RAC

When the jobs you are running depend on things other than just the database, it might be interesting to know what these other things are, and where they can be found. There are several ways to control which instance a job is going to run, some of which are:

• Define instance_id in the job definition

• Use instance_stickiness in the job definition

• Define service_name in the job class definition for the job

To be able to show what exactly happens, we need an RAC installation. The RAC installation used here runs on two nodes—oralin01 and oralin02—using the shared storage that is exported as iSCSI disks from the of01 server running openfiler. The database called "shields" is using the iSCSI disks through the ASM. The instances of this database are shields1 and shields2. In the default setup, every instance of the cluster supports the service whose name is the same as the database name. For demonstration purposes, the services departures and arrivals have been defined. The preferred instance...