1. Local high-availability solutions
2. Live-active solution in the same city
3. The master end of the master-standby solution is disconnected, and the slave end needs to manually activate the disaster recovery software
Active-active solution
1. Configure two storage
Primary storage
1 . Create-drive field, storage pools, L UN , L UN group, host, group, mapping
2 . Host the I the SCSI attached storage, partitioning, formatting, loading, write data
Slave storage
1 . Create a hard domain, storage pool, other content will be synced over from the main terminal storage
Two, configure the arbitration server
1 . Install the software arbitration
2 . Configuration arbitration services
3 . Configuration seki
Machine-to-machine account:
User: m m_user
Password: mm_user@storage
Three, install and configure the business host
Fourth, install multi-path software
5. Test (disconnect one of the data centers or one storage)
upadmin set hypermetro workingmode=priority primary_array_id=0
The default is priority mode, that is, data will be written to the priority site, and the priority site will be synchronized to the non-priority site
upadmin set hypermetro workingmode=balance primary_array_id=0
Set to balanced mode, that is, data will be double-written to two storages at the same time
Active-active: Two L UNs become active-active L UNs . After being merged through multipath on the host side, it will be a VLUN to the user.
Remote Copy: Only the master can be mapped to the host, not the mapping from the end, once the primary-side fault, from the need to manually L UN mapped to the host
Online migration (hot migration)
Virtualization platform FusionCompute VMWARE Vsphere
cloud platform
Traditional data center physical environment scenarios
P2V technology physical machine migration to virtual machine platform or cloud platform
V 2V technology migrates business from one virtualization platform to another virtualization platform
OpenStack
1 . Virtualization platform or a cloud, to install a corresponding version of the virtual machine
Centos6.5 MySQL
C entOS6.5 + MySQL + import data
Use MySQL's own migration tool for migration
2. Rainbow migration solution
Physical machine migration to F usionCompute
Migrate the physical machine to O penstack
V2V
迁移流程
1. 安装迁移服务器,Windows平台
2. 在虚拟化平台上准备一台空的虚拟机 CentOS7.6系统
3. 开始迁移
4. 同步数据
5. 迁移完成,验证业务
winPE.iso PE光盘
LiveCD.iso 用来启动Linux
迁移步骤:
1. 开始迁移任务,系统将ISO(共享)文件挂载到空虚拟机上,使用ISO启动虚拟机(LiveCD)
2. 给虚拟机配置IP地址(手动规划)
3. 将虚拟机的硬盘分区,格式化,挂载到liveCD /mnt文件夹上
4. 使用scp传输文件,将源主机数据通过ssh复制到/mnt/vda1 /mnt/mapper/rhel-root文件中(实际就是目标虚拟机的磁盘)
5. 生成引导程序(grub2-install /dev/vda)
6. 卸载光盘,使用本地硬盘启动
7. 测试
问题一、源主机停止业务了吗?
问题二: 源主机和目标主机IP地址冲突了吗?
问题三: 目标主机的CPU和内存需要与源主机一致吗?
问题四: 目标主机磁盘必须大于或等于源主机吗?
块级迁移: 目标必须大于或等于源磁盘
文件级迁移: 目标大于或者等于源磁盘已使用空间
Linux一切皆文件
windows 选择块级别迁移
问题五:源主机需要安装代理吗?
windows需要安装代理,Linux不需要,为什么?
Linux使用ssh协议
问题六:如果迁移失败怎么办? 还能回退吗?
问题七:迁移过程中未停止业务,迁移过程中产生了新数据怎么办?
rsync 是否安装
同步
第一步: 在线将源端数据复制到目标端 (未停止业务) 10个小时 5G 30分钟
第二步: 同步一次 (30分钟)(未停止业务)
第三步: 停止业务,最终同步一次 2分钟
第四步: 修改目标主机的配置(比如IP地址等),启动服务,客户端验证
主机名
DNS
www.example.com A 192.168.20.119
www.example.com A 192.168.12.66
迁移windows流程
UDP: 137 138
TCP: 139 445
1. 给源主机安装agent,监听8899端口
2. 开始迁移任务,系统将ISO(共享)文件挂载到空虚拟机上,使用ISO启动虚拟机(winPE.iso)
3. 给虚拟机配置IP地址(手动规划)
4. 开启卷影副本功能,使用fastcopy将数据块复制到目标虚拟机,源主机8899端口,TCP 139 TCP 445端口
5. 生成引导程序
6. 卸载光盘,使用本地硬盘启动
7. 测试
卷影副本技术
Rainbow不支持 windowsXP windows2003 迁移
windows7 windows2008
Synchronous use is the f astcopy tool to synchronize data blocks
3. Migrate to the cloud platform
1 . Download L inuxImage.qcow2 file
2 . Uploading the file to G Lance in
3 . Use G Lance Image issuing host cloud, cloud directly distributed to the cloud host hard disk
4 . Source host and a target host communication cloud, the source and destination end via ssh file transfer
5 . Use hard disk to start the cloud
Question 8: Can the virtual machine in v mware virtualization be migrated to f usioncompute ? ?
P2V
V 2V
Unified as a physical machine migration
Migrate from v mware to Huawei