Monday, 2024-02-05

*** jph0 is now known as jph08:25
opendevreviewTom Weininger proposed openstack/octavia-tempest-plugin master: DNM test differen dsvm-noo-api config  https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/90730912:43
opendevreviewTom Weininger proposed openstack/octavia-tempest-plugin master: DNM test differen dsvm-noo-api config  https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/90730913:59
opendevreviewTom Weininger proposed openstack/octavia-tempest-plugin master: Add TERMINATED_HTTPS listener API tests  https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/89306615:33
opendevreviewTom Weininger proposed openstack/octavia-tempest-plugin master: Add TERMINATED_HTTPS listener API tests  https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/89306615:35
opendevreviewTom Weininger proposed openstack/octavia-tempest-plugin master: Add TERMINATED_HTTPS listener API tests  https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/89306615:43
QGHello everyone !16:43
QGIs it a legitimate case when the user creates an lb this way?16:43
QGopenstack loadbalancer create --name lb --vip-port-id bcfe386c-77bd-47b6-b127-48ae5c608fe6 --additional-vip subnet-id=7ada4a32-8479-48b5-b68e-429ec3cb54a616:44
QGi think it's a legitimate case and it ends in Traceback, i will create a bug in launchpad16:47
gthiemongeQG: I don't know if it's legitimate, but it's certainly a corner case.17:07
gthiemongeQG: IMO we should handle all the subnets (as VIP subnet + additional subnets) that are plugged into the provided port, but I'm sure we don't do that17:08
gthiemongeQG: then I believe that providing a port_id + additional vips should raise an ValidationException17:09
tweiningI agree, it should never show a traceback. either it should work, or fail validation17:09
gthiemongeQG: in your case, is the subnet 7ada4a already plugged into the port?17:10
QGgthiemonge: why do you think it should raise an ValidationException ?17:10
gthiemongeQG: because I think a multivip LB should only be create with a vip_subnet_id + additional_vips[].subnet_id or a vip_port_id, and not a mix of port and subnets17:11
gthiemongeQG: anyway, please create a launchpad17:11
QGgthiemonge: Subnet 7ada4a32 belongs to the same network as the port's subnet does 17:12
QGgthiemonge: what do you mean by already plugged into the port ?17:12
QGgthiemonge: yep i have : https://bugs.launchpad.net/octavia/+bug/205245517:13
opendevreviewTom Weininger proposed openstack/octavia-tempest-plugin master: Test new HSTS feature  https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/88154617:14
gthiemongeQG: I mean that your subnet_2 should be alredy a one of fixed_ips of the port beofre creating the LB17:14
gthiemongeQG: I need to doublecheck but I think we don't handle multiple subnets in a vip_port_id, so there may be a bug17:14
gthiemongeQG: we can discuss that in the next weekly meeting17:15
QGgthiemonge: ah ok, so in this case the port subnet and the additional subnet are two different subnets but in the same network17:16
gthiemongeQG: does it work when you pass "--vip-subnet-id subnet_1"?17:39
QGgthiemonge: testing17:39
QGgthiemonge: yes it does 17:42
QGin https://github.com/openstack/octavia/blob/master/octavia/controller/worker/v2/tasks/database_tasks.py#L496  additional_vips is empty when passing the vip_port_id but not when passing the vip-subnet-id 17:43
gthiemongeI think we should validate that vip_subnet_id is passed when using vip_port_id and additional_vips[].subnet_id17:44
QGahhhh but wait17:45
QGi think we are not entering in https://github.com/openstack/octavia/blob/master/octavia/api/v2/controllers/load_balancer.py#L22918:01
QGno actually we are 18:09

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!