Optimize Distributed Workload - Expert Oracle RAC Performance Diagnostics and Tuning - page 496

■ instance failover is only available with the policy-managed configuration. it is not available if the database is
configured with the admin-managed option.

Note

As illustrated in Figure 15-4 , all components depend on the NODE service. For the DATABASE application, the
instances and all services created on the instance will be affected should the DATABASE service fail. For an INSTANCE ,
all services are affected.
How do you monitor and track when these events are fired? Oracle has provided options where a script or
utility or application (called server side callouts), if placed in the $GRID_HOME/racg/usrco directory, will be executed
automatically. For example, the following shell script, when placed in the directory, will write out events generated by
the event manager to the directory specified in the script:

[oracle@prddb2 oracle]$ more callout.sh
#!/bin/ksh
FAN_LOGFILE=$GRID_HOME/racg/log/fan_log_'hostname'.log
echo $* "reported="'date' >> $FAN_LOGFILE &

Similarly, any script or executable that could be invoked automatically when an ONS event condition occurs can
be located in this directory and will be executed. A few example events from a Oracle Database 12c implementation
(formatted for clarity) generated by the preceding callout script includes the following:

DOWN event • RDBMS instance

INSTANCE VERSION=1.0 service=scdb database=scdb instance=SCDB_2 host=ssky1l1p1
status=down reason=USER timestamp=2014-02-04 00:39:07 timezone=-05:00
db_domain= reported=date

UP event

• RDBMS instance

INSTANCE VERSION=1.0 service=scdb database=scdb instance=SCDB_2 host=ssky1l1p1 status=up
reason=USER timestamp=2014-02-04 00:44:12 timezone=-05:00 db_domain= reported=date

DATABASE service UP event •

DATABASE VERSION=1.0 service=scdb database=scdb instance= host=ssky1l1p1 status=up
reason=USER timestamp=2014-02-04 00:39:07 timezone=-05:00 db_domain= reported=date

DATABASE service DOWN event •

DATABASE VERSION=1.0 service=scdb database=scdb instance= host=ssky1l1p1 status=down
reason=USER timestamp=2014-02-04 00:39:07 timezone=-05:00 db_domain= reported=date

SERVICEMEMBER UP event •

SERVICEMEMBER VERSION=1.0 service=TAPS database=scdb instance=SCDB_2 host=ssky1l1p1
status=up reason=USER card=1 timestamp=2014-02-04 00:44:25 timezone=-05:00
db_domain= reported=date

SERVICE TAPS UP event • Application

SERVICE VERSION=1.0 service=TAPS database=scdb instance=SCDB_2 host=ssky1l1p1 status=up
reason=USER timestamp=2014-02-04 00:44:25 timezone=-05:00 db_domain= reported=date