曝光台 注意防骗
网曝天猫店富美金盛家居专营店坑蒙拐骗欺诈消费者
• External, lightweight
“Observer” process ensures
failover rules are satisfied
and eliminates the possibility
of “split brain” processing
• Managed through Data Guard
Broker & Oracle Enterprise
Manager Grid Control
2
systems is essential. Any disruption in system availability directly impacts
Amazon’s bottom line.
Historically, Amazon has used custom-maintained standby databases to protect
business operations and data that would otherwise be affected by system and sitewide
failures. Amazon has built-in automation through custom-developed scripts
and processes that enable business continuity even in the event of complete system
or site failures. Amazon has determined that automation is critical to meeting
aggressive service level agreements.
Fast-Start Failover is designed to meet requirements such as those at Amazon.
Fast-Start Failover quickly and reliably fails over a target standby database to the
primary database role without requiring manual intervention to invoke the failover.
Continually seeking to enhance their high availability architecture, Amazon
evaluated Fast-Start Failover during early beta testing of Oracle Data Guard 10g
Release 2.
Amazon determined that Fast-Start Failover satisfied their High Availability
requirements. For the first time, an out-of-the-box solution is available that
Amazon can use instead of supporting and maintaining custom scripts and
processes. Amazon’s findings were presented at Oracle Open World in
September, 2005. They are:
• Fast-Start Failover reliably executes fast, automatic failover to a standby
site without human intervention.
• Fast-Start Failover reduced failover time from minutes to seconds.
Average failover time was 25 seconds based on a Data Guard
primary/standby pair with round-trip network latency of 0.5ms.
• Fast-Start Failover achieves zero data loss. Failover will not commence
automatically if the failover target is not synchronized with the primary
database.
• Because Fast-Start Failover is based on Data Guard Maximum Availability
protection mode, the primary production database is not affected by
network or standby failures.
• After failover the old primary database can be automatically reinstated as a
new target standby (assuming that the database can be restarted),
following which, Data Guard quickly and automatically resynchronizes
the old primary with the new primary database. It no longer needs to be
restored from a backup of the new primary.
The remainder of this paper provides an overview of Fast-Start Failover. A
detailed discussion of Fast-Start Failover is provided in the Maximum Availability
Architecture (MAA) [2] best practices paper Fast-Start Failover: Oracle Database
10g Release 2 [3] and documentation for Oracle Data Guard Broker [4], and
Oracle Data Guard Concepts and Administration [5].
“The capability of fast,
guaranteed zero-data-loss
failover with Fast-Start
Failover in Oracle Data Guard
takes the availability of an
Oracle database platform to
new levels. Our initial tests
running Oracle Database 10g
Release 2, show that Fast-Start
Failover offers a magnitude of
improvement in availability.”
—Rajesh Sheth,
Manager, Database
Engineering
Amazon.com
3
OVERVIEW: FAST-START FAILOVER
There are three essential participants (figure 1) in a Fast-Start Failover
configuration:
• The primary database
• The target standby database
• The Fast-Start Failover Observer
The target standby database will become the new primary database following a
fast-start failover (note there can be multiple standby databases in a Data Guard
configuration).
Fast-Start Failover is used in a Data Guard configuration under the control of the
Data Guard Broker. The Data Guard Broker provides centralized management of
all resources within a Data Guard configuration.
Fast-Start Failover is managed using the Data Guard Broker command line
interface, DGMGRL, or Oracle Enterprise Manager 10g Grid Control. Enterprise
Manager provides a GUI interface that interacts with the Data Guard Broker. Not
only does this provide a GUI interface for monitoring and control, it also enables
centralized management of resources in one or more Data Guard configurations.
The Observer is a separate process incorporated into the DGMGRL client that
continuously monitors the primary database and the target standby database for
possible failure conditions. Installing the Observer is simple – all that is required is
Primary Site Standby Site
中国航空网 www.aero.cn
航空翻译 www.aviation.cn
本文链接地址：
航空资料10(2)
