Corosync high availability order constraints

At the moment I'm working on a high availability environment with corosync, with multiple resource groups. I'm struggling with the order in which these groups should start (this is my first time working with pacemaker and corosync).

My desired configuration resembles example 6.17 from the Pacemaker/Corosync docs, yet in reverse order and the resources are groups:

• Example 6.17, original docs
• Desired configuration, imgur (note, my config prefixed group before the capital letters)

The colocation constraints work, yet the order in which groups without inter-dependencies can start/stop don't.

I have 2 goals:

• I would like to stop groupC without stopping groupD&E
• Understand how I could write this in crm shell syntax instead of xml

I'm kind of clueless here, and was hoping someone could point me out what I'm doing wrong.

• bash commands, example of what isn't working, sprunge
• $ crm configure show, sprunge
• $ cibadmin -Q, sprunge

Thanks in advance!

ps. this is a mockup configuration. In the final configuration groupA will be a drbd device, groupB will mount this device and start several services like nfs & samba. GroupC-Z will contain different apache servers and mountpoints. Group A-Z should all run on the same node.