| opendevreview | ZhouHeng proposed openstack/neutron master: [ovn]l3 plugin support floating ip distributed attribues https://review.opendev.org/c/openstack/neutron/+/856955 | 06:33 |
|---|---|---|
| opendevreview | Merged openstack/neutron master: Removed duplicated job call in experimental queue https://review.opendev.org/c/openstack/neutron/+/872274 | 06:47 |
| opendevreview | Luis Tomas Bolivar proposed openstack/ovn-octavia-provider stable/zed: Remove LB from LS belonging to provider networks https://review.opendev.org/c/openstack/ovn-octavia-provider/+/872577 | 06:48 |
| opendevreview | Maurice Escher proposed openstack/neutron master: allow manila ports to do multiple port binding for ML2 https://review.opendev.org/c/openstack/neutron/+/869295 | 08:21 |
| lajoskatona | slaweq, ykarel: Hi, Could you please check this small one for n-d-r when you have some free time: https://review.opendev.org/c/openstack/neutron-dynamic-routing/+/872384 ? This is another one for qslalchemy 2 | 08:31 |
| ykarel | lajoskatona, ack | 08:34 |
| lajoskatona | ykarel: thanks | 08:39 |
| opendevreview | ZhouHeng proposed openstack/neutron master: [ovn]l3 plugin support floating ip distributed attribues https://review.opendev.org/c/openstack/neutron/+/856955 | 08:46 |
| *** gibi_pto is now known as gibi | 08:47 | |
| opendevreview | Slawek Kaplonski proposed openstack/neutron-tempest-plugin master: [Secure RBAC] Enable scope enforcing in jobs run on master branch https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/867518 | 09:12 |
| opendevreview | Slawek Kaplonski proposed openstack/neutron-tempest-plugin master: Clean definitions of the CI jobs for EOL'ed branches https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/872736 | 09:15 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [OVN] Skip ports with VNIC type direct and no port binding https://review.opendev.org/c/openstack/neutron/+/872737 | 09:30 |
| opendevreview | Tom Weininger proposed openstack/ovn-octavia-provider master: Replace python-neutronclient with openstacksdk https://review.opendev.org/c/openstack/ovn-octavia-provider/+/870514 | 09:54 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Filter subnets by "enable_dhcp" flag using the correct type https://review.opendev.org/c/openstack/neutron/+/872745 | 10:18 |
| opendevreview | Krzysztof Tomaszewski proposed openstack/neutron stable/victoria: [ovn]neutron agent show real heartbeat_timestamp https://review.opendev.org/c/openstack/neutron/+/872683 | 10:32 |
| opendevreview | Tom Weininger proposed openstack/ovn-octavia-provider master: Replace python-neutronclient with openstacksdk https://review.opendev.org/c/openstack/ovn-octavia-provider/+/870514 | 10:44 |
| opendevreview | Merged openstack/ovn-octavia-provider stable/xena: Remove LB from LS belonging to provider networks https://review.opendev.org/c/openstack/ovn-octavia-provider/+/872579 | 11:18 |
| opendevreview | Merged openstack/ovn-octavia-provider stable/wallaby: Remove LB from LS belonging to provider networks https://review.opendev.org/c/openstack/ovn-octavia-provider/+/872580 | 11:19 |
| opendevreview | Merged openstack/ovn-octavia-provider stable/zed: Remove LB from LS belonging to provider networks https://review.opendev.org/c/openstack/ovn-octavia-provider/+/872577 | 11:19 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Adapt deploy_rootwrap filters path for tox4 https://review.opendev.org/c/openstack/neutron/+/867554 | 11:28 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: DNM Trying to get fullstack logs with tox4 https://review.opendev.org/c/openstack/neutron/+/868380 | 11:28 |
| opendevreview | Tom Weininger proposed openstack/ovn-octavia-provider master: Replace python-neutronclient with openstacksdk https://review.opendev.org/c/openstack/ovn-octavia-provider/+/870514 | 11:33 |
| opendevreview | Felix Huettner proposed openstack/neutron-lib master: Update status code for ExternalGatewayForFloatingIPNotFound https://review.opendev.org/c/openstack/neutron-lib/+/872748 | 12:23 |
| opendevreview | Lajos Katona proposed openstack/neutron-lib master: model_query: Improve perf with large # of RBAC https://review.opendev.org/c/openstack/neutron-lib/+/817459 | 12:31 |
| opendevreview | Tom Weininger proposed openstack/ovn-octavia-provider master: Replace python-neutronclient with openstacksdk https://review.opendev.org/c/openstack/ovn-octavia-provider/+/870514 | 13:16 |
| opendevreview | Merged openstack/os-vif master: Increase the swap size to 8GB in tempest jobs https://review.opendev.org/c/openstack/os-vif/+/872655 | 13:16 |
| *** dasm|off is now known as dasm|rover | 13:52 | |
| bcafarel | ykarel: hi I was checking ussuri https://review.opendev.org/c/openstack/neutron/+/871989 (broken grenade) - is there a LP for the paramiko train issue? Or fix in progress? | 13:58 |
| bcafarel | ykarel: given ussuri EM status, scaling back grenade to non-voting may also be option | 13:59 |
| ykarel | bcafarel, i see you rechecked, is it still failing? | 13:59 |
| bcafarel | ykarel: yes https://zuul.opendev.org/t/openstack/build/2248c27a9d5f4dae895058f0e8589115 | 14:01 |
| ykarel | bcafarel, ack let me check if its same issue or not | 14:03 |
| elodilles | bcafarel: ussuri and train is broken for multiple projects due to cryptography (bcrypt) installation fails, if i'm not mistaken: https://zuul.opendev.org/t/openstack/build/2248c27a9d5f4dae895058f0e8589115/log/job-output.txt#15037 | 14:04 |
| elodilles | it is probably because some old pre-built wheel disappeared from pypi, and now cryptography requires rust compiler to build | 14:06 |
| ykarel | bcafarel, yes it's same | 14:07 |
| elodilles | i don't know yet how to fix it... (with newer pip the error disappears though) | 14:07 |
| ykarel | elodilles, yes with newer pip it works | 14:09 |
| ykarel | seems you trying to fix this issue only with https://review.opendev.org/q/topic:ensure-rust-cryptography ? | 14:09 |
| elodilles | ykarel: i tried that but it was not enough | 14:09 |
| elodilles | if you have a better/working fix then please let me know o:) (as multiple projects are failing on it, as I said :S) | 14:12 |
| ykarel | iirc older pip don't recognize newer wheels that's why it's not pulling wheels | 14:13 |
| ykarel | https://pypi.org/project/cryptography/39.0.0/#files | 14:13 |
| elodilles | hmmm, i see, thx for the info! | 14:16 |
| opendevreview | Merged openstack/ovsdbapp master: Define "LbAddIpPortMappingCommand" with latin chars https://review.opendev.org/c/openstack/ovsdbapp/+/872200 | 14:25 |
| ykarel | jfr found https://github.com/pypa/pip/issues/9077 | 14:27 |
| amarao | I'm trying to find a way to use segments (?) to allow automatic selection of different subnetworks for a given network. | 14:36 |
| amarao | Basically I have a normal (neutron managed/routed) network with multiple small (/28) subnetworks, and I want to be able to use a single network name during VM creation, so nova (neutron) are choosing subnet for me. | 14:36 |
| amarao | What extension can do this? 'segments' is tightly bounded by 'provider: physical_network' (which is None in my case). | 14:36 |
| felixhuettner[m] | Hi everyone, i just started working on https://bugs.launchpad.net/neutron/+bug/2006122 and noticed that i would need a cyclic "depends-on" for that as i need to touch neutron-lib and tempest at the same time. Could you help me with how i could accomplish that? | 14:38 |
| labedz_ | amarao: maybe Routed provider networks are for you: https://docs.openstack.org/neutron/latest/admin/config-routed-networks.html ? | 14:38 |
| ralonsoh | amarao, what exactly do you need? to be able to provide an IP on a subnet? | 14:38 |
| ralonsoh | if this is the case, you don't need segments | 14:39 |
| amarao | I have host bgp to terminate routes on the network host, and then there is openstack router doing routing for clients, so routed-networks is not working for that case (at least not within docs scope). | 14:39 |
| ralonsoh | right so what you should need is to create a port with an assigned IP on the required subnet | 14:41 |
| ralonsoh | and then create the VM with this port | 14:41 |
| amarao | The problem I'm trying to solve is that multiple tenants (without NAT) in the same overlay network is a mess, so we need to keep networks small. But this creates API problem: which network to create instance in. We want one network 'internet' with multiple small subnets. I've tried, but nova is actively resisted: NovaException: None of the segments of network {uuid} provides a physical_network | 14:41 |
| ralonsoh | if you are using L3 in Neutron, do not use multiple segments | 14:42 |
| amarao | And there is no physical_network, because it's not 'physical L2 networks', as they are created and managed by OS. | 14:42 |
| ralonsoh | those are two different concepts | 14:42 |
| ralonsoh | when do you have this message? | 14:42 |
| ykarel | elodilles, what's missing in those fixes? i only see options:- upgrade pip or ensure rust or somehow pin paramkiko<3.0.0 | 14:43 |
| amarao | I've created vxlan network with muliple segments (neutron router from 'connection' network to OS network), and tried to create instance in it. | 14:43 |
| ralonsoh | amarao, please check the documentation for routed segment networks | 14:44 |
| amarao | Trace (zed): https://gist.github.com/amarao/3f894326707e85ebd08b7af06059f758 | 14:44 |
| ralonsoh | this is the Nova exception | 14:45 |
| ralonsoh | that gives no information aboyt the issue in Neutron | 14:45 |
| ralonsoh | amarao, please open a launchpad bug describing the issue you have | 14:45 |
| ralonsoh | providing logs and how to reproduce it | 14:45 |
| amarao | Yes, it's hardcoded in Nova, I've checked. | 14:45 |
| amarao | But I thought it's neutron domain. | 14:45 |
| amarao | I've read routed-networks, they propose to use segments with provider-network, and I have none for my network. I've just come to ask, may be there is some other extention to work with L3 but with a single network name... | 14:46 |
| amarao | Btw, are there list of extentions for neutron with extended description? Just listing through API is a bit too concise. | 14:48 |
| ralonsoh | no that I'm aware, you can list those supported and check in the documentation for more info | 14:50 |
| ralonsoh | felixhuettner[m], you can re-implement temporarily "ExternalGatewayForFloatingIPNotFound" in Neutron code, inheriting from BadRequest instead of NotFound with a note saying to bump the n-lib version | 14:51 |
| ralonsoh | felixhuettner[m], be careful if this exception is used anywhere else | 14:51 |
| ralonsoh | https://codesearch.openstack.org/?q=ExternalGatewayForFloatingIPNotFound&i=nope&literal=nope&files=&excludeFiles=&repos= | 14:51 |
| amarao | So, basically, it's either provider L2 (and then there are segments), or no luck, because L3 routed network shouldn't have segments... | 14:52 |
| opendevreview | yatin proposed openstack/neutron stable/ussuri: [DNM] Check upgrades https://review.opendev.org/c/openstack/neutron/+/872756 | 14:52 |
| ralonsoh | amarao, please, describe what are you trying to achieve. If that is written in a launchpad bug, that will be better and we'll help us you in a better way | 14:53 |
| amarao | I created configuration where I have L2 'connection network' between host and openstack router (it's 'flat' network for openstack). In reality this network is routing traffic via bgp with provider. Actual networks for tenants are created in the neuton in the same address scope as 'connection network', so there is no NAT, only routing. | 14:56 |
| amarao | After that I start to wonder if I can replace multiple 'networks' in API, with one 'the network' to connect to, but with subnets (and openstack-managed segments underneath). | 14:56 |
| felixhuettner[m] | <ralonsoh> "felix.huettner, you can re-..." <- thanks for the hint | 15:00 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: rbacs: filter out model that are already owned by context https://review.opendev.org/c/openstack/neutron/+/871113 | 15:23 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: rbacs: tests mix on networks and users https://review.opendev.org/c/openstack/neutron/+/872761 | 15:23 |
| sahid | ralonsoh: o/ - just small heads-up regarding the change for rbacs, i have added a functional test, but it is not part of the change, so we can test the current behavior, against the change | 15:25 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: rbacs: add functionl test that asserts behavior on networks/users https://review.opendev.org/c/openstack/neutron/+/872761 | 15:34 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: rbacs: filter out model that are already owned by context https://review.opendev.org/c/openstack/neutron/+/871113 | 15:34 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: rbacs: fix typo, s/cxt/ctx https://review.opendev.org/c/openstack/neutron/+/872762 | 15:34 |
| *** marlinc is now known as Guest3760 | 15:46 | |
| ykarel | elodilles, i did a quick test at https://review.opendev.org/c/openstack/neutron/+/872756 and with that it works | 15:50 |
| elodilles | ykarel: paramiko is pinned (ussuri: paramiko===2.7.1; train: paramiko===2.6.0). i've tried the ensure-rust so far and it seemed to be not working. i haven't tried to use newer pip (i remember something was failing when we started to use newer pip version, that's why it is pinned too). | 15:58 |
| elodilles | ykarel: so your patch is different than mine by adding ensure-rust directly to the job definition in neutron repository | 15:59 |
| elodilles | ykarel: while i added it in the common job definition | 15:59 |
| ykarel | elodilles, yes it's pinned but in tempest-26.1.0(train version) it's set to paramiko>=2.7.0 so it attempts to install latest available paramiko(which then attempts to source(because of old pip) install cryptography>=3.3) | 16:01 |
| ykarel | elodilles, you did it in master branch of grenade but job is ussuri/train so need to patch those branches | 16:02 |
| elodilles | oh, hmmmm. so you say tempest does not use upper constraints | 16:02 |
| ykarel | also need to fix non grenade jobs too in train | 16:02 |
| ykarel | elodilles, it does while setting up deps but then while installing tempest itself use_develop=True , paramiko requirement is not matched so it tries to udpate | 16:04 |
| elodilles | ouch :S | 16:07 |
| elodilles | hmm, yes, you are right with my change: i've added it on the master branch (somehow i remembered i added it some branchless repo - like openstack-zuul-jobs - but grenade is clearly branched) | 16:09 |
| ykarel | yeap | 16:09 |
| elodilles | ykarel: in that case the grenades will work hopefully. but as you say, we still have the tempest jobs on train :/ | 16:10 |
| ykarel | yes for that need to do this in some other playbook | 16:13 |
| ykarel | running in those train jobs | 16:14 |
| * ykarel need to drop now will catch up tomorrow | 16:16 | |
| elodilles | ykarel: thx for the heads up, have a nice evening o/ | 16:21 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [OVN] Skip ports with VNIC type direct and no port binding https://review.opendev.org/c/openstack/neutron/+/872737 | 16:27 |
| opendevreview | Merged openstack/os-vif master: Implement "BaseCommand" result property https://review.opendev.org/c/openstack/os-vif/+/872391 | 18:32 |
| *** dmitriis9 is now known as dmitriis | 21:51 | |
| *** seebaer is now known as seba | 21:51 | |
| *** EugenMayer43 is now known as EugenMayer4 | 21:51 | |
| *** noonedeadpunk_ is now known as noonedeadpunk | 21:51 | |
| *** marlinc is now known as Guest3907 | 21:57 | |
| opendevreview | Miguel Lavalle proposed openstack/neutron master: [WIP][DNM][OVN] Change the default firewall policy https://review.opendev.org/c/openstack/neutron/+/839066 | 23:35 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!