| opendevreview | Merged openstack/neutron master: agent/linux: remove eventlet usage from of_monitor https://review.opendev.org/c/openstack/neutron/+/940684 | 00:43 |
|---|---|---|
| opendevreview | yatin proposed openstack/neutron master: [OVN] Add option to allow configuring dns ovn-owned https://review.opendev.org/c/openstack/neutron/+/942373 | 02:40 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: DNM - Test errors in ``neutron-functional-with-pyroute2-master`` https://review.opendev.org/c/openstack/neutron/+/942859 | 06:49 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: DNM - Test errors in ``neutron-functional-with-pyroute2-master`` https://review.opendev.org/c/openstack/neutron/+/942695 | 06:50 |
| opendevreview | Dmitriy Chubinidze proposed openstack/neutron master: Update CIDR notation for loopback address according to actual RFC (https://datatracker.ietf.org/doc/html/rfc5735) https://review.opendev.org/c/openstack/neutron/+/942763 | 08:03 |
| opendevreview | Dmitriy Chubinidze proposed openstack/neutron master: Update CIDR notation for loopback and multicast address according to actual RFC (https://datatracker.ietf.org/doc/html/rfc5735) https://review.opendev.org/c/openstack/neutron/+/942763 | 08:09 |
| opendevreview | Vasyl Saienko proposed openstack/neutron master: Inherit parent port vif|vnic_type for subports https://review.opendev.org/c/openstack/neutron/+/942837 | 08:46 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [FT] Mark some ``TestNeutronServer`` tests as unstable https://review.opendev.org/c/openstack/neutron/+/942883 | 09:02 |
| ralonsoh | slaweq, lajoskatona ^ please check it to unblock the FT CI. More info provided in the LP bug (olso.service related) | 09:03 |
| slaweq | ralonsoh +2 | 09:05 |
| lajoskatona | ralonsoh: thanks for checking | 09:10 |
| ralonsoh | ah, oslo if you have time: https://review.opendev.org/c/openstack/neutron/+/938541 | 09:11 |
| ralonsoh | very short patch | 09:11 |
| opendevreview | Lajos Katona proposed openstack/neutron master: WIP: [eventlet-removal] Remove the usage of eventlet in the SR-IOV agent https://review.opendev.org/c/openstack/neutron/+/942884 | 09:20 |
| opendevreview | Lajos Katona proposed openstack/neutron master: [eventlet-removal] Remove eventlet from DHCP agent https://review.opendev.org/c/openstack/neutron/+/942393 | 09:20 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [eventlet-removal] Remove the usage of eventlet in the DHCP agent https://review.opendev.org/c/openstack/neutron/+/942530 | 09:21 |
| opendevreview | Vasyl Saienko proposed openstack/neutron master: Inherit parent port vif|vnic_type for subports https://review.opendev.org/c/openstack/neutron/+/942837 | 09:23 |
| rxiao | Hi, I'd appreciate a review on this if anyone has time: https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/940906 | 10:02 |
| ralonsoh | rxiao, let me check | 10:18 |
| ralonsoh | rxiao, is there a job where this test is being executed for ML2/OVN with snat_rules_apply_to_nested_networks=True? | 10:23 |
| rxiao | The test is not in CI yet, and I've only tried it manually. | 10:39 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: DNM - Test errors in ``neutron-functional-with-pyroute2-master`` https://review.opendev.org/c/openstack/neutron/+/942695 | 10:40 |
| ralonsoh | rxiao, the test is being executed in the CI | 10:41 |
| ralonsoh | but with the config value to false | 10:41 |
| ralonsoh | https://fecebc633a48a866ba5a-cd4be6953fb399b0e556e02be7ab8b9a.ssl.cf5.rackcdn.com/940906/10/check/neutron-tempest-plugin-ovn/aa96b8b/testr_results.html | 10:41 |
| ralonsoh | test_snat_external_ip[id-3d73ea1a-27c6-45a9-b0f8-04a283d9d764] | 10:41 |
| ralonsoh | pass | 10:41 |
| rxiao | Is it possible to add the upstream patch as a Depends-On in Gitlab? If yes, would triggering one with snat_rules_apply_to_nested_networks=True on gitlab help? | 10:46 |
| ralonsoh | I'm not talking about any D/S repository, but the U/S one | 10:47 |
| rxiao | Could you tell me how to do that in upstream? | 10:50 |
| ralonsoh | rxiao, I've commented that in the patch. You need to change the job definition, but only for Epoxy. That means you'll need to set it to false in the stable releases | 10:50 |
| ralonsoh | let me update your patch | 10:51 |
| ralonsoh | is that ok for you? | 10:51 |
| rxiao | Sure, thank you | 10:51 |
| opendevreview | Maor Blaustein proposed openstack/neutron-tempest-plugin master: [DNM] Testing 3rd party whitebox gate vote https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/942887 | 10:56 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: DNM - Test errors in ``neutron-functional-with-pyroute2-master`` https://review.opendev.org/c/openstack/neutron/+/942695 | 11:03 |
| opendevreview | Renjing Xiao proposed openstack/neutron-tempest-plugin master: Cover both enabled and disabled cases in nested snat validation test https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/940906 | 11:07 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: DNM - Test errors in ``neutron-functional-with-pyroute2-master`` https://review.opendev.org/c/openstack/neutron/+/942695 | 11:25 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: DNM - Test errors in ``neutron-functional-with-pyroute2-master`` https://review.opendev.org/c/openstack/neutron/+/942695 | 11:47 |
| opendevreview | Renjing Xiao proposed openstack/neutron-tempest-plugin master: Cover both enabled and disabled cases in nested snat validation test https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/940906 | 11:49 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: DNM - Test errors in ``neutron-functional-with-pyroute2-master`` https://review.opendev.org/c/openstack/neutron/+/942695 | 11:51 |
| ralonsoh | rxiao, I never suggested to create a new job. We have too many in n-t-p | 12:02 |
| ralonsoh | what we need is to enable this flag for ovn in master and disable it in stable jobs | 12:03 |
| ralonsoh | same as you did in PS11 but also adding the needed configuration for the stable jobs | 12:03 |
| rxiao | Sorry in which stable jobs should it be disabled? | 12:26 |
| ralonsoh | rxiao, in any job inheriting from neutron-tempest-plugin-ovn, in a stable release | 13:21 |
| ralonsoh | for example (one sec) | 13:21 |
| ralonsoh | https://github.com/openstack/neutron-tempest-plugin/blob/master/zuul.d/2024_2_jobs.yaml#L202 | 13:22 |
| ralonsoh | https://github.com/openstack/neutron-tempest-plugin/blob/master/zuul.d/2024_1_jobs.yaml#L204 | 13:22 |
| ralonsoh | etc | 13:22 |
| rxiao | Thank you | 13:32 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: DNM - Test errors in ``neutron-functional-with-pyroute2-master`` https://review.opendev.org/c/openstack/neutron/+/942859 | 13:32 |
| opendevreview | Renjing Xiao proposed openstack/neutron-tempest-plugin master: Cover both enabled and disabled cases in nested snat validation test https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/940906 | 13:49 |
| opendevreview | Merged openstack/neutron master: [FT] Mark some ``TestNeutronServer`` tests as unstable https://review.opendev.org/c/openstack/neutron/+/942883 | 13:52 |
| opendevreview | Luca Del Monte proposed openstack/neutron master: [FT] Added test_ha_chassis_keep_priority_external_ports function https://review.opendev.org/c/openstack/neutron/+/942901 | 14:40 |
| lucadelmonte | hello ralonsoh, i think i managed to create a working functional tests that proves what i described you yesterday, let me know if it is good enough ^^ | 14:43 |
| ralonsoh | lucadelmonte, I'll check it later today or tomorrow | 15:00 |
| ralonsoh | lucadelmonte, but you are not testing the issue you reported | 15:01 |
| ralonsoh | this test is not creating a port inside a txn and then creating the HA_Chassis_Group inside it | 15:03 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: DNM - Test errors in ``neutron-functional-with-pyroute2-master`` https://review.opendev.org/c/openstack/neutron/+/942695 | 15:07 |
| lucadelmonte | sorry maybe i did explain myself clearly yesterday, the issue i am incurring is what i am testing in the function, the transtaction that is creating the problem is inside _sync_ha_chassis_group, the command which is supposed to get if an ha_chassis_group is already present always fails because the ovsdbapp does not return the result you expect with may_exists=True because the transaction is not commited | 15:09 |
| lucadelmonte | and my tests shows that the master of the chassis changes when 2 external ports are created because _sync_ha_chassis_group thinks this is a brand new chassis group instead of an already existing one | 15:10 |
| ralonsoh | lucadelmonte, a create method with may_exists=True should always return the value | 15:10 |
| ralonsoh | the new one or the existing one | 15:10 |
| lucadelmonte | according to my tests this is not the case | 15:11 |
| ralonsoh | if that is not happening, this could be an issue in ovsdbapp | 15:11 |
| ralonsoh | did you create a LP bug? | 15:11 |
| ralonsoh | just to have the reference | 15:11 |
| lucadelmonte | nope, will do | 15:12 |
| opendevreview | Slawek Kaplonski proposed openstack/neutron master: Add .mypy_cache directory to the gitignore file https://review.opendev.org/c/openstack/neutron/+/938123 | 15:17 |
| ralonsoh | otherwiseguy, qq, about the issue with ``_sync_ha_chassis_group`` in Neutron | 15:18 |
| ralonsoh | we have this: https://github.com/openstack/neutron/blob/ec257eb74a64e3782c052f975942e17ba085eeb3/neutron/common/ovn/utils.py#L1117-L1123 | 15:19 |
| ralonsoh | and we also have this: | 15:19 |
| * otherwiseguy looks | 15:19 | |
| ralonsoh | https://github.com/openstack/ovsdbapp/blob/master/ovsdbapp/schema/ovn_northbound/commands.py#L2039 | 15:20 |
| ralonsoh | if the HA_Chassis_Group exists, we should have it in .result | 15:20 |
| lucadelmonte | https://bugs.launchpad.net/neutron/+bug/2100505 | 15:22 |
| otherwiseguy | It looks like it is looking for the result before the command runs? | 15:22 |
| otherwiseguy | since run_idl() would be run when we commit the txn | 15:23 |
| ralonsoh | otherwiseguy, right so this won't work as is now | 15:24 |
| ralonsoh | it would be needed another transaction to create/retrieve the ha_chassis_group at this moment | 15:24 |
| ralonsoh | actually something similar to what was done before | 15:26 |
| ralonsoh | https://review.opendev.org/c/openstack/neutron/+/894767/16/neutron/common/ovn/utils.py#1042 | 15:26 |
| ralonsoh | ok, I'll add this comment to the bug | 15:26 |
| lucadelmonte | yeah this is how i fixed it locally in prod :) | 15:27 |
| ralonsoh | I'll push a patch tomorrow | 15:27 |
| otherwiseguy | ralonsoh: if we didn't pass a txn around to a bunch of methods, then storing the cmd and grabbing the specific result would be a lot easier. | 15:27 |
| lucadelmonte | thanks | 15:27 |
| ralonsoh | otherwiseguy, yes, that's right | 15:28 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: DNM == WIP == [eventlet-removal] Remove eventlet in the Metadata Agent https://review.opendev.org/c/openstack/neutron/+/942916 | 15:47 |
| cardoe | hello all. I'm hoping that someone can point me to the correct documentation. Using Nova and Neutron. I've got an external network called PUBLICNET which I'm looking to have router and floatips assigned out of. So that subnet has service_type of network:router_gateway and the floating IP one. But then I'm also looking to create another network per tenant that I'm calling OUTSIDE. | 18:28 |
| cardoe | I'm using network rbac create to give just the one project access_as_external for OUTSIDE. But what would be the approach that I can use to allow nova servers to be built directly connected to OUTSIDE but not to PUBLICNET? | 18:31 |
| cardoe | I'm wondering if I need to set some kind of other kind of service_type on subnets in OUTSIDE? | 18:31 |
| opendevreview | Mohammed Naser proposed openstack/ovn-octavia-provider master: Add define_plugin and plugin_requires for DevStack https://review.opendev.org/c/openstack/ovn-octavia-provider/+/942940 | 19:19 |
| haleyb | cardoe: having no service_type implies all can use it, i guess you could restrict with compute:nova if you really wanted, but if only one tenant has access don't know why you'd need to? | 19:25 |
| cardoe | Just wanting to make sure I'm not opening a hole is why I was asking. | 19:27 |
| cardoe | So then follow on question if I can haleyb. Let's say I create a /24 subnet pool and say the min/max it can carve out being a /29. Now I go to add a subnet to OUTSIDE. Is there anyway to restrict OUTSIDE to only taking subnet's from a specific subnet pool? | 19:45 |
| cardoe | Or maybe an address scope? | 19:46 |
| haleyb | hmm, i don't think you can force it to a single subnet pool, i typically only see admins adding subnets to external networks so never thought about it | 20:09 |
| jamesdenton | cardoe do you have some kind of L3 driver configuring the subnet for OUTSIDE on some device (other than a neutron router)? I don't know that I've seen subnet pools used in that way (for use as external provider networks) because that is not really tenant-driven. | 20:25 |
| jizaymes | Cant get the neutron metadata service to show anything but error 404. ERROR neutron.agent.ovn.metadata.server [-] No port found in network None with IP address 103.229.152.80 -- ovn-nbctl show shows it existing and on a network. | 20:39 |
| jizaymes | somehow it stopped functioning entirely despite showing running and restarting cleanly (kolla ansible docker containers) Not sure what I'm doing wrong as to why it shows no network even though it has an IP and port assigned | 20:40 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: DNM - Test errors in ``neutron-functional-with-pyroute2-master`` https://review.opendev.org/c/openstack/neutron/+/942695 | 21:02 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: DNM - Test errors in ``neutron-functional-with-pyroute2-master`` https://review.opendev.org/c/openstack/neutron/+/942859 | 21:04 |
| opendevreview | Vasyl Saienko proposed openstack/neutron master: Inherit vif|vnic_type and host for trunk subports https://review.opendev.org/c/openstack/neutron/+/942837 | 21:04 |
| ralonsoh | haleyb, hello! Can you +W this patch? https://review.opendev.org/c/openstack/neutron/+/938541 | 21:05 |
| ralonsoh | thanks | 21:05 |
| haleyb | ralonsoh: will look | 21:06 |
| opendevreview | Vasyl Saienko proposed openstack/neutron master: Inherit vif|vnic_type and host for trunk subports https://review.opendev.org/c/openstack/neutron/+/942837 | 21:24 |
| haleyb | jizaymes: it looks like a network ID was not provided, which is normally in the request header. What release? | 22:35 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!