/init.d/keepalived startStarting keepalived:                                       [  OK  ][root@Keepalived_MASTER ~]# ps -ef|grep keepalivedroot     28718     1  0 13:09 ?        00:00:00 keepalived -Droot     28720 28718  0 13:09 ?        00:00:00 keepalived -Droot     28721 28718  0 13:09 ?        00:00:00 keepalived -Droot     28763 27466  0 13:09 pts/0    00:00:00 grep keepalived    查看vip[root@Keepalived_MASTER ~]# ip addr1: 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000    link/ether 52:54:00:09:21:60 brd ff:ff:ff:ff:ff:ff    inet 182.48.115.233/27 brd 182.48.115.255 scope global eth0    inet 182.48.115.239/32 scope global eth0    inet6 fe80::5054:ff:fe09:2160/64 scope link       valid_lft forever preferred_lft forever

2)Keepalived_BACKUP(mfs master)机器上的操作

在另一篇文档里,该机器是作为metalogger元数据日志服务器的,那么在这个高可用环境下,该台机器改为Keepalived_BACKUP使用。即去掉metalogger的部署,直接部署mfs master(部署过程参考另一篇文档)。下面直接说下Keepalived_BACKUP下的Keepalived配置: 安装Keepalived[root@Keepalived_BACKUP ~]# yum install -y openssl-devel popt-devel[root@Keepalived_BACKUP ~]# cd /usr/local/src/[root@Keepalived_BACKUP src]# wget http://www.keepalived.org/software/keepalived-1.3.5.tar.gz[root@Keepalived_BACKUP src]# tar -zvxf keepalived-1.3.5.tar.gz[root@Keepalived_BACKUP src]# cd keepalived-1.3.5[root@Keepalived_BACKUP keepalived-1.3.5]# ./configure --prefix=/usr/local/keepalived[root@Keepalived_BACKUP keepalived-1.3.5]# make && make install      [root@Keepalived_BACKUP keepalived-1.3.5]# cp /usr/local/src/keepalived-1.3.5/keepalived/etc/init.d/keepalived /etc/rc.d/init.d/[root@Keepalived_BACKUP keepalived-1.3.5]# cp /usr/local/keepalived/etc/sysconfig/keepalived /etc/sysconfig/[root@Keepalived_BACKUP keepalived-1.3.5]# mkdir /etc/keepalived/[root@Keepalived_BACKUP keepalived-1.3.5]# cp /usr/local/keepalived/etc/keepalived/keepalived.conf /etc/keepalived/[root@Keepalived_BACKUP keepalived-1.3.5]# cp /usr/local/keepalived/sbin/keepalived /usr/sbin/[root@Keepalived_BACKUP keepalived-1.3.5]# echo "/etc/init.d/keepalived start" >> /etc/rc.local       [root@Keepalived_BACKUP keepalived-1.3.5]# chmod +x /etc/rc.d/init.d/keepalived      #添加执行权限[root@Keepalived_BACKUP keepalived-1.3.5]# chkconfig keepalived on                   #设置开机启动[root@Keepalived_BACKUP keepalived-1.3.5]# service keepalived start                   #启动[root@Keepalived_BACKUP keepalived-1.3.5]# service keepalived stop                    #关闭[root@Keepalived_BACKUP keepalived-1.3.5]# service keepalived restart                 #重启     配置Keepalived[root@Keepalived_BACKUP ~]# cp /etc/keepalived/keepalived.conf /etc/keepalived/keepalived.conf-bak[root@Keepalived_BACKUP ~]# vim /etc/keepalived/keepalived.conf! Configuration File for keepalivedglobal_defs {  notification_email {    root@localhost    }  notification_email_from keepalived@localhostsmtp_server 127.0.0.1smtp_connect_timeout 30router_id MFS_HA_BACKUP}  vrrp_script chk_mfs {                             script "/usr/local/mfs/keepalived_check_mfsmaster.sh"  interval 2  weight 2}  vrrp_instance VI_1 {  state BACKUP  interface eth0  virtual_router_id 51  priority 99  advert_int 1  authentication {    auth_type PASS    auth_pass 1111}  track_script {    chk_mfs}virtual_ipaddress {    182.48.115.239}notify_master "/etc/keepalived/clean_arp.sh 182.48.115.239"}  接着编写监控脚本[root@Keepalived_BACKUP ~]# vim /usr/local/mfs/keepalived_notify.sh#!/bin/bashA=`ps -C mfsmaster --no-header | wc -l`if [ $A -eq 0 ];then/etc/init.d/mfsmaster startsleep 3   if [ `ps -C mfsmaster --no-header | wc -l ` -eq 0 ];then      /usr/bin/killall -9 mfscgiserv      /usr/bin/killall -9 keepalived   fifi [root@Keepalived_BACKUP ~]# chmod 755 /usr/local/mfs/keepalived_notify.sh设置更新虚拟服务器(VIP)地址的arp记录到网关脚本(Haproxy_Keepalived_Master 和 Haproxy_Keepalived_Backup两台机器都要操作)[root@Keepalived_BACKUP ~]# vim /etc/keepalived/clean_arp.sh #!/bin/shVIP=$1GATEWAY=182.48.115.254                                      /sbin/arping -I eth0 -c 5 -s $VIP $GATEWAY &>/dev/null 启动keepalived[root@Keepalived_BACKUP ~]# /etc/init.d/keepalived startStarting keepalived:[root@Keepalived_BACKUP ~]# ps -ef|grep keepalivedroot     17565     1  0 11:06 ?        00:00:00 keepalived -Droot     17567 17565  0 11:06 ?        00:00:00 keepalived -Droot     17568 17565  0 11:06 ?        00:00:00 keepalived -Droot     17778 17718  0 13:47 pts/1    00:00:00 grep keepalived 要保证Keepalived_BACKUP机器的mfs master服务和keepalived服务都要启动!

3)chunkServer的配置

只需要将mfschunkserver.cfg文件中的MASTER_HOST参数配置成182.48.115.239,即VIP地址。其他的配置都不需要修改。然后重启mfschunkserver服务

4)clinet客户端的配置

只需要见挂载命令中的元数据ip改为182.48.115.249即可![root@clinet-server ~]# mkdir /mnt/mfs                [root@clinet-server ~]# mkdir /mnt/mfsmeta [root@clinet-server ~]# /usr/local/mfs/bin/mfsmount /mnt/mfs -H 182.48.115.239mfsmaster accepted connection with parameters: read-write,restricted_ip,admin ; root mapped to root:root[root@clinet-server ~]# /usr/local/mfs/bin/mfsmount -m /mnt/mfsmeta -H 182.48.115.239mfsmaster accepted connection with parameters: read-write,restricted_ip[root@clinet-server ~]# df -hFilesystem            Size  Used Avail Use% Mounted on/dev/mapper/VolGroup-lv_root                      8.3G  3.8G  4.1G  49% /tmpfs                 499M  228K  498M   1% /dev/shm/dev/vda1             477M   35M  418M   8% /boot/dev/sr0              3.7G  3.7G     0 100% /media/CentOS_6.8_Final182.48.115.239:9421   107G   42G   66G  39% /mnt/mfs[root@clinet-server ~]# mount /dev/mapper/VolGroup-lv_root on / type ext4 (rw)proc on /proc type proc (rw)sysfs on /sys type sysfs (rw)devpts on /dev/pts type devpts (rw,gid=5,mode=620)tmpfs on /dev/shm type tmpfs (rw,rootcontext="system_u:object_r:tmpfs_t:s0")/dev/vda1 on /boot type ext4 (rw)none on /proc/sys/fs/binfmt_misc type binfmt_misc (rw)gvfs-fuse-daemon on /root/.gvfs type fuse.gvfs-fuse-daemon (rw,nosuid,nodev)/dev/sr0 on /media/CentOS_6.8_Final type iso9660 (ro,nosuid,nodev,uhelper=udisks,uid=0,gid=0,iocharset=utf8,mode=0400,dmode=0500)182.48.115.239:9421 on /mnt/mfs type fuse.mfs (rw,nosuid,nodev,allow_other)182.48.115.239:9421 on /mnt/mfsmeta type fuse.mfsmeta (rw,nosuid,nodev,allow_other)验证下客户端挂载MFS文件系统后的数据读写是否正常[root@clinet-server ~]# cd /mnt/mfs[root@clinet-server mfs]# echo "12312313" > test.txt[root@clinet-server mfs]# cat test.txt12312313[root@clinet-server mfs]# rm -f test.txt [root@clinet-server mfs]# cd ../mfsmeta/trash/[root@clinet-server trash]# find . -name "*test*"./003/00000003|test.txt[root@clinet-server trash]# cd ./003/[root@clinet-server 003]# ls00000003|test.txt  undel[root@clinet-server 003]# mv 00000003\|test.txt undel/[root@clinet-server 003]# ls /mnt/mfstest.txt[root@clinet-server 003]# cat /mnt/mfs/test.txt 12312313以上说明挂载后的MFS数据共享正常

5)Keepalived_MASTER和Keepalived_BACKUP的iptales防火墙设置

Keepalived_MASTER和Keepalived_BACKUP的防火墙iptables在实验中是关闭的。如果开启了iptables防火墙功能,则需要在两台机器的iptables里配置如下:可以使用"ss -l"命令查看本机监听的端口[root@Keepalived_MASTER ~]# ss -lState       Recv-Q Send-Q                                     Local Address:Port                                         Peer Address:Port   LISTEN      0      100                                                    *:9419                                                    *:*       LISTEN      0      100                                                    *:9420                                                    *:*       LISTEN      0      100                                                    *:9421                                                    *:*       LISTEN      0      50                                                     *:9425                                                    *:*       LISTEN      0      128                                                   :::ssh                                                    :::*       LISTEN      0      128                                                    *:ssh                                                     *:*       LISTEN      0      100                                                  ::1:smtp                                                   :::*       LISTEN      0      100                                            127.0.0.1:smtp                                                    *:*    [root@Keepalived_MASTER ~]# vim /etc/sysconfig/iptables........-A INPUT -s 182.148.15.0/24 -d 224.0.0.18 -j ACCEPT       #允许组播地址通信。注意设置这两行,就会在Keepalived_MASTER故障恢复后,将VIP资源从Keepalived_BACK那里再转移回来-A INPUT -s 182.148.15.0/24 -p vrrp -j ACCEPT             #允许VRRP(虚拟路由器冗余协)通信-A INPUT -m state --state NEW -m tcp -p tcp --dport 9419 -j ACCEPT -A INPUT -m state --state NEW -m tcp -p tcp --dport 9420 -j ACCEPT  -A INPUT -m state --state NEW -m tcp -p tcp --dport 9421 -j ACCEPT-A INPUT -m state --state NEW -m tcp -p tcp --dport 9425 -j ACCEPT [root@Keepalived_MASTER ~]# /etc/init.d/iptables start

6)故障切换后的数据同步脚本

上面的配置可以实现Keepalived_MASTER机器出现故障(keepalived服务关闭),VIP资源转移到Keepalived_BACKUP上;当Keepalived_MASTER机器故障恢复(即keepalived服务开启),那么它就会将VIP资源再次抢夺回来!但是只是实现了VIP资源的转移,但是MFS文件系统的数据该如何进行同步呢?下面在两台机器上分别写了数据同步脚本(Keepalived_MASTER和Keepalived_BACKUP要提前做好双方的ssh无密码登陆的信任关系)Keepalived_MASTER机器上[root@Keepalived_MASTER ~]# vim /usr/local/mfs/MFS_DATA_Sync.sh #!/bin/bashA=`ip addr|grep 182.48.115.239|awk -F" " '{print $2}'|cut -d"/" -f1`if [ $A == 182.48.115.239 ];then   /etc/init.d/mfsmaster stop    /bin/rm -f /usr/local/mfs/var/mfs/*   /usr/bin/rsync -e "ssh -p22" -avpgolr 182.48.115.235:/usr/local/mfs/var/mfs/* /usr/local/mfs/var/mfs/   /usr/local/mfs/sbin/mfsmetarestore -m   /etc/init.d/mfsmaster -a   sleep 3   echo "this server has become the master of MFS"   if [ $A != 182.48.115.239 ];then   echo "this server is still MFS's slave"   fifiKeepalived_BACKUP机器上[root@Keepalived_BACKUP ~]# vim /usr/local/mfs/MFS_DATA_Sync.sh #!/bin/bashA=`ip addr|grep 182.48.115.239|awk -F" " '{print $2}'|cut -d"/" -f1`if [ $A == 182.48.115.239 ];then   /etc/init.d/mfsmaster stop    /bin/rm -f /usr/local/mfs/var/mfs/*   /usr/bin/rsync -e "ssh -p22" -avpgolr 182.48.115.233:/usr/local/mfs/var/mfs/* /usr/local/mfs/var/mfs/   /usr/local/mfs/sbin/mfsmetarestore -m   /etc/init.d/mfsmaster -a   sleep 3   echo "this server has become the master of MFS"   if [ $A != 182.48.115.239 ];then   echo "this server is still MFS's slave"   fifi即当VIP资源转移到自己这一方时,执行这个同步脚本,就会将对方的数据同步过来了。

7)故障切换测试

1)关闭Keepalived_MASTER的mfsmaster服务由于keepalived.conf文件中的监控脚本定义,当发现mfsmaster进程不存在时,就会主动启动mfsmaster。只要当mfsmaster启动失败,才会强制killall掉keepalived和mfscgiserv进程[root@Keepalived_MASTER ~]# /etc/init.d/mfsmaster stopsending SIGTERM to lock owner (pid:29266)waiting for termination terminated发现mfsmaster关闭后,会自动重启[root@Keepalived_MASTER ~]# ps -ef|grep mfsroot     26579     1  0 16:00 ?        00:00:00 /usr/bin/python /usr/local/mfs/sbin/mfscgiserv startroot     30389 30388  0 17:18 ?        00:00:00 /bin/bash /usr/local/mfs/keepalived_check_mfsmaster.shmfs      30395     1 71 17:18 ?        00:00:00 /etc/init.d/mfsmaster start默认情况下,VIP资源是在Keepalived_MASTER上的[root@Keepalived_MASTER ~]# ip addr1: 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000    link/ether 52:54:00:09:21:60 brd ff:ff:ff:ff:ff:ff    inet 182.48.115.233/27 brd 182.48.115.255 scope global eth0    inet 182.48.115.239/32 scope global eth0    inet6 fe80::5054:ff:fe09:2160/64 scope link        valid_lft forever preferred_lft forever在client端挂载后(通过VIP地址挂载),查看数据[root@clinet-server ~]# cd /mnt/mfs[root@clinet-server mfs]# lltotal 4-rw-r--r-- 1 root root 9 May 24 17:11 grace-rw-r--r-- 1 root root 9 May 24 17:11 grace1-rw-r--r-- 1 root root 9 May 24 17:11 grace2-rw-r--r-- 1 root root 9 May 24 17:11 grace3-rw-r--r-- 1 root root 9 May 24 17:10 kevin-rw-r--r-- 1 root root 9 May 24 17:10 kevin1-rw-r--r-- 1 root root 9 May 24 17:10 kevin2-rw-r--r-- 1 root root 9 May 24 17:10 kevin3当keepalived关闭(这个时候mfsmaster关闭后就不会自动重启了,因为keepalived关闭了,监控脚本就不会执行了)[root@Keepalived_MASTER ~]# /etc/init.d/keepalived stopStopping keepalived:                                       [  OK  ][root@Keepalived_MASTER ~]# ip addr1: 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000    link/ether 52:54:00:09:21:60 brd ff:ff:ff:ff:ff:ff    inet 182.48.115.233/27 brd 182.48.115.255 scope global eth0    inet6 fe80::5054:ff:fe09:2160/64 scope link        valid_lft forever preferred_lft forever发现,Keepalived_MASTER的keepalived关闭后,VIP资源就不在它上面了。查看系统日志,发现VIP已经转移[root@Keepalived_MASTER ~]# tail -1000 /var/log/messages.......May 24 17:11:19 centos6-node1 Keepalived_vrrp[29184]: VRRP_Instance(VI_1) Sending/queueing gratuitous ARPs on eth0 for 182.48.115.239May 24 17:11:19 centos6-node1 Keepalived_vrrp[29184]: Sending gratuitous ARP on eth0 for 182.48.115.239May 24 17:11:19 centos6-node1 Keepalived_vrrp[29184]: Sending gratuitous ARP on eth0 for 182.48.115.239May 24 17:11:19 centos6-node1 Keepalived_vrrp[29184]: Sending gratuitous ARP on eth0 for 182.48.115.239May 24 17:11:19 centos6-node1 Keepalived_vrrp[29184]: Sending gratuitous ARP on eth0 for 182.48.115.239然后到Keepalived_BACKUP上面发现,VIP已经过来了[root@Keepalived_BACKUP ~]# ip addr1: 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000    link/ether 52:54:00:82:69:69 brd ff:ff:ff:ff:ff:ff    inet 182.48.115.235/27 brd 182.48.115.255 scope global eth0    inet 182.48.115.239/32 scope global eth0    inet6 fe80::5054:ff:fe82:6969/64 scope link        valid_lft forever preferred_lft forever查看日志,也能看到VIP转移过来了[root@Keepalived_BACKUP ~]# tail -1000 /var/log/messages.......May 24 17:27:57 centos6-node2 Keepalived_vrrp[5254]: VRRP_Instance(VI_1) Received advert with higher priority 100, ours 99May 24 17:27:57 centos6-node2 Keepalived_vrrp[5254]: VRRP_Instance(VI_1) Entering BACKUP STATEMay 24 17:27:57 centos6-node2 Keepalived_vrrp[5254]: VRRP_Instance(VI_1) removing protocol VIPs.再次在clinet客户端挂载后,查看数据[root@clinet-server mfs]# ll[root@clinet-server mfs]#发现没有数据,这就需要执行上面的那个同步脚本[root@Keepalived_BACKUP ~]# sh -x /usr/local/mfs/MFS_DATA_Sync.sh 再次在clinet客户端挂载后,查看数据[root@clinet-server mfs]# lltotal 4-rw-r--r--. 1 root root 9 May 24 17:11 grace-rw-r--r--. 1 root root 9 May 24 17:11 grace1-rw-r--r--. 1 root root 9 May 24 17:11 grace2-rw-r--r--. 1 root root 9 May 24 17:11 grace3-rw-r--r--. 1 root root 9 May 24 17:10 kevin-rw-r--r--. 1 root root 9 May 24 17:10 kevin1-rw-r--r--. 1 root root 9 May 24 17:10 kevin2-rw-r--r--. 1 root root 9 May 24 17:10 kevin3发现数据已经同步过来,然后再更新数据[root@clinet-server mfs]# rm -f ./*[root@clinet-server mfs]# echo "123123" > wangshibo[root@clinet-server mfs]# echo "123123" > wangshibo1[root@clinet-server mfs]# echo "123123" > wangshibo2[root@clinet-server mfs]# echo "123123" > wangshibo3[root@clinet-server mfs]# echo "123123" > wangshibo4[root@clinet-server mfs]# lltotal 3-rw-r--r--. 1 root root 7 May 24 17:26 wangshibo-rw-r--r--. 1 root root 7 May 24 17:26 wangshibo1-rw-r--r--. 1 root root 7 May 24 17:26 wangshibo2-rw-r--r--. 1 root root 7 May 24 17:26 wangshibo3-rw-r--r--. 1 root root 7 May 24 17:26 wangshibo42)恢复Keepalived_MASTER的keepalived进程[root@Keepalived_MASTER ~]# /etc/init.d/keepalived startStarting keepalived:                                       [  OK  ][root@Keepalived_MASTER ~]# ip addr1: 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000    link/ether 52:54:00:09:21:60 brd ff:ff:ff:ff:ff:ff    inet 182.48.115.233/27 brd 182.48.115.255 scope global eth0    inet 182.48.115.239/32 scope global eth0    inet6 fe80::5054:ff:fe09:2160/64 scope link        valid_lft forever preferred_lft forever发现Keepalived_MASTER的keepalived进程启动后,VIP资源又抢夺回来。查看/var/log/messages日志能看出VIP资源转移回来再次查看Keepalived_BACKUP,发现VIP不在了[root@Keepalived_BACKUP ~]# ip addr1: 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000    link/ether 52:54:00:82:69:69 brd ff:ff:ff:ff:ff:ff    inet 182.48.115.235/27 brd 182.48.115.255 scope global eth0    inet6 fe80::5054:ff:fe82:6969/64 scope link        valid_lft forever preferred_lft forever在client端挂载后(通过VIP地址挂载),查看数据[root@clinet-server mfs]# lltotal 4-rw-r--r--. 1 root root 9 May 24 17:11 grace-rw-r--r--. 1 root root 9 May 24 17:11 grace1-rw-r--r--. 1 root root 9 May 24 17:11 grace2-rw-r--r--. 1 root root 9 May 24 17:11 grace3-rw-r--r--. 1 root root 9 May 24 17:10 kevin-rw-r--r--. 1 root root 9 May 24 17:10 kevin1-rw-r--r--. 1 root root 9 May 24 17:10 kevin2-rw-r--r--. 1 root root 9 May 24 17:10 kevin3发现数据还是旧的,需要在Keepalived_MASTER上执行同步[root@Keepalived_MASTER ~]# sh -x /usr/local/mfs/MFS_DATA_Sync.sh 再次在clinet客户端挂载后,查看数据,发现已经同步[root@xqsj_web3 mfs]# lltotal 3-rw-r--r-- 1 root root 7 May 24 17:26 wangshibo-rw-r--r-- 1 root root 7 May 24 17:26 wangshibo1-rw-r--r-- 1 root root 7 May 24 17:26 wangshibo2-rw-r--r-- 1 root root 7 May 24 17:26 wangshibo3-rw-r--r-- 1 root root 7 May 24 17:26 wangshibo4
©著作权归作者所有:来自51CTO博客作者80民工的原创作品,如需转载,请注明出处,否则将追究法律责任