Working with Enterprise Manager

Oracle Enterprise Manager (OEM) is useful for troubleshooting and diagnosing issues communicating with the head end system. In particular, security and other communications issues can be discerned.

To locate the instance of a service that is showing a problem, open the Dashboard view of the service. Each run instance is time-indexed. The State column in the first table contains the most important information for each instance. Completed messages show up with a green check icon. Terminated instances are indicated with a grey stop icon. Instances that are still running are marked as such. Runtime exceptions are in the bottom window.

When you click on the instance ID, a new window opens showing the entire process flow. This view is particularly useful in debugging.

Note the State column showing the status of each instance within the flow. SOA processes are made up of several calls to SOA components and web services. The Instance column shows a rough ordering of the operations. A typical approach to troubleshooting would find the “lowest” instance of an error or termination. This would be the error that is deepest within the process and is usually the source of the problem. In the above case, the deepest error is in the very first composite for the process. More information can be found by clicking into the link within the Instance column.

The level of detail is dependent on the auditing level set on the server. In this case, the error is in the response from the head end system.

OEM also contains a centralized location to control security. On a Dashboard screen for a composite, the Policies tab shows the OWSM policies attached to the composite. Typically, a log policy is placed on any inbound or outbound communication. Also, all services or references are delivered with attached basic http security policies.