Configuring system data persistence using an Oracle database

Content

To configure Decision Server Insights persistence using an Oracle database, use the following information instead of the DB2 topics that are currently in the IBM Knowledge Center:

Setting up a database
Instead of Setting up a database, set up a database for your container servers to persist entity data and system state data. This task is not applicable to the other server types.

Before you begin
A relational database can be configured to store the runtime state while Insight Server is running. After a system outage, the stored runtime state can be recovered from the database. The runtime state includes entity business information, in-progress transactions, and the history of events and event data.

Decide on whether to persist the runtime data in a database by balancing the following factors:
• The value of the business information in the runtime state of the system is sufficiently valuable.
• The risk of a full system outage is sufficiently high.
• The performance impact of this feature to the business case is sufficiently small.

About this task
If persistence is required, you must set up a database, and you must define a data source that connects to that database. The database can exist on a separate computer, or on a computer that hosts a server.

Decision Server Insights includes an Oracle database script OracleDistrib.sql, which is in <InstallDir>/runtime/ia/persistence/sql/Oracle.

Tip: Because site-specific requirements vary greatly, the OracleDistrib.sql file specifies only the basic structure of the required tables. Attributes such as table spaces and LOB locations have a default value, or are not specified. A database administrator might want to modify a copy of this file to define these performance-related attributes.

To set up a database, you run the SQL statements from a command window or a database client.

Results
After the SQL script is run, an administrator can check the database to validate the following tables are created:

Database tables	Description
DELAYEDEVENTS	Data about the events that are due to be processed at a specific time.
ENTITIES	Data about the entities in the run time.
EVENTSTORE	Data about the event history that is related to entities.
INBOUNDEVENTS	Data about the events that are waiting to go into an agent.
JOBHISTORY	Data about the stored global aggregates.
JOBRESULTS	Data about the stored global aggregates.
OUTBOUNDEVENTS	Data about the events that are waiting to be sent, or in progress of being sent, to an outbound endpoint.
RULESETS	Data about the rule engine state.

What to do next
When a server that is configured for database persistence is cloned and ready to start, run the server start command. After the server is started, run the dataLoadManager finalize command to activate the persistence.

Configuring Decision Server Insights persistence in JDBC
Instead of Configuring Decision Server Insights persistence in JDBC, add a Java™ database connectivity (JDBC) data source to determine the database to use for persisting system data. JDBC is an essential prerequisite for communicating with the database. The data source is looked up by a Java Naming Directory Interface (JNDI) process.

Procedure:
1) Uncomment the feature manager for JDBC in the <InstallDir>/runtime/wlp/usr/servers/your_cisContainer_name/server.xml file.

<featureManager>
       <feature>jdbc-4.0</feature>
    </featureManager>

2) Create a JDBC data source, and configure the JDBC driver for your database.

You must ensure that the data source references the database where your map data is persisted. For more information, see Adding a data source by using the developer tools.

For more information, see Editing the Liberty profile configuration by using developer tools.

3) Complete the following entries for the JDBC data source:

• JNDI name
• Path to the JDBC .jar directory
• Space-separated list of the JDBC .jar files to include
• Database name
• User name
• Password
• Server name
• Database schema name
• Port number

The following code is an example of an Oracle configuration:

<dataSource jndiName="jdbc/oracle_datasource">
   <connectionManager/>
   <jdbcDriver>
      <library>
         <fileset dir="/opt/oracle/java" includes="ojdbc6.jar" />
   </library>
   </jdbcDriver>
   <properties.oracle
  databaseName="SAMPLEDB"
  user="user1"
  password="pwd1"
  serverName="localhost"
  portNumber="1521" />
</dataSource>

Add a persistence configuration item, and set the data source JNDI name. You can set the maximum batch size and maximum cache age attributes. The following code is an example of an Oracle configuration:

<ia_persistence databaseType="ORACLE" datasourceName="jdbc/oracle_datasource" maxBatchSize="10000" maxCacheAge="1000" />

Where:
datasourceName (mandatory)
The JNDI reference of the data source that you created in step 1.

maxBatchSize (optional)
The maximum number of entries that can be restored to the system in a single transaction if recovery occurs. The rows are read in batches of the size that is configured. The default is 10,000.

The maxBatchSize value can affect the performance of the recovery process. However, there is no risk of data loss because it is a read only operation.

maxCacheAge (optional)
The maximum time in milliseconds that the system caches data source information before it looks up the data source again. The default is 1,000.

The system checks the server.xml for new database details at the specified rate.

What to do next
Update the object grid to enable database persistence. See Configuring the object grid to persist data.