Regarding the behavior of the IP availability API, I think total_ips needs to be fixed.
In my understanding, "total_ips" is the size of the allocation pools (the number of IP addresses in the allocation pools) and "user_ips" is the number of IP address used by neutron ports.
My understanding above looks correct based on my test http://paste.openstack.org/show/775414/.
I created a subnet with an allocation pool which has 100 IP addresses. I got 100 as total_ips.
In the case of this bug report, the size of the allocation pools is zero, so "total_ips" should be zero.
Note that we can create a port with IP address outside of the allocation pools so the number of available IP addresses does not match "total_ips - used_ips" necessarily.
Regarding the behavior of the IP availability API, I think total_ips needs to be fixed.
In my understanding, "total_ips" is the size of the allocation pools (the number of IP addresses in the allocation pools) and "user_ips" is the number of IP address used by neutron ports.
My understanding above looks correct based on my test http:// paste.openstack .org/show/ 775414/.
I created a subnet with an allocation pool which has 100 IP addresses. I got 100 as total_ips.
In the case of this bug report, the size of the allocation pools is zero, so "total_ips" should be zero.
Note that we can create a port with IP address outside of the allocation pools so the number of available IP addresses does not match "total_ips - used_ips" necessarily.