Projects
The bottom section of the screen shows details of the projects included in this application. Many applications will only have one project entry listed, but for more complex applications there can be multiple projects included. This screen provides the ability to edit the existing project details, and to add and remove entries, but in most cases you will not need to do this, The name attribute should contain the two part identifier of the project to include. This consists of the product name (e.g. 'mvc' or 'pxp') followed by the project name as shown in the previous example. The location attribute should contain the directory in which this project has been deployed.
XPlatform Modes
Each Morphyc application can run in one of three modes controlled by setting the mode attribute on the xplatform tag. Alternatively, this can be set using the XPlatform Run Mode drop-down on the screen.
static - This is the recommended mode for live applications. In this mode, the platform will initialise all agents and engines when the platform first starts up. Any requests received will reuse these engines. Depending on the pooling settings more engines could still be created under heavy load. The platform will also try to cache assets such as XSLT transforms in this mode to improve performance.
on-demand - In this mode no agents will be initialised at start up, but instead this process will occur for each agent on the first request that requires it. Once an agent has been initialised it performs in the same way as the static mode. This could therefore be used in the situation were quick start up time is important.
dynamic - This is the recommended mode for developing applications. In this mode, the platform will not cache or reuse any agent or asset information. Each agent and engine will be recreated for each request, and so any changes to the application specification will be automatically picked up.
Pooling Options
The pooling section of the Morphyc file controls how many engines can be maintained by each agent (node) in the application. These settings are useful for concurrent multi user applications. There will always be min engines available, but this may be increased to max if the load requires it. Alternatively, engine pooling can be turned off by setting the pooled attribute to false. In this situation, engines will not be maintained, and a new one will be used for each request.
All of these pooling options can be overridden for each agent (node) in the application as required by adding a pooling override as shown in the previous example. The node ids that need to be entered to indicate which agent to override can be found from the dashboard log views, or alternatively from the Node Viewer screen within the studio.
We would not recommend using a very large value for the max number of agents unless the engine cleardown option (see below) is enabled. Otherwise a large number of engines could be created which take up resources, and will never get removed until the application is restarted.
Platform Cleanup
You can instruct XPlatform to automatically perform cleanup operations. This can provide better resource utilisation by releasing engines that are not busy. The Cleanup Enabled option is used to instruct the platform to enable or disable engine cleanup. If this option is enabled, the Cleanup Run Interval should be entered to specify how often the process will run. This is in milliseconds, so a value of 300000represent checking every 5 minutes. As engines are pooled upto the stated maximum during peak loads, there may be the need to clean down pooled engines to release resources once they are no longer being used. This can be managed using the Engine Cleardown option. This sets the idle time (in milliseconds) before enginescan be removed. Every time the cleanup process is run, all engines that have been idle for the entered time or longer will be removed (subject to the min pooling constraints set for the agent).