PRIMECLUSTER Wizard for Oracle is a software product that facilitates cluster operation management for Oracle operating on a PRIMECLUSTER system.
The scalable operation requires Oracle Real Application Clusters (Oracle9i RAC, Oracle RAC 10g or later). The standby operation requires Oracle cold standby instances.
Module structure
The modules for operating Oracle on PRIMECLUSTER are as follows:
Environment setup Wizard | Environment setup tool to enable Oracle operation on PRIMECLUSTER |
Detector | Module to monitor Oracle |
Scripts | Control Oracle startup and stop |
Setup/operation command | Commands used to set up and operate |
Environment setup
The environment setup tool provides “ORACLE” Wizard that generates userApplication in the environment configuration by executing the “hvw” command of PRIMECLUSTER RMS Wizard.
Monitoring
A detector monitors Oracle instances and Oracle listeners.
The detector connects itself to Oracle as a SYSTEM user to monitor the Oracle instance. Then, it periodically creates, updates, and deletes a table, which is created in the SYSTEM user’s default table space by executing the SQL command so that it can monitor the process state of the Oracle instance as well as logical failures. If the Oracle instance hangs, so the SQL command is not returned within a specified time, then the detector will time out to notify a user of the error.
The two types of monitoring the Oracle listener are process detection and the “tnsping” command.
In operation with Oracle RAC 10g or later, a detector monitors Oracle instances and Oracle listeners. If the Oracle instances and listeners fail, they will be recovered by Oracle Clusterware.
Startup and stop control
Along with the state transition of a cluster system, scripts automatically start or stop Oracle instances or Oracle listeners. These scripts first check and then start up the Oracle instances. If damage occurs, the scripts automatically recover the Oracle instances. For example, if the Oracle instance fails during online backup, the scripts automatically execute “end backup” to start up the Oracle instance (When the AutoRecover is set to “Yes”.) When stopping the Oracle instance, the scripts stop it in the immediate mode (default). If this does not occur properly, the scripts will stop the instance in the abort mode. This means operating node switchover can be performed at high-speed.
In operation with Oracle RAC 10g or later, start and stop of Oracle Clusterware are managed.
