Hi..
we use a two node hyper-v cluster.
On each node running 5 vm that shared the 64GB system memory.
The vms are configured to use dynamic memory. The sum of the startup (min) memory of the 10 vms are currently 50 GB.
So now our problem:
If we try to drain roles on node A for maintenance purpose two vms are still on node A because they wasted currently each 25GB memory and so on node B is not enough free memory to host this vms.That soundslogical
But as all vms configured to use dynamic memory they should shrink their memory to get enough space for all vms? If we shutdown all vms, move it to one node, then everything works fine.
Is it by design that live migration do not trigger the vm to shrink their memory?
Best regards, Alexander Zirbes