Parameters Required for Starting an ESP Node

When registering an SCC agent for an ESP node, there are three optional parameters that you can specify: node configuration file path, log configuration file path, and startup folder path.

The node configuration file path is the path to the node configuration file that the agent should use to start the ESP node. This is the equivalent to the –cluster-node parameter for the esp_server command.

The log configuration file path is the path to the log configuration file. This is the equivalent to the –cluster-log-properties parameter for the esp_server command.

The startup folder path is the path to the folder from where the ESP node was started. The agent looks for the startup folder because that is where the cluster log resides. Note that this is not necessarily the location of the esp_server binary. For example, if the esp_server binary is located under $ESP_HOME/bin, but your current directory is /myhome/myserverrun, you can:

> cd /myhome/myserverrun 
> $ESP_HOME/bin/esp_server ..

In this example, /myhome/myserverrun would be the startup folder.

If you do not provide these parameters, the SCC agent attempts to discover the correct node configuration file and derives the other two parameters from this. The agent iterates over all subfolders under $ESP_HOME/cluster/nodes and looks for ".xml". It parses these files looking for those matching the node's port (as specified in node registration in SCC) and name (as specified during installation of Event Stream Processor). If it finds exactly one matching file, that is assumed to be the desired node configuration file.

If multiple matching configuration files are found this way, discovery fails and you will need to specify the node configuration manually. The same happens if the agent does not find any matching configuration files.

Once a node configuration file is found, or if the node configuration was specified in the agent registration, the agent attempts to locate the other parameters, if they are left blank. The startup location is assumed to be the folder where the node configuration file is located, and the log configuration is assumed to reside in the same folder as the node configuration file. Any *.log.properties files located in the same folder are used for log configuration, with preference given to the cluster.log.properties file. If multiple .log.properties files exist in the folder, or if none are found, the server starts without the --cluster-log-config option.