I have encountered a few issues regarding the resource tracking and hopefully 
someone can help clarify here. From what I understand we just seem to ignore 
the actual used disk and memory from the hypervisor and calculate it according 
to the allocated instances. For example, if the hypervisor has cached images or 
volume support, then this disk usage is ignored.
I have posted https://review.openstack.org/#/c/52900/ but I am not 100% sure 
that this is the correct way of addressing this (here we need to ensure that 
the hypervisor is not returning the disk spaces used by instances, similarly 
with memory)
OpenStack-dev mailing list

Reply via email to