I am looking to learn more about Qlik's ability to handle simultaneous scheduling requests. Does anyone have insight into how a 64-bit enabled architecture with numerous cores can process and handle simultaneous publisher task requests?
This is automatically handled by the QDS. If you are trying to start them at the same time, QDS will add a buffer zone with 0,1 seconds in between. The QVB's that are running the reload will not start more often than 100 ms then. According to a setting in the QMC for the QDS, it allows a specific number of QVB's to run reloads, which is by default 4.
When the QDS detects that a task is to run (either manually or triggered), the QDS adds the task at the end of a queue that is based on the “first come, first served” principle. The QDS then goes through the queue and starts up the needed number of QVBs. If there are more tasks in the queue than QVB engines allowed by this setting (default: 4), the QDS waits until a currently running QVB finishes and then starts up a new QVB to handle the outstanding tasks in the queue.
I think the answer to this is the time between firing the EDX tasks. If you fire one and then another too soon it will queue the task with the one fired before it. I look at this a while back and I can't remember the time between calls, but you should be able to figure it out. Just watch the Task Manager and look for your QVB processes. Once the timing is correct you should see more than one QVB.