Here is where nova API checks for request limit < 0:
https://github.com/openstack/nova/blob/2014.1.b3/nova/api/openstack/common.py#L210
and get_limit_and_marker is used by the servers API:
https://github.com/openstack/nova/blob/2014.1.b3/nova/api/openstack/common.py#L222
https://github.com/openstack/nova/blob/2014.1.b3/nova/api/openstack/compute/servers.py#L595
So it would be easy enough to just detect limit == 0 in the servers API (v2 and v3) and handle it there by returning instance_list = [].
Here is where nova API checks for request limit < 0:
https:/ /github. com/openstack/ nova/blob/ 2014.1. b3/nova/ api/openstack/ common. py#L210
and get_limit_ and_marker is used by the servers API:
https:/ /github. com/openstack/ nova/blob/ 2014.1. b3/nova/ api/openstack/ common. py#L222
https:/ /github. com/openstack/ nova/blob/ 2014.1. b3/nova/ api/openstack/ compute/ servers. py#L595
So it would be easy enough to just detect limit == 0 in the servers API (v2 and v3) and handle it there by returning instance_list = [].