Testing for Availability - Expert Oracle RAC Performance Diagnostics and Tuning - page 78

Step 4
Test all key functional areas of the application in a similar manner. Also test each instance and server for similar failures.
Testing the application by functional areas can give a better control over what and where things need to be fixed.

RAP Phase III—HA
In machine-critical applications, failover of users due to instance or node failures is just one part of the puzzle. There
is something beyond failure of instances, nodes, and services—for example, failure of the data center itself and
media failures—that needs some additional attention and cannot be made transparent unless some customization
and or implementation planning is done. There are several applications that require continuous availability with
minimal or no downtime. Such as applications used by financial trading institutions cannot tolerate downtimes and
for that matter have poor response times of over 6-7 seconds during critical business periods. Applications have
expanded on some of the already available RAC features, which provide automatic failover between two different
locations, synchronizing data using Oracle streams, or golden gate on a real-time basis. Applications such as these
are affected by upgrades, power outages, natural disasters, and so forth. Business requirements for these types of
applications require configuration of disaster recovery (DR) sites. It's important that before the production goes live,
the application and database functionality for DR is also tested and the SLA requirements validated.

Workshop
In this workshop, two tests have to be completed: switchover (or planned failover) and failover from the primary site
to the DR site.

■ the steps required to configure a Dr site is beyond the scope of this topic. in a raC environment, where
application downtime is critical, Dr sites need to be configured with identical hardware, application, and database
configurations.

Note

Step 1—Switchover
Switchover operations are required when the primary database is required to be brought down for hardware and/or
software maintenance. The DBA will perform a manual switchover of the environment from primary to standby/DR.
Switchover is performed using data guard broker (DGB) using the SWITCHOVER TO “<database >” command. To start
the switchover operations, all instances except one should be shut down using the SHUTDOWN IMMEDIATE option. The
last instance should be kept available until such time when both the primary database and standby database have
SYNCHRONIZED the activities. The failover status can be determined using the following query:

Script: MVRACPDnTap_dgsynchcheck.sql
INST INSTN FFS FFCT FFT FFOP OM DUN DBR
---- ------ ------------- ------- --- ----- ---------- ------ ---------
1 SSKY1 SYNCHRONIZED SSKYDG 60 YES READ WRITE SSKYDB PRIMARY
2 SSKY2 SYNCHRONIZED SSKYDG 60 YES READ WRITE SSKYDB PRIMARY