Database Reference
In-Depth Information
RAC resource mastering is performed only on the Hub node instances, not on a Leaf node. This ensures that
the failure of a Leaf node does not require remastering and also ensures that masters have affinity to the Hub node
instances.
Oracle RAC and Oracle Clusterware also work together to allow application connections to perform seamless
failover from the failed instance to the surviving instance. The applications can use these technologies to implement
smooth failover for their database operations such as query or transactions. These technologies include:
1.
Transparent Application Failover (TAF)
2.
Fast Connect Failover (FCF)
3.
Better Business continuity & HA using Oracle 12c Application Continuity (AC)
Transparent Application Failover (TAF)
Transparent Application Failover (TAF) is a feature that helps database connection sessions fail over to a surviving
database instance during an instance failure. This is a client-side failover. With this feature, you can specify how to
fail over the session and re-establish the session on another instance, and how the query of the original database
connection continues after the connection gets relocated to the new database instance. It should be mentioned that
only a query operation such as a select statement gets replayed after the connection is relocated to the new database
instance. However, active transaction operations such as DML statements will not be failed over and replayed, as TAF
can't preserve these active transactions. During an instance failure, these transactions will be failed and rolled back,
and the application will receive an error message about the transaction failure.
The configuration of TAF is done through the tnsname.ora file on the client side without a need for any
application code change.
KHDB_Sales =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = kr720n-scan)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME =
khdb_sales.dblab.com
)
(FAILOVER_MODE =
(TYPE=session)
(METHOD=basic)
(RETRIES=10)
(DELAY=10)
) ))
The failover mode is specified by the TYPE parameter, with three possible values: “session” for the session
mode; “select” for the select mode; “none” for deactivating the failover. The session mode is a basic configuration.
In this mode, TAF just reconnects the session to another instance and the new session has to start over again. In the
select mode, TAF reconnects the session to another instance and allows the query to reuse the open cursor from the
previous session. You can deactivate the failover if you put “none” or just skip the FAILOVER_MODE clause.
The options for the failover method are “basic” or “preconnect.” Using the basic failover method, TAF
re-establishes the connection to another instance only after the instance failure. In the preconnect method, the
application preconnects a session to a backup instance. This will speed up failover and thus avoid the huge sudden
reconnection storm that may happen to the surviving instance in the basic failover method. This is especially serious
in a two-node RAC, where all the connections on the failed instance will need to be reconnected to the only surviving
instance during an instance failure.
