01Data backup and disaster recovery

R PO: RPO (Recovery Point Objective):    5 minutes for data recovery point objective

When a disaster occurs, the data can be restored to the state it was 5 minutes ago

R PO=0 When a   disaster occurs, zero data loss

R TO : Application recovery time ( RTO )

RTO (Recovery Time Objective):  5分钟

When a disaster occurs, business can be restored within 5 minutes

RTO = 0   Hyperactive Active – Active  

Disaster recovery  : Generally speaking, disaster recovery is a combination of disaster recovery and backup, that is, local backup combined with remote data replication to achieve complete data protection. Remote replication database file system state snapshot media packaged

DR: disaster recovery system means in places far apart, or to establish two sets of the same functionality as the IT system, with each other may be health monitoring and switching functions, as a system due to accidental (such as fires, earthquakes Etc.) When it stops working, the entire application system can be switched to another place, so that the system function can continue to work normally

Backup: disaster recovery backup are the basis, through often referred to in the data center, all or part of the data  set from the  copy of the host application to a hard disk or other storage array storage medium process

1. Backup mode

1 . Full backup: all data backed up once all  100TB

Frequency: 1 month and a half month, 1 week, 1 day   

2 . Incremental backups

Data from the last backup

3. Differential backup

Data from the full set

image.png 

 

Two, backup media

1 . CD 20G 2003 Nian  CD-ROM 600M 80 Zhang  DVD-ROM 4.7G

2 . Tape the mainstream media backup  TB   stored in off-site slow write once   

3 . Virtual tape library common memory + virtual tape library software   

4 . Cloud scene storage backup to a storage D2D 

D2Tape   D2V VirtualTape   D2V2T   D2D

3. The backup window is  8 hours and  50TB

During the backup period, the impact on the business is within the acceptable range of the customer

Four, backup software

1 . Symantec

NBU: NetworkBackUp FOR Linux Unix

Veritas: windows

2. Commvault Simpana

3 . Love the number of British

5. What to back up?

1 . Data word excel pdf file unstructured data

2 . Business: the Oracle database  table oracle

 10TB single file incremental difference in database / oradata directory  

Can the backup software read the table structure in the database? If one of the tables changes, only this table needs to be backed up, not the database file

Mail Server:

NBU for Oracle Agent

3 . Virtual machine machine: cloud virtual machines

( 1) Treat the virtual machine as a physical machine

( 2) Backup the whole virtual machine

 VMWARE

Huawei:  ebackup   Huawei self-developed virtual machine backup solution

 

Six, backup networking

 image.pngimage.png

 image.png

image.pngimage.png 

 

 

 

 

 

 

 

Seven, install and configure e backup

OceanStor BCManager

OceanStor BCManagerEbackup

OceanStor BCManager Ereplication

 

1. 物理部署

1)安装OSRHEL,SUSEOpenEuler

2)安装ebackup软件

sh  hcp_utilities.sh config

2. 虚拟化部署

使用模板部署

3. 清空防火墙规则

#iptables -F

#service iptables save

4. 通过web界面访问ebackup

image.png

八、备份虚拟机

1. ebackup添加磁盘,用于备份虚拟机

配置存储,在存储上创建LUN10TB,映射给ebackup物理机

备份存储  业务存储  SAS SSD 独立,不要复用

备份存储性能低于业务存储,但容量要大 SATA RAID5 RAID6 顺序IO

备份原理

1. 首次备份,会生成CBT1,系统开始备份

2. 系统自动创建快照,生成一个差分盘,用以存放备份期间产生的数据

3. 内存位图开始记录执行快照以后的数据变更

4. 备份时会将CBT1一并备份到ebackup服务器上

 

 

第二次备份(增量)

1. 系统会将CBT1复制为CBT1’ , CBT1’+内存位图==CBT2

2. 系统比对CBT2CBT1差异,仅备份变更的数据块

3. 备份完成后,内存位图全部清空

第三次备份(增量)

1 . The system will CBT2 replication is C BT2 ', CBT2' + memory bitmap == CBT3

2 . System comparison C BT3 and C BT2 difference, only the changed data blocks backup

C BT is used to record the difference data from the last backup

The differential disk is used to store the data generated during the backup. Once the backup is completed, the differential disk will be automatically merged


Guess you like

Origin blog.51cto.com/14234935/2642359