How to Implement Clustering

Clustering › How to Implement Clustering

How to Implement Clustering

As a CA Service Catalog administrator, you can optionally use clustering for CA Service Catalog to improve performance and provide failover protection. Here, the term clustering means multiple computers in a group that perform the same or similar function, essentially acting as one virtual computer. Specifically, clustering here refers to CA Service Catalog running on two or more computers. Failover protection means that if one computer malfunctions, becomes heavily loaded, or loses power, its workload is transferred to the other computers in the cluster. These computers retain and complete the active sessions.

Another advantage of clustering is load-balancing: If one of the cluster components is busy processing a request, the load is redirected to another component in the cluster. Users of the system see no interruption of access. CA Service Catalog processing continues. The loss of performance on users and business functions is minimized, even when computer availability is lost or reduced.

Important! These instructions apply only to the Apache Tomcat web server version supplied with CA Service Catalog. If you are using any other web server instead, skip these instructions and see your web server documentation for instructions to implement clustering.

To implement clustering for CA Workflow and CA Service Catalog, perform these tasks.

Perform the preliminary tasks.
Meet the prerequisites.
Set up horizontal clustering for CA Service Catalog
Set up NTLM authentication for each cluster, if applicable.
Configure NTLM with Apache Load Balancer, if applicable.
Optionally set up horizontal clustering for CA Workflow.
Set up load balancing.
If you are implementing horizontal clustering on multiple computers, use the computers that have the same default gateway. To obtain the default gateway, enter the ipconfig command at the DOS prompt.
To record cluster-related logging information in tomcat_view.log, do the following:
1. Add the following section to the log4j.xml file for each CA Service Catalog computer:
```
<logger name="org.apache.catalina.cluster" additivity="false">
  <level value="TRACE" />
  <appender-ref ref="tomcat_view" />
</logger>
```
2. Set the log level for each log4j.xml file to meet your needs.
  For example, use a high log level while you are troubleshooting. Conversely, use a low log level, or comment out the previous lines, at other times.
  
  Note: When you use the previous lines, each communication between CA Service Catalog computers is logged in the log4j.xml file.
Also, consider setting up a single location for shared files. While this task is not directly related to clustering, it does help improve overall system performance and helps reinforce best practices.