How can I trigger an instant rerun of Check_MK data collection through Icinga?

I am using Check_MK to generate checks to make Icinga (a Nagios fork) monitor a bunch of machines. After starting up Icinga, after some time, Check_MK starts to collect data from the configured machines. The order and the amount of time in between the machines which are checked by Check_MK feels very random and sometimes it takes several minutes until the checks start and even more time until all the machines have been checked once. Also it seems that commanding Icinga to recheck all hosts doesn't start the Check_MK checks.

How can I force Check_MK to process the checks for a specific or all of the configured machines in a way so that Icinga becomes aware of the results instantly after this happens?

The machines run Debian 7 "Wheezy" with Icinga version 1.7 and Check_MK 1.2.2.