PRIMECLUSTER Wizard for Oracle is a software product that facilitates cluster operation management for Oracle database operating on a PRIMECLUSTER system.
The scalable operation requires Oracle Real Application Clusters (Oracle RAC). The standby operation with Oracle database is enabled requires Oracle cold standby instances. The single-node cluster operation requires an operating node in which a cluster consists of one node.
This software provides the following features:
Feature | Description |
---|---|
Environment setup | Environment setup tool to enable clustering of Oracle database |
Monitoring | Daemon processes to monitor Oracle instance, Listener and detect their failures |
Start and Stop Control | Scripts used for startup and stop of Oracle instances, Listeners |
Setup and Operation Commands | Commands used for setup and operations |
Environment setup
To enable clustering of Oracle database, it is required that Oracle instances and Listeners are registered on userApplication (cluster application) as a "resource". PRIMECLUSTER Wizard for Oracle enables it. For details, refer to "Chapter 2 Environment setup".
Monitoring
PRIMECLUSTER Wizard for Oracle provides daemon processes to monitor Oracle instances and Listeners and detect their failures. If errors are detected in monitoring, in standby operation, Oracle instances and Listeners are restarted and a failover occurs automatically. In scalable operation, Oracle Clusterware automatically restarts Oracle instances and listeners. For details, refer to "3.1 Oracle database Control through PRIMECLUSTER Wizard for Oracle".
Monitoring Oracle instances
The daemon processes monitor Oracle instances in the following way:
Monitoring processes
The daemon process connects to the Oracle instance as the SYSTEM user and cyclically monitors background processes (PMON, SMON, DBWn, LGWR and CKPT).
Monitoring a table
The daemon process creates a table for monitoring and cyclically manipulates data on the table with SQL statements (INSERT, UPDATE, DELETE and COMMIT). It checks the SQL execution result to monitor Oracle instances. Therefore, it means that not only the background process failures but also logical failures can be detected.
Monitoring Listener
The daemon process monitors a Listener process in the following way:
Monitoring processes
The daemon process cyclically monitors Listener processes.
Tnsping command
In standby operation, The daemon process cyclically executes the tnsping command to check that Oracle Net Services are running. In scalable operation, "tnsping" command is not executed.
Monitoring Oracle Clusterware
Oracle Clusterware is not monitored by PRIMECLUSTER Wizard for Oracle.
Startup and stop control
PRIMECLUSTER Wizard for Oracle provides scripts to start up and stop Oracle instances, Listeners and Oracle Clusterware automatically, according to the state transition of userApplication. For details, refer to "3.1 Oracle database Control through PRIMECLUSTER Wizard for Oracle".
Starting and Stopping Oracle instances
In standby operation, the script starts up the Oracle instance, checking the state and recovering. For example, if the Oracle instance had failed during online backup, the scripts automatically execute "END BACKUP" during the startup. When stopping, the scripts stop the instance in the IMMEDIATE mode. If it fails, the script immediately tries in the ABORT. It means that the failover of userApplication can be performed at high-speed.
In scalable operation with Oracle RAC, Oracel RAC instances are started and stopped using srvctl command.
Starting and Stopping Listeners
In standby operation, the script starts up and stops the Listener with "lsnrctl start/stop" command, checking the state.
In scalable operation, Listeners are started by the script.
Starting and Stopping Oracle Clusterware
In scalable operation with Oracle RAC, Oracel Clusterware are started and stopped using crsctl command.