opendevreview | weilingfei proposed openstack/neutron master: Update url and package name https://review.opendev.org/c/openstack/neutron/+/863013 | 06:20 |
---|---|---|
*** ministry is now known as __ministry | 08:11 | |
opendevreview | Slawek Kaplonski proposed openstack/neutron-tempest-plugin master: [Stateless SG] Add test to check connectivity between vms https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/874820 | 08:11 |
opendevreview | Slawek Kaplonski proposed openstack/neutron-tempest-plugin master: [Stateless SG] Ensure replies won't work without ingress rule https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/877092 | 08:13 |
lajoskatona | ralonsoh (& all): Hi, if there are others with owner membership on pypi for projects yesterday fungi/jayf and others helped to solve the issue, see the irc log: https://meetings.opendev.org/irclogs/%23openstack-infra/%23openstack-infra.2023-03-22.log.html#t2023-03-22T17:24:12 | 08:14 |
lajoskatona | ralonsoh (&all) : and the related doc clarification: https://review.opendev.org/c/opendev/infra-manual/+/878240 | 08:15 |
lajoskatona | ralonsoh (&all): so if there's another owner the original human project owner can remove him/herself from the group and the other owner must be openstackci | 08:16 |
opendevreview | Max proposed openstack/neutron master: ovn metadata agent: stop fetching all portbindings https://review.opendev.org/c/openstack/neutron/+/878320 | 09:46 |
opendevreview | Dmitriy Rabotyagov proposed openstack/neutron-vpnaas master: Remove `restart-by-peer` from dpd actions https://review.opendev.org/c/openstack/neutron-vpnaas/+/878321 | 10:07 |
opendevreview | Dmitriy Rabotyagov proposed openstack/neutron-vpnaas master: Add dpdaction `none` https://review.opendev.org/c/openstack/neutron-vpnaas/+/878322 | 10:21 |
frickler | so for qos, max_burst_kbps actually is kilobits, not kilobits/s, right? and the name of the parameter is a mistake that can't be fixed due to api stability? | 11:14 |
frickler | the option in OSC is named --max-burst-kbits, I wonder if it would be less confusing to also change the field name in OSC output at least | 11:16 |
lajoskatona | frickler: by api-ref it's kilobits | 12:23 |
frickler | lajoskatona: yes, but why isn't the attribute named max_burst_kbits then? | 12:33 |
ralonsoh | we had this discussion some time (years) ago | 12:35 |
ralonsoh | that requires a CLI, database and API change | 12:35 |
ralonsoh | this is indeed an error introduced initially | 12:35 |
ralonsoh | but we didn't want to spend time changing that | 12:36 |
slaweq | ha, I remember that I even had patches proposed for that but then we didn't had shim extensions and there wasn't way to make this change discoverable | 12:37 |
slaweq | so we decided to abandon those changes :) | 12:37 |
ralonsoh | right, that was you | 12:37 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: Add neutron-lib project to the SQLAlchemy master branch CI jobs https://review.opendev.org/c/openstack/neutron/+/878359 | 12:40 |
frickler | kind of like I expected, thx for confirming | 12:40 |
opendevreview | Slawek Kaplonski proposed openstack/neutron-tempest-plugin master: [Stateless SG] Test if TCP packets with various conn state are allowed https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/877106 | 13:28 |
opendevreview | Slawek Kaplonski proposed openstack/neutron-tempest-plugin master: [Stateless SG] Test fragmented traffic is allowed by stateless SG https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/877480 | 13:28 |
slaweq | ralonsoh mlavalle hi, I addressed Your comments in https://review.opendev.org/q/topic:issue/OSP-20968+status:open - please review it again once CI will finish its work and You will have some time :) | 13:30 |
slaweq | thx in advance | 13:30 |
ralonsoh | yw! | 13:30 |
mlavalle | slaweq: ok. I think I don't have any concerns. I just have a different opinion regarding https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/877092. So I +2ed it. Let's see what ralonsoh thinks | 13:33 |
slaweq | mlavalle Your opinion is similar to mine there :) | 13:33 |
mlavalle | :-) | 13:33 |
ralonsoh | slaweq, qq, how that was working before? in particular for the negative cases | 13:40 |
ralonsoh | https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/877092/7/neutron_tempest_plugin/scenario/test_security_groups.py | 13:41 |
slaweq | ralonsoh I don't think we were using nc_client for negative cases before | 13:42 |
slaweq | it's first test like that if I'm not mistaken | 13:42 |
ralonsoh | ok then that makes sense | 13:43 |
ralonsoh | let me check again the code | 13:43 |
slaweq | thx ralonsoh | 13:56 |
opendevreview | Dmitrii Shcherbakov proposed openstack/neutron master: Add extra router attributes for ECMP and BFD https://review.opendev.org/c/openstack/neutron/+/874797 | 14:00 |
opendevreview | Arnaud Morin proposed openstack/neutron master: Add new DEFAULT option named "my_ip" https://review.opendev.org/c/openstack/neutron/+/878070 | 14:04 |
opendevreview | Arnaud Morin proposed openstack/neutron master: Fix doc links for networking option 2 https://review.opendev.org/c/openstack/neutron/+/878417 | 14:08 |
opendevreview | Arnaud Morin proposed openstack/neutron master: Add new DEFAULT option named "my_ip" https://review.opendev.org/c/openstack/neutron/+/878070 | 14:15 |
opendevreview | Arnaud Morin proposed openstack/neutron master: Fix doc links for networking option 2 https://review.opendev.org/c/openstack/neutron/+/878417 | 14:15 |
haleyb | amorin: you didn't need to rebase ^^ it could merge on its own as it's unrelated to my_ip change | 14:16 |
haleyb | i think at least | 14:17 |
haleyb | oh, now i see all the other changes, never mind | 14:17 |
haleyb | i guess i'm back to my original assertion it didn't need a rebase | 14:21 |
opendevreview | Luis Tomas Bolivar proposed openstack/neutron master: WIP: ensure redirect-type=bridged not used for geneve networks https://review.opendev.org/c/openstack/neutron/+/878450 | 15:51 |
dtantsur | Greetings folks! We're seeing failures in some of the Ironic grenade runs, where the floating IP cannot be pinged, although the instance seems up and healthy. | 15:56 |
dtantsur | All I kind find so far is guru meditation in q-l3 and q-agt like this: https://zuul.opendev.org/t/openstack/build/d90374e9b6554704a2f84b7fe8a9d411/log/controller/logs/screen-q-l3.txt | 15:56 |
dtantsur | and example run is https://zuul.opendev.org/t/openstack/build/d90374e9b6554704a2f84b7fe8a9d411/logs | 15:56 |
dtantsur | any ideas? | 15:56 |
dtantsur | seems like it started yesterday: https://zuul.opendev.org/t/openstack/builds?job_name=ironic-grenade&project=openstack/ironic | 15:57 |
opendevreview | Arnau Verdaguer proposed openstack/neutron master: [OVN] Avoid setting requested-chassis on trunk subports https://review.opendev.org/c/openstack/neutron/+/877006 | 16:15 |
haleyb | dtantsur: i don't have much time to help, but the q-svc log is showing some port binding failures, don't know if it's related. https://7c830b912b5951e3da88-240d4aa44573f96d2050363ab0d496a6.ssl.cf5.rackcdn.com/877600/4/check/ironic-grenade/d90374e/controller/logs/screen-q-svc.txt | 16:24 |
dtantsur | I don't know to which extent port binding is fatal for bare metal (at some point we did not support it) | 16:28 |
opendevreview | Merged openstack/neutron master: Add neutron-lib project to the SQLAlchemy master branch CI jobs https://review.opendev.org/c/openstack/neutron/+/878359 | 16:32 |
ralonsoh | dtantsur, the VM is not created | 16:51 |
ralonsoh | Mar 23 16:37:48.259425 np0033552931 devstack@n-api.service[84571]: INFO nova.api.openstack.wsgi [None req-2df3fe9d-cc57-42da-acb0-5cfa26c9fafe nova_grenade nova_grenade] HTTP exception thrown: Instance nova_server1 could not be found. | 16:51 |
ralonsoh | what is grenade doing at this point? | 16:51 |
dtantsur | ralonsoh: it is created just some minutes before the failure: https://zuul.opendev.org/t/openstack/build/d90374e9b6554704a2f84b7fe8a9d411/log/controller/logs/grenade.sh_log.txt#1665 | 16:53 |
opendevreview | Merged openstack/neutron stable/wallaby: Fix default value for MTUs, when not provided https://review.opendev.org/c/openstack/neutron/+/878142 | 16:55 |
ralonsoh | dtantsur, no, this VM is never created | 16:55 |
ralonsoh | https://zuul.opendev.org/t/openstack/build/d90374e9b6554704a2f84b7fe8a9d411/log/controller/logs/screen-n-api.txt#3776 | 16:55 |
opendevreview | Merged openstack/neutron stable/wallaby: [OVN] Explicitly define the fixed IPs for the metadata port https://review.opendev.org/c/openstack/neutron/+/877985 | 16:55 |
dtantsur | wow, fun | 16:57 |
dtantsur | then I don't quite understand how " openstack server show nova_server1 -f value -c id" succeeds just 2 seconds after this failure... | 16:58 |
dtantsur | I wonder if Nova supports names in URL, like in "GET /compute/v2.1/servers/nova_server1" | 16:58 |
ralonsoh | because the server is being created at this point | 16:58 |
ralonsoh | and has an ID | 16:58 |
ralonsoh | but the VM is not ready (nor created) | 16:59 |
dtantsur | There are no VMs involved | 16:59 |
ralonsoh | ok, the server | 16:59 |
dtantsur | What I don't understand is why GET with UUID works, but with name does not | 16:59 |
ralonsoh | what command? server show? | 16:59 |
dtantsur | both "server create --wait" and "server show" return successfully | 17:00 |
ralonsoh | actually the server create is using wait | 17:00 |
ralonsoh | and OS-EXT-STS:power_state | Running | 17:00 |
dtantsur | "GET /compute/v2.1/servers/ae01b6c4-74fa-4e14-980d-051f52403072" works, "GET /compute/v2.1/servers/nova_server1" does not | 17:00 |
ralonsoh | this could be related with the API | 17:00 |
ralonsoh | that is only accepting the UUID | 17:01 |
ralonsoh | nothing else | 17:01 |
ralonsoh | this is a red herring | 17:01 |
dtantsur | hmmm, yeah, maybe it's OSC or SDK trying to fetch my name (dunno why) | 17:02 |
dtantsur | User-Agent: python-novaclient | 17:02 |
ralonsoh | this is an "error" in the CLI | 17:03 |
ralonsoh | GET /compute/v2.1/servers/nova_server1 | 17:03 |
ralonsoh | vs | 17:03 |
ralonsoh | GET /compute/v2.1/servers?name=nova_server1 | 17:03 |
dtantsur | right, I see | 17:03 |
ralonsoh | we handle this correctly in the neutron cli (checking the string is a UUID or not) | 17:03 |
dtantsur | Floating ip 22ba5523-6b33-4318-8de1-1a76d86f36ab added, status ACTIVE | 17:07 |
dtantsur | (looking at the l3 logs) | 17:08 |
ralonsoh | dtantsur, what is is owner of the ironic port? | 17:10 |
ralonsoh | 7c306c7b-3a1e-4b61-9117-3cc192cfd19f | 17:10 |
ralonsoh | baremetal:xxx | 17:11 |
ralonsoh | right? | 17:11 |
dtantsur | ironic ports don't have owners really. or do I misunderstand your quetion? | 17:12 |
ralonsoh | no, that's ok | 17:12 |
ralonsoh | the port is unbound | 17:12 |
ralonsoh | Mar 23 08:29:13.249970 np0033547003 neutron-server[85439]: DEBUG neutron.plugins.ml2.plugin [None req-64c4002f-5399-442e-8ddf-48e591f82ff2 None None] Port 7c306c7b-3a1e-4b61-9117-3cc192cfd19f cannot update to ACTIVE because it is not bound. {{(pid=85439) _port_provisioned /opt/stack/old/neutron/neutron/plugins/ml2/plugin.py:361}} | 17:12 |
dtantsur | it's interesting, networking-baremetal is supposed to fix up the port binding.. should not be fatal though | 17:13 |
dtantsur | cc hjensas who actually understands these things | 17:14 |
ralonsoh | dtantsur, | 17:25 |
ralonsoh | Mar 23 08:29:55.004621 np0033547003 neutron-server[85437]: DEBUG neutron.plugins.ml2.drivers.mech_agent [req-2204d326-80b3-4faf-9ac1-bbcf6cdd69bc req-3bdc0205-363c-4346-9aff-74a83f52e50f service ironic] Attempting to bind port 96721476-7aed-4be0-a718-486a300e1f5b on network 53c0501a-6c76-408d-b319-0e2c0a1addac {{(pid=85437) bind_port /opt/stack/old/neutron/neutron/plugins/ml2/drivers/mech_agent.py:91}} | 17:25 |
ralonsoh | Mar 23 08:29:55.004621 np0033547003 neutron-server[85437]: DEBUG neutron.plugins.ml2.drivers.mech_agent [req-2204d326-80b3-4faf-9ac1-bbcf6cdd69bc req-3bdc0205-363c-4346-9aff-74a83f52e50f service ironic] Refusing to bind due to unsupported vnic_type: baremetal {{(pid=85437) bind_port /opt/stack/old/neutron/neutron/plugins/ml2/drivers/mech_agent.py:98}} | 17:25 |
dtantsur | ralonsoh: yeah, it's probably a red herring as well | 17:26 |
dtantsur | I wonder why it does not happen with networking-baremetal, but we'd worked like that for years before networking-baremetal | 17:26 |
frickler | dtantsur: two days ago grenade switched from zed as base to 2023.1 https://review.opendev.org/c/openstack/grenade/+/877998 not sure if that is related to your issue though | 17:37 |
dtantsur | Could be | 17:37 |
frickler | method: <bound method ConsoleOutputController.get_console_output ... exception: The requested functionality is not supported. | 17:41 |
frickler | sounds more like a nova issue? | 17:41 |
opendevreview | Merged openstack/ovn-octavia-provider stable/xena: Fix broken pep8 jobs due to bandit 1.7.5 updated version https://review.opendev.org/c/openstack/ovn-octavia-provider/+/877465 | 17:44 |
opendevreview | Merged openstack/ovn-octavia-provider stable/wallaby: Fix broken pep8 jobs due to bandit 1.7.5 updated version https://review.opendev.org/c/openstack/ovn-octavia-provider/+/877466 | 17:46 |
gmann | slaweq: do you want to run enforce scope zed job on check/gate pipeline? we can do in periodic/experimental also, I am doing same in tempest https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/874709/7/zuul.d/project.yaml#150 | 17:47 |
dtantsur | frickler: no, it happens while collecting debug data after failure | 18:00 |
frickler | ah, I need to better correlate timestamps I guess | 18:06 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: [sqlalchemy-20] query.join should define one table/column per call https://review.opendev.org/c/openstack/neutron/+/878479 | 19:20 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: [sqlalchemy-20] Do not use strings for aatribute names in loader options https://review.opendev.org/c/openstack/neutron/+/878480 | 19:20 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: [sqlalchemy-20] query.join should define one table/column per call https://review.opendev.org/c/openstack/neutron/+/878479 | 19:36 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: [sqlalchemy-20] Do not use strings for aatribute names in loader options https://review.opendev.org/c/openstack/neutron/+/878480 | 19:36 |
opendevreview | Merged openstack/ovn-octavia-provider stable/zed: Fix broken pep8 jobs due to bandit 1.7.5 updated version https://review.opendev.org/c/openstack/ovn-octavia-provider/+/877463 | 21:57 |
*** promethe- is now known as prometheanfire | 23:08 | |
*** seebaer is now known as seba | 23:19 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!