7 Cluster Considerations

Modern business application requirements are classified by the abilities that the system must provide. This list of abilities such as availability, scalability, reliability, audit ability, recoverability, portability, manageability, and maintainability determine the success or failure of a business.

With a clustered system many of these business requirement abilities gets addressed without having to do lots of development work within the business application. Clustering directly addresses availability, scalability, recoverability requirements which are very attractive to a business. In reality though it is a tradeoff, clustered system increases complexity, is normally more difficult to manage and secure, so one should evaluate the pros and cons before deciding to use clustering.

Oracle provides many clustering solutions and options; those relevant to BDI are Oracle database cluster (RAC) and WebLogic Server clusters.

Scaling BDI

BDI needs to be scaled horizontally to handle large number of concurrent jobs. Single instances of Scheduler and Process Flow can be used since they are not resource intensive. Job Admin can be very resource intensive. To handle large number of concurrent jobs, multiple instances of Job Admin can be used to distribute jobs. WebLogic Server cluster that consists of multiple managed server instances provide horizontal scalability for Job Admin.

BDI on Cluster

As recommended above, for scaling BDI for large number of jobs, BDI components should be deployed to cluster. Following are some considerations to be taken into account when deploying BDI on cluster.

Logging

Issue

The "System Logs" tab in Scheduler, Process Flow, and Job Admin UIs show only logs from the server that UI is connected to.

Solution

Use a common log directory for each of the BDI components.

BDI components use the following directory structure for creating log files.

$DOMAIN_HOME/logs/<server name>/<app name>

Example

$DOMAIN_HOME/logs/server1/bdi-rms-job-admin.war

$DOMAIN_HOME/logs/server2/bdi-rms-job-admin.war

Create a common log directory (e.g. /home/logs/jobadmin) for each BDI application.

Create symbolic links to the common log directory for each server using the below command from $DOMAIN_HOME/logs directory.

ln -s  /home/logs/jobadmin
       server1/bdi-rms-job-admin.war
ln -s /home/logs/jobadmin
       server2/bdi-rms-job-admin.war

If the directory $DOMAIN_HOME/logs/<server>/<app> already exists, it needs to be deleted before symbolic link is created.
App needs to be restarted after symbolic link is created.

When WebLogic managed servers are in different machines a shared network disk has to be used.

Update Log Level

Issue

When log level is updated through UI or REST end point, it updates the log level only on the server it is connected to.

Solution

Log level needs to be updated through the URL of all the nodes in the cluster using UI or REST endpoint.

Example

http://server1:port1/bdi-rms-batch-job-admin/resources/system-setting/system-logs

http://server2:port2/bdi-rms-batch-job-admin/ resources/system-setting/system-logs

Create/Update/Delete System Options

Issue

When system options are created/updated/deleted using UI or REST end point, the changes are reflected only on the server that client is connected to.

Solution

The reset-cache REST endpoint needs to be invoked on the other nodes in the cluster for that application in BDI.

Example

http://server1:port1/bdi-rms-batch-job-admin/ resources/system-setting/reset-cache

Use curl command to reset cache as:

curl --user userId:password -i -X POST -H "Cot-Type:application.json"   
http://server1:port1/bdi-rms-batch-job-admin/resources/system-setting/reset-cache

Create/Update/Delete System Credentials

Issue

When system credentials are created/updated/deleted using REST endpoint, the credentials are created/updated/deleted only on the server that client is connected to.

Solution

The REST endpoint that creates/updates/deletes credentials need to be invoked on all the nodes in the cluster for that application in BDI.

Example

http://server1:port1/bdi-rms-batch-job-admin/resources /system-setting/system-credentials
http://server2:port2/bdi-rms-batch-job-admin/resources /system-setting/system-credentials
 
        Use curl command to create credentials on other nodes in the cluster as:
 
        curl --user userId:password -i -X PUT -H "Content-Type:application/json" 
        http://server1:port1/bdi-rms-batch-job-admin/resources/system-setting/system-credentials  
        -d '{"userAlias":"rmsappJobAdminBaseUrlUserAlias", "userName":"rmsjobadmin" ,    "userPassword":"xyzxyz"}'

Use curl command to update credentials on other nodes in the cluster as:

        curl --user userId:password -i -X POST -H "Content-Type:application/json"       http://server1:port1/bdi-rms-batch-job-admin/resources/system-setting/system-credentials
         -d '{"userAlias":" reimappJobAdminBaseUrlUserAlias", "userName":"reimjobadmin" ,        "userPassword":"wwwqqqq"}'

Use curl command to delete credentials on other nodes in the cluster as:

        curl --user userId:password -i -X DELETE  -H "Content-Type:application/json" 
        http://server1:port1/bdi-rms-batch-job-admin/resources/system-setting/system-credentials
         -d '{"key":"rmsappJobAdminBaseUrl"}'

Scheduler Configuration Changes for Cluster

Perform the following procedure to cluster the Job Scheduler Data Source:

Two data sources need to be created for scheduler on cluster in the Admin Console.
- Create a non-XA data source (SchedulerTimerDs) pointing to the schema that contains the WEBLOGIC_TIMERS table. This is the schema with the WLS suffix, created using RCU.
  
  Specify this schema in the scheduling tab of cluster configuration in WebLogic console.
- Create a non-XA data source (SchedulerRuntimeDs) pointing to schema that contains ACTIVE table. This is the schema with the WLS_RUNTIME suffix, created using RCU.
  
  Specify this schema in the Migration tab of cluster configuration in the WebLogic console.
Perform the following steps to configure the data sources:
1. Specify the data source for schedule timers in the Admin Console.
2. Login to Admin Console.
3. Click Lock & Edit (For Production Mode only).
4. Click Environment -> Clusters.
5. Click the cluster name.
6. Click Scheduling.
7. Select SchedulerTimerDs for the Data Source For Job Scheduler field.
8. Click Save.
9. Click Migration.
10. Select Migration Basis: DataBase, and Data Source For Automatic Migration: SchedulerRuntimeDs.
11. Click Save.
12. Verify Auto Migration Table Name populated with ACTIVE.
13. Click Activate Changes.

Update the weblogic-ejb-jar.xml in WEB-INF folder of the bdi-scheduler-ui-<version>.war in <bdi-home>/dist folder with the contents shown (The entry in red is the change from the existing contents of the file)

Instructions to update

cd dist
jar xf bdi-scheduler-ui-<version>.war WEB-INF/weblogic-ejb-jar.xml
Update the WEB-INF/weblogic-ejb-jar.xml with the contents below
jar uf bdi-scheduler-ui-<version>.war WEB-INF/weblogic-ejb-jar.xml
Delete dist/WEB-INF folder
Deploy the scheduler application

<weblogic-ejb-jar xmlns="http://xmlns.oracle.com/weblogic/weblogic-ejb-jar" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
    <security-role-assignment>
        <role-name>AdminRole</role-name>
        <principal-name>BdiSchedulerAdminGroup</principal-name>
    </security-role-assignment>
 
 
    <security-role-assignment>
        <role-name>OperatorRole</role-name>
        <principal-name>BdiSchedulerOperatorGroup</principal-name>
    </security-role-assignment>
    <security-role-assignment>
        <role-name>MonitorRole</role-name>
        <principal-name>BdiSchedulerMonitorGroup</principal-name>
    </security-role-assignment>
    <timer-implementation>Clustered</timer-implementation>
</weblogic-ejb-jar>