# 灾难恢复规划
[root@dr-server ~]# cat > /root/dr-plan.txt << 'EOF' 灾难恢复规划 ============ 1. RTO/RPO定义 - RTO（恢复时间目标）：系统恢复所需时间 - RPO（恢复点目标）：可接受的数据丢失量 2. 灾难恢复级别 - 级别1：实时复制，RTO<1小时，RPO=0 - 级别2：异步复制，RTO<4小时，RPO<1小时 - 级别3：定期备份，RTO<24小时，RPO<24小时 3. 灾难恢复架构 - 同城双活：两个数据中心同时提供服务 - 同城灾备：主备模式，备用数据中心待命 - 异地灾备：跨地域数据复制 4. 故障切换流程 - 故障检测 - 业务评估 - 决策切换 - 执行切换 - 业务验证 5. 回切流程 - 数据同步 - 业务验证 - 决策回切 - 执行回切 EOF # 创建灾难恢复目录 [root@dr-server ~]# mkdir -p /dr/{scripts,logs,config}

# 配置DRBD复制
[root@primary ~]# dnf install -y drbd drbd-utils

# 配置DRBD资源
[root@primary ~]# cat > /etc/drbd.d/r0.res << 'EOF' resource r0 { protocol C; meta-disk internal; device /dev/drbd0; disk /dev/sdb1; net { cram-hmac-alg sha1; shared-secret "drbd-secret"; } on primary.fgedu.net.cn { address 192.168.1.10:7789; } on secondary.fgedu.net.cn { address 192.168.1.11:7789; } } EOF # 初始化DRBD [root@primary ~]# drbdadm create-md r0 initializing activity log initializing bitmap (320 KB) New drbd meta data block successfully created. [root@secondary ~]# drbdadm create-md r0 initializing activity log initializing bitmap (320 KB) New drbd meta data block successfully created. # 启动DRBD [root@primary ~]# drbdadm up r0 [root@secondary ~]# drbdadm up r0 # 设置主节点 [root@primary ~]# drbdadm primary --force r0 # 查看状态 [root@primary ~]# drbdadm status r0 r0 role:Primary disk:UpToDate secondary.fgedu.net.cn role:Secondary replication:Established peer-disk:UpToDate # 格式化并挂载 [root@primary ~]# mkfs.xfs /dev/drbd0 [root@primary ~]# mount /dev/drbd0 /data # 测试数据同步 [root@primary ~]# echo "test data" > /data/test.txt
[root@primary ~]# cat /data/test.txt
test data

# 切换到备节点
[root@primary ~]# umount /data
[root@primary ~]# drbdadm secondary r0
[root@secondary ~]# drbdadm primary r0
[root@secondary ~]# mount /dev/drbd0 /data
[root@secondary ~]# cat /data/test.学习交流加群风哥QQ113257174txt
test data

# 安装Pacemaker和Corosync
[root@primary ~]# dnf install -y pacemaker corosync pcs

# 设置密码
[root@primary ~]# echo “hacluster” | passwd –stdin hacluster
Changing password for user hacluster.
passwd: all authentication tokens updated successfully.

# 启动服务
[root@primary ~]# systemctl enable –now pcsd
Created symlink /etc/systemd/system/multi-user.更多视频教程www.fgedu.ne学习交流加群风哥微信: itpux-co更多学习教程公众号风哥教程itpux_commt.cntarget.wants/pcsd.service → /usr/lib/systemd/system/pcsd.service.

# 认证节点
[root@primary ~]# pcs cluster auth primary secondary -u hacluster -p hacluster
primary: Authorized
secondary: Authorized

# 创建集群
[root@primary ~]# pcs cluster setup dr-cluster primary secondary
Destroying cluster on nodes: primary, secondary…
primary: Stopping Cluster (pacemaker)…
primary: Stopping Cluster (corosync)…
secondary: Stopping Cluster (pacemaker)…
secondary: Stopping Cluster (corosync)…
primary: Successfully destroyed cluster
secondary: Successfully destroyed cluster

Sending ‘pacemaker Remote auth’ to ‘primary’, ‘secondary’
primary: successful distribution of the file ‘pacemaker Remote auth’
secondary: successful distribution of the file ‘pacemaker Remote auth’
Sending cluster config files to the nodes…
primary: Succeeded
secondary: Succeeded

Synchronizing pcsd SSL certificates on nodes primary, secondary…
primary: Success
secondary: Success
Restarting pcsd on the nodes in order to reload the SSL certificates…
primary: Success
secondary: Success

# 启动集群
[root@primary ~]# pcs cluster start –all
primary: Starting Cluster (corosync)…
secondary: Starting Cluster (corosync)…
primary: Starting Cluster (pacemaker)…
secondary: Starting Cluster (pacemaker)…

# 查看集群状态
[root@primary ~]# pcs status cluster
Cluster Status:
Cluster Summary:
* Stack: corosync
* Current DC: primary (version 2.1.2-4.el9) – partition with quorum
* Last updated: Fri Apr 4 22:20:00 2026
* Last change: Fri Apr 4 22:19:00 2026 by hacluster via crmd on primary
* 2 nodes configured
* 0 resource instances configured

# 禁用STONITH
[root@primary ~]# pcs property set stonith-enabled=false

# 创建DRBD资源
[root@primary ~]# pcs resource create drbd_r0 ocf:linbit:drbd drbd_resource=r0 op monitor interval=30s

# 创建主从资源
[root@primary ~]# pcs resource promotable drbd_r0 promoted-max=1 promoted-node-max=1 clone-max=2 clone-node-max=1 notify=true

# 创建文件系统资源
[root@primary ~]# pcs resource create fs_data ocf:heartbeat:Filesystem device=”/dev/drbd0″ directory=”/data” fstype=”xfs”

# 设置约束
[root@primary ~]# pcs constraint colocation add fs_data with drbd_r0-clone INFINITY with-rsc-role=Promoted
[root@primary ~]# pcs constraint order promote drbd_r0-clone then start fs_data

# 创建虚拟IP资源
[root@primary ~]# pcs resource create vip ocf:heartbeat:IPaddr2 ip=192.168.1.100 cidr_netmask=24 op monitor interval=30s

# 设置约束
[root@primary ~]# pcs constraint colocation add vip with fs_data INFINITY
[root@primary ~]# pcs constraint order fs_data then vip

# 查看资源状态
[root@primary ~]# pcs status resources
* Clone Set: drbd_r0-clone [drbd_r0] (promotable):
* Promoted: [ primary ]
* Unpromoted: [ secondary ]
* fs_data (ocf:heartbeat:Filesystem): Started primary
* vip (ocf:heartbeat:IPaddr2): Started primary

# 手动切换资源
[root@primary ~]# pcs cluster standby primary
[root@primary ~]# pcs status resources
* Clone Set: drbd_r0-clone [drbd_r0] (promotable):
* Promoted: [ secondary ]
* Unpromoted: [ primary ]
* fs_data (ocf:heartbeat:Filesystem): Started secondary
* vip (ocf:heartbeat:IPaddr2): Started secondary

# 恢复节点
[root@primary ~]# pcs cluster unstandby primary
[root@primary ~]# pcs status resources
* Clone Set: drbd_r0-clone [drbd_r0] (promotable):
* Promoted: [ primary ]
* Unpromoted: [ secondary ]
* fs_data (ocf:heartbeat:Filesystem): Started primary
* vip (ocf:heartbeat:IPaddr2): Started primary

# 创建故障切换脚本
[root@primary ~]# cat > /dr/scripts/failover.sh << 'EOF' #!/bin/bash ACTION=$1 NODE=$2 case $ACTION in failover) echo "Failing over from $NODE..." pcs cluster standby $NODE sleep 5 pcs status resources ;; failback) echo "Failing back to $NODE..." pcs cluster unstandby $NODE sleep 5 pcs status resources ;; status) pcs status resources ;; *) echo "Usage: $0 {failover|failback|status} [node]" exit 1 ;; esac EOF [root@primary ~]# chmod +x /dr/scripts/failover.sh # 测试故障切换 [root@primary ~]# /dr/scripts/failover.sh failover primary Failing over from primary... primary: successfully placed node in standby mode * Clone Set: drbd_r0-clone [drbd_r0] (promotable): * Promoted: [ secondary ] * Unpromoted: [ primary ] * fs_data (ocf:heartbeat:Filesystem): Started secondary * vip (ocf:heartbeat:IPaddr2): Started secondary