Asks the discovery component to discover all of its resources. The plugin container may or may not have already {@link ResourceDiscoveryContext#getAutoDiscoveredProcesses() auto-discovered} some resources for this componentalready. In this case, this discovery component should take those auto-discovered resources from the context and "prepare them" to be included as part of the returned set of resources, unless there is some reason it doesn't want to include those resources. By "prepare them", it means to make sure to set their {@link DiscoveredResourceDetails#setResourceKey(String) key}, {@link DiscoveredResourceDetails#setResourceName(String) name} and other details.
ClassLoader Note:There are two main usages of this method: first, when top-level resources need to be discovered (i.e. servers running as direct children underneath the top-level platform) and second, when discovering child services and child servers running inside those top-level server resources that were previously discovered.
For the first case, this discoverResources method implementation can't know where any specific client jars are that may be needed to connect to the managed resource, because the managed resource hasn't been discovered yet! Therefore, when discovering top-level servers, this method cannot import or use connection classes that are available from a managed resource's client jars, because it isn't yet known where they are! (see {@link ClassLoaderFacet} forinformation on client jar detection). In this case, this method is called within the context of the main plugin classloader.
In the second case, this method is called within the context of its parent resource thus giving this method access to connection classes that are found within that parent resource's client jars. The plugin container will ensure that this method is invoked with a context classloader that has the appropriate client jars available.
What the above means is that this class must not import or directly access client jar classes, because within some contexts, it will not have access to those jars. The easy way to think of this is the following: this discovery component must not directly import, load or access any class that is found in any jar that is returned by this object's implementation of {@link ClassLoaderFacet#getAdditionalClasspathUrls(ResourceDiscoveryContext,DiscoveredResourceDetails)}.
@param context the discovery context that provides the information to the component that helps it perform itsdiscovery @return a set of discovered resource details that were discovered and can be imported/merged into inventory @throws InvalidPluginConfigurationException (used only be the deprecated manual add API) if the pluginconfiguration found in the context was somehow invalid and thus caused a failure to connect to a managed resource @throws Exception if a generic error occurred that caused the discovery to abort | |
| |