On 5 December 2016 at 07:59, Bence Romsics <bence.romsics at gmail.com> wrote: > Hi, > > I measured how the new trunk API scales with lots of subports. You can > find the results here: > > https://wiki.openstack.org/wiki/Neutron_Trunk_API_Performance_and_Scaling > > Hope you find it useful. There are several open ends, let me know if > you're interested in following up some of them. > I looked into [1] a little bit as it bothered me :) Here's my findings: - openstack port list is slower than neutron port-list as it looks like the command goes to Nova first [2], which is where the overhead comes from. - when you have subports the port-list response gets bigger because of the bigger response due to the trunk-details extension. - However, the bulk of the time is spent in [3], when building the payload for a port involved as a parent port. In no other case you will see the overhead as in no other case the loop over subports is executed. The hook should be optimized! Cheers, Armando [1] https://wiki.openstack.org/wiki/Neutron_Trunk_API_ Performance_and_Scaling#surprise_effect_on_filtered_port_listings [2] http://paste.openstack.org/show/591882/ [3] https://github.com/openstack/neutron/blob/master/neutron/services/trunk/plugin.py#L42 > Cheers, > Bence Romsics > > __________________________________________________________________________ > OpenStack Development Mailing List (not for usage questions) > Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20161208/0c3552c7/attachment.html>