With the introduction of the new Headless Client type build, we have a lot more flexibility to run server logic than ever before. I'd like to examine launching headless clients as needed, for example to run server logic on instances of a game.
Right now we need to launch these manually. Ideally we'd be able to launch them from another job and pass in some parameters which tell the new process which area/session it needs to load up, or otherwise how to configure itself.
We don't have the ability that I can see to launch new programs/processes from within Shiva without a plugin, and plugins aren't available yet on Linux headless clients.
I was thinking at this point I'd need a separately written non-Shiva process which periodically polls a database to look for a record which indicates that a new headless client is needed. When it is found, it could then launch the new headless client process in the background. When the new headless client is launched, it can examine the database and grab the most recent configuration data, and then pop that data off the list.
Any other methods that you guys can think of?