1 copyright © 2012, oracle and/or its affiliates. all ... · ebay, the no. 1 online auctioneer,...


Oracle最大高可用架构王睿

高级销售顾问


什么是最大高可用架构 Maximum Availability Architecture?


数据

– 停电后的数据损失对你造成怎样的影响?

时间

– 在停电后你能在多短的时间内快速恢复业务?

成本

– 你的所有系统是否存在空闲的组件?

三个维度

It’s all about reducing risks for running IT operations …

高可用需要考虑的问题


Monetary

Impact!

Public Sector

Impact!

Impact on the

Cloud!

为何高可用如此重要? Your Business Is At Stake


任何时间都可能发生数据损坏

随着增加的数据量和复杂的IO子系统，数据损坏变得不可避免

低成本的硬件增加了出问题的可能性 – Schroeder and Gibson, Disk Failures in the Real World: What Does

an MTTF of 1,000,000 Hours Mean to You?, 2007

– Pinheiro, et al, Failure Trends in a Large Drive Population, 2007

– Elerath, Hard Disk Drives: the Good, the Bad, and the Ugly! 2007

– Krioukov, et al, Parity Lost and Parity Regained, 2008

– Bairavasundaram, et al, An Analysis of Data Corruption in the Storage Stack, 2008

– Jiang, et al, Are Disks the Dominant Contributor for Storage Failures? 2008

– Schroeder et al, Understanding Latent Sector Errors and How to Protect Against Them, 2010

除了磁盘损坏，还有更多潜在的数据损坏


eBay

08/11/99 12:00 AM PT, E-Commerce Times

eBay, the No. 1 online auctioneer, had yet another system crash which lasted more than nine hours -- pushing shares of its stock down 10 percent. … This outage comes on the heels of a number of such failures in recent months, including four outages in July and a 21-hour outage on June 10. eBay announced that one will slice a cool $5 million (US$) from its second-quarter revenue

1999年8月11日

eBay，全球最大的网上拍卖商，系统崩溃持续超过9小时-导致股价下跌百分之十。 ...此故障只是最近几个月里发生的故障之一，包括7月份的4次中断和一次在6月10日发生的21小时的停机。eBay宣布，这将导致第二季度的营收减少500万美元

…


申银万国由于硬盘故障导致交易瘫痪

Oracle Confidential

2009年9月16日

故障因小概率硬盘故障导致，16日下午14时25分-14时52分，申银万国全国各地超过100个营业部均发生了故障。


民生银行由于数据库维护失误全行业务瘫痪

Oracle Confidential

2010年2月3日

从上午11:00到下午15:30，由于数据库维护失误，导致全行系统瘫痪持续四个多小时。

虽然该行有灾备系统，但最终“没敢切换”。


星展银行由于存储系统故障导致系统瘫痪

Oracle Confidential

2010年7月5日

由于存储系统的宕机故障使星展银行银行的IT系统在7个小时内无法正常运行，1000

多台自动提款机和银行服务瘫痪长达7小时。

新加坡金融管理局（MAS）谴责星展银行监管不足，造成其银行服务一度瘫痪，并要该行再拨出2.3亿元的营运风险监管资本。


3rd party volume manager

and file system… 3rd party

backup software

3rd party cold

failover cluster

空闲的

备用切换

服务器生产服务器

Identical storage

空闲的灾备系统

3rd party remote mirroring

•大量的空闲组件

•当你需要切换时无法知道备用系统是否正常工作

传统的高可用架构


3rd party volume manager

and file system… 3rd party

backup software

3rd party cold

failover cluster

Idle

Failover

Server Production

Server

Identical storage

Idle DR

3rd party remote mirroring

• Lot of idle components

• Impossible to know if it all works

until you attempt a failover

数据丢失, 时间的浪费, 系统的闲置!

Risky!!

传统的高可用架构


最新高可用架构的要求 Oracle 用户

需求解决方案

断电保护任何类型、任何时间

典型的恢复时间从秒级到分钟级别

测试的频度任何时间

典型的数据丢失无数据丢失或者秒级别

实施的复杂程度容易

堆栈整合的需要预先集成的

ROI “Show me the money!”


Oracle 高可用架构的设计原则

完整最小化所有计划内和计划外的停机时间

标准的,经过验证的最大高可用架构

面向应用的保护并能恢复应用的对象

能够实现应用的在线更改

可扩展的横向扩张的计算和存储资源

由智能软件实现架构中所有组件都是可用

集成和易实施内置的高可用性与可插拔组件；

自动的 – 消除手工操作


RAC – Scalability

– Server HA

ASM – Volume Management Online Redefinition,

Edition-based Redefinition,

Data Guard, GoldenGate – Minimal downtime maintenance,

upgrades, and migrations

生产环境

Active Data Guard – Data Protection, DR

– Query Offload

GoldenGate – Active-active

– Heterogeneous

Oracle Secure Backup – Backup to tape / cloud

活动备份环境

Flashback – Human error

correction

RMAN & Fast Recovery Area – On-disk backups

Oracle最大高可用架构(MAA) 低成本, 集成, 双活, 高的 ROI


Online Redefinition,

Edition-based Redefinition,

Data Guard, GoldenGate – Minimal downtime maintenance,

upgrades, and migrations

Production Site

RAC – Scalability

– Server HA

Flashback – Human error

correction

Active Data Guard – Data Protection, DR

– Query Offload

GoldenGate – Active-active

– Heterogeneous

Active Replica

Oracle Secure Backup – Backup to tape / cloud

ASM – Volume Management

RMAN & Fast Recovery Area – On-disk backups

MAA With Exadata Low-Cost, Integrated, Fully Active, High ROI


集成最大高可用架构的软件

Flexible

Maintenance

and

Migrations

Media and

Storage

Failures

Human or

Application

Error

Database,

System,

Cluster, Site,

and

Geographic

Outages

Server

Failure

•Storage failure

•Data recovery

•Backups

Automatic

Storage

Management

Recovery

Manager

(RMAN)

Oracle Secure

Backup

• Instance failure

•Server failure

•Rolling

maintenance

•Performance

scale-out

•Online instance

relocation

•Consolidation

Oracle RAC

•Fast point-in-

time recovery

•Granular

repair of

logical

corruptions

•Transaction

•Table

•Database

Flashback

•Database failure

•System failure

•Site failure

•Zero data loss

•Automatic failover

•Corruption

protection

•Rolling upgrade

•Read-only offload

•Backup offload

Active Data Guard

同时解决计划内和计划外的停机

•Heterogeneous

migrations

•Bi-directional

and multi-master

replication

•Zero downtime

maintenance

GoldenGate

Edition Based

Redefinition

•Zero downtime

application

upgrades


Oracle数据库的高可用架构


Oracle 数据库集成高可用解决方案

计划外停机

计划内停机

Real Application Clusters (RAC)

Flashback

RMAN & Oracle Secure Backup

ASM

Data Guard & GoldenGate

Online Reconfiguration

Rolling Upgrades

Edition-based Redefinition

Ora

cle

MA

A B

est P

ractic

es

Online Redefinition

系统变更

应用变更

数据变更

服务器高可用性

数据高可用性


服务器的高可用性和扩展

计划外停机

计划内停机

Real Application Clusters (RAC)

系统变更

应用变更

数据变更


数据高可用性


HR SALES FINANCE

• 通过低成本的服务器来扩展工作负载

• 运行所有的Oracle应用程序

• 将数据库整合到少量的服务器和数据库中

• 内置的高可用性来支持关键的工作负载

真正应用集群(RAC)


数据的高可用性

Flashback

RMAN & Oracle Secure Backup

ASM

Data Guard & GoldenGate

Protection from

Data Corruptions

Protection from

Storage / Site Failures

Enabling Active-

Active Data Centers

Protection from

Human Errors

计划外停机


数据高可用性


革命性改进错误修复

– 查看过去某个时间点的”好的” 数据

纠正错误的时间等于产生错误的时间，与数据大小无关

Simple: SQL> flashback database to <timestamp>;

Flexible: Flashback Query, Table, Transaction, Database, Drop

0

20

40

60

Correction Time = Error Time + f(DB_SIZE)

闪回技术 Fast, Granular Error-Correction


Recovery Manager (RMAN): Oracle数据库备份恢复的基础架构

RMAN is aware of database file formats &

best recovery procedures, supporting:

– 数据块的校验

– 在线块级恢复

– 未使用块的重复数据删除

– 在线的多通道的备份

– 本机的加密和压缩

与磁盘备份集成: Fast Recovery Area

集成的磁带和云备份:

Oracle Secure Backup (OSB)

Oracle Enterprise

Manager

RMAN

Data Files

Fast Recovery

Area

Tape Drive

Oracle Secure

Backup

Cloud

(Amazon S3)

备份& 恢复


Oracle 安全备份(OSB) 集中式磁带管理备份

RMAN – Oracle Recovery Manager, MEB – MySQL Enterprise Backup, SBT – Oracle’s API for integration with media managers

Protects Entire IT

Environment

• Oracle 数据库从11g到9i

• 比同类产品快25 – 40%

• 可以备份MySQL Enterprise Edition

• 支持异构的文件系统(UNIX/ Linux /

Windows) and NAS devices

• Built-in Oracle Integration

• 在分布式环境中进行集中管理

• 比同类的竞争产品便宜75%


ASM 能支持所有类型的文件– database files, filesystems, Clusterware files (OCR, Voting Disk)

内置的镜像保护机制来保护磁盘实效

利用镜像副本自动修复坏块

3rd Party FS Application

Automatic Storage Management (ASM)

ASM Cluster & Single Node File System (ACFS)

Database

ACFS Snapshot

ASM Disk

Group

DB Datafiles, OCR and Voting Files Oracle Binaries 3rd Party File Systems

Dynamic Volume Manager

ASM Instance Managing

Oracle DB Files

自动存储管理ASM


Data Guard Broker

Enterprise Manager Cloud Control

专门针对Oracle数据库的数据高可用和数据保护方案

在单一的配置环境中最多可以配置30个备用库

备用库可以用于查询、报表和备份等只读操作

Active Data Guard

备份数据库生产数据库

Sync / Async

Redo Transport

Oracle Data Guard 最佳数据保护方案


• Data Guard通过网络直接从SGA中传递 redo blocks 类似于memcpy ；

• Redo 的接收和应用是通过运行的Oracle实例: 持续的数据库集成和数据的验证

System Memory (SGA)

Oracle

Database

Architecture To Standby

Databases

TCP/IP

1.Better performance since no disk I/O

2.Better isolation from lower layer faults

3.Better network utilization: only redo sent

4.Transactional consistency: always

5.Corrupted blocks auto-repaired

6.Database-integrated application failover

数据库的集成复制 Data Guard: Why A Big Deal


1. 内置在Oracle数据库融合事务的一致性

2. 超高的复制性能

3. 集成的应用程序故障切换

4. 透明操作，支持所有数据库特性和数据类型

5. Combined HA/DR solution

6. 确保故障隔离，防止数据损坏

7. 确保零数据库丢失

8. 灾备服务器可以用于实时的操作

9. 同时解决计划内和计划外的停机

10. 不需要绑定存储供应商

11. 最小的网络带宽占用

12. 没有距离限制

LAN & MAN deployments provide Local HA and DR

Extend to a Wide Area Network and add remote DR

Data Guard: 高可用基础架构


Real-time

Queries

Active Standby Database Production Database

Continuous Redo

Shipment and Apply Offload Fast

Incremental

Backups

卸载只读操作

• Automatically monitor

and respond to query lag

• DML redirection

Active Data Guard 保护灾备设备的投资


App 1

App n

Active Data Guard 1

App 2

App 3

Primary Database ADG 2

ADG 3

ADG 4

…

SYNC

ASYNC

Load

Balancer

Up to 30

standby

databases

Scale out for

high look-to-

book ratio

Active Data Guard: Reader Farm Scale-Out


自动块修复

– 当Oracle检测到在生产库上有坏块产生，就会自动利用灾备库上的好的数据块进行修复 (& 反之亦然)

– 修复过程对用户和应用程序是透明的

Active Standby

Database

Primary

Database

Read/Write

Workload

Continuous redo shipping, validation & apply

Real-time

Reporting

Active Data Guard: 自动块修复通过自动块修复提供数据的高可用性


存储镜像: 冗余的存储用于存储实效的保护：

– 不能保护数据的物理损坏

– 灾备系统在镜像期间是离线的 – 不能提供实时的数据验证

– Limited distance, storage vendor lock-in, manual failover, no rolling upgrades, high network use

See: http://www.oracle.com/technetwork/database/features/availability/dataguardremotemirroring-086151.html

Database

Instance

Database

Storage

Primary Site: All

Servers Active

DR Site:

All Servers

Inactive

Storage Mirroring

采用存储镜像的数据保护


Network I/O

Much more

network volume

Much more

network impact

Zero Oracle

awareness

Poor fault

isolation

Idle standby

systems

Also: database block corruptions propagated to target

Target

Volumes Primary

Volumes

Log

Buffer

Online Logs

Archive Logs

Flashback Logs

Control Files

Data Files

SYSTEM

USER

TEMP

UNDO

存储镜像的缺点


Online

Logs

Archive Logs

Flashback Logs

Control Files

Data Files

SYSTEM

USER

TEMP

UNDO

Log

Buffer

Also: standby database protected from block corruptions

Active

Standby

Database

Storage agnostic

Minimal

network impact

Automatic

block repair

Strong fault

isolation

Active standby

systems

Primary

Database

Redo block transmission Oracle Apply

& Validation

Data Guard 的优势


• 业界最佳的实时数据集成解决方案

• 是现有Oracle产品的一个补充

• 对异构的系统提供持续的可用性

• 为报表、BI/EPM等系统提供实时数据抽取

• 为数据库、硬件和应用提供零宕机时间的迁移升级等

• 全球超过1200个客户，包括世界500强的用户，用户遍布金融服务、电信、健康卫生、公共服务和零售等行业

• 4 of top 5 largest commercial banks

• 3 of top 3 busiest ATM networks

• 5 of top 5 financial data services companies

• 5 of top 5 telecommunications providers

• 3 of top 5 largest food & drug stores

Oracle GoldenGate

http://brandsoftheworld.com/brands/0002/6850/Liberty_Mutual.eps

http://images.google.com/imgres?imgurl=http://hrff.jalmc.org/_Images/News/adp.jpg&imgrefurl=http://www.ortop.org/sd.htm&h=671&w=1500&sz=74&hl=en&start=1&um=1&tbnid=lSxuKOYqXblBlM:&tbnh=67&tbnw=150&prev=/images?q=adp&imgsz=xxlarge&svnum=10&um=1&hl=en&rlz=1T4GZHY_enUS230US230&sa=N

http://www.google.com/imgres?imgurl=http://www.androidguys.com/wp-content/uploads/2011/01/sprint-logo.jpg&imgrefurl=http://www.androidguys.com/2011/01/18/sprint-add-10-fee-smart-phones-effective-january-30th/sprint-logo/&h=325&w=600&sz=94&tbnid=4KxUWz0_-gpywM:&tbnh=73&tbnw=135&prev=/search?q=Sprint+logo&tbm=isch&tbo=u&zoom=1&q=Sprint+logo&usg=__fOAuD9MkRB0yPYgzD3mJWRvnMS8=&sa=X&ei=avRqTuS1JqvciALEkoCTDg&ved=0CBQQ9QEwAA

http://www.siriusxm.com/wo/i/logodownload/img/logo.eps

http://www.google.com/imgres?imgurl=http://logok.org/wp-content/uploads/2010/07/dell_logo.jpg&imgrefurl=http://logok.org/dell/&h=480&w=640&sz=33&tbnid=vDCkrslBpqF2fM:&tbnh=103&tbnw=137&prev=/search?q=Dell+logo&tbm=isch&tbo=u&zoom=1&q=Dell+logo&usg=__nk3s3cLpJsfrEW05IseCXCl1_bA=&sa=X&ei=5_RqTt6HCePZiALb0PC-Dg&ved=0CBgQ9QEwAg

http://www.google.com/imgres?imgurl=http://www.credit-card-logos.com/images/visa_credit-card-logos/visa_logo_6.gif&imgrefurl=http://www.credit-card-logos.com/&h=113&w=180&sz=2&tbnid=6RgFXMad8jPZ6M:&tbnh=63&tbnw=101&prev=/search?q=Visa+logo&tbm=isch&tbo=u&zoom=1&q=Visa+logo&usg=__gUcOb1jLcLWi-BDQpHJ5scq4BUo=&sa=X&ei=DfVqTq_PDonXiAL6k9DBDg&ved=0CBUQ9QEwAQ

http://www.google.com/imgres?imgurl=http://upload.wikimedia.org/wikipedia/en/7/76/Mayo-clinic-logo.png&imgrefurl=http://en.wikipedia.org/wiki/File:Mayo-clinic-logo.png&h=300&w=300&sz=19&tbnid=A5QcsnkrPLNbvM:&tbnh=116&tbnw=116&prev=/search?q=mayo+clinic+logo&tbm=isch&tbo=u&zoom=1&q=mayo+clinic+logo&usg=__SrXRqCQTmrVAupL7I6lKwG4Nex0=&sa=X&ei=vfVqToP1BY_RiAKctqWhDg&ved=0CBoQ9QEwAg

http://www.google.com/imgres?imgurl=http://www.kokkada.com/wp-content/images/2009/04/bse-logo.jpg&imgrefurl=http://www.kokkada.com/sensex-bombay-stock-exchange-bse/&h=288&w=432&sz=31&tbnid=NlkoKyI0y0z2CM:&tbnh=84&tbnw=126&prev=/search?q=bombay+stock+exchange+logo&tbm=isch&tbo=u&zoom=1&q=bombay+stock+exchange+logo&usg=__xDyibMVps8n3-4Qr6pGr0LkVtxc=&sa=X&ei=B_ZqToD9N-nXiAKLjZnGDg&ved=0CBoQ9QEwAw

http://www.google.com/imgres?imgurl=http://news.walgreens.com/images/20007/wag_logo2.gif&imgrefurl=http://news.walgreens.com/photo+archive/logos/&h=58&w=229&sz=2&tbnid=WY8d8gw2DxRGXM:&tbnh=27&tbnw=108&prev=/search?q=walgreens+logo&tbm=isch&tbo=u&zoom=1&q=walgreens+logo&usg=__YxNNh5bwFk3VPCwMW13a0WwyXyM=&sa=X&ei=XfZqTobII9LWiALdjcWmDg&ved=0CCoQ9QEwBA

http://www.google.com/imgres?imgurl=http://www.meddutch.com/sites/default/files/pictures/faa.gif&imgrefurl=http://www.meddutch.com/?q=content/faa-medicals&h=998&w=967&sz=312&tbnid=uukb5SmuS8jCaM:&tbnh=149&tbnw=144&prev=/search?q=FAA+logo&tbm=isch&tbo=u&zoom=1&q=FAA+logo&usg=__q7RPSPNK3OV4Y5Y-dOMkBySBIgw=&sa=X&ei=t_ZqTvvSIKfYiAL52-DTDg&ved=0CBUQ9QEwAQ

http://www.google.com/imgres?imgurl=http://www.yourlogoresources.com/wp-content/uploads/2011/08/walmart-logo.png&imgrefurl=http://www.yourlogoresources.com/walmart-logo/&h=787&w=2846&sz=70&tbnid=38GNj6G-fd-PLM:&tbnh=41&tbnw=150&prev=/search?q=Walmart+logo&tbm=isch&tbo=u&zoom=1&q=Walmart+logo&usg=__XIZZd_QfnsMe8X81zdMNviqsJGo=&sa=X&ei=pfdqTtv_CMniiALb-Ki-Dg&ved=0CBcQ9QEwAg


Capture: committed transactions are captured (and can be filtered) as they occur by reading redo logs.

Trail: stages and queues data for routing.

Pump: distributes data for routing to target(s).

Route: data is compressed, encrypted for routing to target(s).

Delivery: applies data with transaction integrity,

transforming the data as required.

Source

Oracle & Non-Oracle Database(s) Target

Oracle & Non-Oracle Database(s)

Capture

Delivery

Trail

Files

Pump

Trail

Files Pump

Delivery

Capture

Bi-directional

LAN / WAN / Internet

Over TCP/IP

Trail

Files

Trail

Files

GoldenGate: 双活的复制机制


38

GoldenGate: 支持异构的环境

数据库操作系统或平台

Oracle GoldenGate Capture and Delivery:

Oracle

DB2

Microsoft SQL Server

Sybase ASE

Teradata

Enscribe

SQL/MP

SQL/MX

MySQL v 5.5

JMS message queues

Oracle GoldenGate Delivery:

All listed above, plus:

TimesTen, Postgres

Netezza, Greenplum, HP Neoview

ETL products

Linux

Oracle Solaris 11

Windows 2000, 2003, XP

HP NonStop

HP-UX

IBM AIX v7.1

IBM z Series

zLinux

IBM i Series


GoldenGate

- Information Distribution

- Heterogeneous

Bi-directional

Replication Subsetting MySQL

Standby

Database

Active Data Guard

- DR & Data Protection

- Real-time Query

Primary

Database

RAC

- Scalability

- Server HA

强强联合集成的高可用性、数据保护和双活中心


灾备/ 数据保护

• Active Data Guard:完整的Oracle数据库保护、只读的备份库

Oracle内置的高效、简便的灾备解决方案

Zero data loss,零数据库丢失、数据保护，提供, switchover / failover功能；

为各种数据类型各个应用和负载提供灾备

双活中心, 维护、迁移、数据集成、非Oracle数据库

• GoldenGate:, 双活中心、迁移、异构、目标库的可读写

所有节点都是可用(read/write): 需要规避访问和管理的冲突

可以通过建立第二套数据库的方式来实现零宕机时间的数据库和应用的迁移

支持异构系统

Active Data Guard & GoldenGate: 补充说明


在线计划内维护

计划外停机

计划内停机

Online Reconfiguration

Rolling Upgrades

Edition-based Redefinition

Online Redefinition

系统变更

应用变更

数据变更


数据高可用性


服务器

– 在线的增加或删除RAC节点

– 无需数据的迁移

存储

– 在线增加ASM磁盘和磁盘组

– 自动的rebalance

Clusterware, ASM

– 在线的升级Cluster和ASM

Database

Storage

Online Reconfiguration 根据需求的扩展


在线升级或打补丁

简单的一次性补丁可以应用在运行的Oracle实例上

– Linux-x86, Solaris 10, HP-UX 11i

– Windows 32-bit and 64-bit, AIX v6.1 [TL2 SP1]

多数更复杂的补丁可以用过RAC滚动升级

数据版本和补丁集可以通过Data Guard或者Golden Gate

来进行升级– 具体参考Note 1265700.1

完全的数据中心或者SAN的迁移可以利用Data Guard /

GoldenGate来实现最小宕机时间的迁移


所有的索引操作都可以实现在线

表的重新定义或者重新组织可以利用with package DBMS_REDEFINITION

– 更改位置、表类型、分区、列、列类型

– 可以在复制过程中对内容进行转换

Source Table

Update

Tracking

Transform Copy

Table

Transform

Updates

Result Table

Continuous

Queries &

Updates

Store Updates

在线表的重定义


WAN

• Comprehensive protection from failures: server, storage, network, site, corruptions

• Protection from human errors: database, table, row, transaction

• Active DR: Real-time remote standby open for query offload

• Online indexing and table redefinition

• Online patching and upgrades

Real Application Clusters

ASM

RMAN / Fast

Recovery Area

Active Data Guard /

GoldenGate

Oracle Secure

Backup

MAA With Exadata Complete, Integrated Availability


Oracle最大可用架构的用户成功案例

MAA Best Practices at Bank of America

Oracle Footprint

• 1000’s of Oracle Databases

• Oracle Database Enterprise Edition, Enterprise Manager Grid

Control, E-Business Suite, Oracle CRM, Hyperion, Fusion

Middleware

• Oracle Database 11g Release 2

– The bank’s core RDBMS version as of July 1

– 11gR2 upgrade ramping now, 10% of databases upgraded

Seven Application Tiers

• Tiers 0 and 1

– Full MAA configuration

• Tiers 2 through 6

– Single instance MAA configurations (non-RAC)

• Business continuity implemented for all tiers

– Data Guard is implemented on 99% of all production databases

– GoldenGate is used for multi-master configurations

• e.g. one critical in-house application with 3-way replication

between continents

Conclusion

• It would be difficult for the Bank of America to meet its

availability and data protection requirements without the

Oracle Maximum Availability Architecture.


Solution

• GoldenGate between Siebel 7.8 and Siebel 8.0 to enable a zero downtime Siebel upgrade

• GoldenGate configured with application upgrade & downgrade logic to enable both systems to be in continuous synchronization

Return on Investment

• Zero database downtime

• Reduced user downtime

• Mitigated risk to the business by having prior

Siebel environment in-sync in case of fallback

Business Challenges

1. Highly critical Siebel CRM call center application

2. 15,000 call center agents per day, 25TB+ database

3. Upgrade both with no downtime while maintaining fallback solution

利用GoldenGate 实现Siebel的零停机时间的升级


成功案例：United Parcel Service (UPS) 使因升级而导致的停机时间减少 93%

UPS 的任务关键航班操作和日程安排应用程序：

– 世界上第 9 大航空公司

– 为全球 800 个机场提供服务。

计划停机时间 SLA — 低于 15 分钟

– Oracle 数据库 11g 升级测试显示: 传统的升级方法无法满足 UPS SLA

Data Guard 滚动升级

– 总应用程序停机时间低于 4 分钟

– 升级备用数据库，进行切换，然后升级原始主数据库

– 应用程序在切换期间之外的所有时间内均可用

23

0

10

20

30

40

50

60

Min

ute

s

10.2.0.3 to 10.2.0.4 10.2.0.4 to 11.1.0.6

Traditional Upgrade Rolling Upgrade

总应用程序停机时间

15

4

55

4

Data Guard（主数据库和备用数据库相距 800 英里）、RMAN、闪回数据库、Streams


中华人民共和国海关总署

远距离容灾

低带宽

低延时

可随时接管的双业务中心

快速业务切换

Oracle 9i

RAC

Oracle 9i

RAC

Oracle 9i

RAC

决策支持库

GoldenGate单向复制

GoldenGate单向复制

H2000生产库

Oracle 9i

全国各地41个口岸H2000应用

GoldenGate双向复制

广州北京

H2000生产库

决策支持库

网络


MAA Deployments on Exadata

PayPal: Exadata MAA for Mission Critical OLTP

55

GoldenGate Real-time Data Integration *

Extreme Performance

• 300+K executions/sec

• Real Time analysis of 99.99% of

critical transactions.

• Avg 40 ms response for 99.99%

• 10 X performance compared to

pre-Exadata system

HA and MAA

• 99.99% Availability

• All disk groups using high

redundancy

• Active Data Guard for auto block

corruption repair and DR

• Rolling upgrade using ASM,

CRS, Data Guard, and

GoldenGate

• Flashback enabled

Mission-critical

Databases

Primary Data Center

DR Data Center

Data Guard ASYNC Redo Transport

ETL

Targets Production Databases • 2 X Exadata X2-8

• 2 X Full Storage Expansion

Active Data Guard Standby • Offload queries and reads

• Corruption Protection

• Symmetric System

WAN, 650+ miles (30ms)

Test/Dev

Production and Standby Clusters = 8 Exadata Racks

3 identical Architectures = 24 Exadata Racks + Test/Dev Resources supporting our Critical Applications.


OTN HA Portal:

http://www.oracle.com/goto/availability

Maximum Availability Architecture (MAA):

http://www.oracle.com/goto/maa

MAA Blogs:

http://blogs.oracle.com/maa

Exadata on OTN:

http://www.oracle.com/technetwork/database/exadata/index.html

Oracle HA Customer Success Stories on OTN:

http://www.oracle.com/technetwork/database/features/ha-casestudies-

098033.html

Resources


Q & A

1 copyright © 2012, oracle and/or its affiliates. all ... · ebay, the no. 1 online auctioneer,...

Documents