opendevreview | Miguel Lavalle proposed openstack/neutron master: Test https://review.opendev.org/c/openstack/neutron/+/821143 | 01:47 |
---|---|---|
opendevreview | yatin proposed openstack/neutron-tempest-plugin master: [WIP] Switch scenario jobs to virt nodes https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/821067 | 04:38 |
opendevreview | Merged openstack/neutron master: Improve Router callback system's publish events https://review.opendev.org/c/openstack/neutron/+/804846 | 05:58 |
opendevreview | yatin proposed openstack/neutron-tempest-plugin master: Update irrelevant-files for api/scenario jobs https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/821160 | 07:13 |
opendevreview | yatin proposed openstack/neutron-tempest-plugin master: Update irrelevant-files for api/scenario jobs https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/821160 | 07:20 |
*** gibi_ is now known as gibi | 07:52 | |
opendevreview | Maor Blaustein proposed openstack/neutron-tempest-plugin master: Add test_create_port_with_dns_name https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/820456 | 08:00 |
lajoskatona | ralonsoh: Hi, am I right that we don't need another discussion for https://bugs.launchpad.net/neutron/+bug/1953170 ([RFE] Unify quota engine API) ? | 08:16 |
ralonsoh | lajoskatona, that's my opinion, I think community accepted the steps provided | 08:17 |
ralonsoh | to move to the new quota behaviour | 08:17 |
lajoskatona | ralonsoh: thanks, that was my understanding also | 08:19 |
opendevreview | Oleg Bondarev proposed openstack/neutron master: Add Local IP fullstack test cases https://review.opendev.org/c/openstack/neutron/+/816327 | 08:20 |
slaweq | lajoskatona that was also my understanding as well :) | 08:22 |
slaweq | ralonsoh lajoskatona can You check https://review.opendev.org/c/openstack/neutron/+/799625 when You will have some time? | 08:32 |
ralonsoh | sure | 08:32 |
slaweq | it should helps with FIPS enabled jobs (maybe it's not critical but will be needed with: https://review.opendev.org/c/openstack/devstack/+/799621 ) | 08:32 |
slaweq | ralonsoh also https://review.opendev.org/c/openstack/devstack/+/799621 seems to be fine now | 08:33 |
ralonsoh | checking now | 08:33 |
slaweq | I will do some DNM patch in Neutron to test it with all our jobs but zuul on that patch is happy now | 08:33 |
opendevreview | Slawek Kaplonski proposed openstack/neutron master: DNM It's just test of the Devstack patch https://review.opendev.org/c/openstack/neutron/+/821163 | 08:35 |
opendevreview | yatin proposed openstack/neutron stable/wallaby: Add retry when executing OF commands if "InvalidDatapath" https://review.opendev.org/c/openstack/neutron/+/821174 | 08:39 |
opendevreview | yatin proposed openstack/neutron stable/victoria: Add retry when executing OF commands if "InvalidDatapath" https://review.opendev.org/c/openstack/neutron/+/821175 | 08:40 |
opendevreview | Maor Blaustein proposed openstack/neutron-tempest-plugin master: Add test_create_port_with_dns_name https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/820456 | 08:41 |
opendevreview | Slawek Kaplonski proposed openstack/neutron master: Add FIPS enabled scenario jobs https://review.opendev.org/c/openstack/neutron/+/797537 | 08:43 |
slaweq | lajoskatona just FYI, I read pain points etherpad today and I opened https://bugs.launchpad.net/neutron/+bug/1953716 | 08:48 |
slaweq | I will work on it | 08:48 |
slaweq | I hope it should be easy to do | 08:48 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: Quota engine to accept --force parameter in limit update https://review.opendev.org/c/openstack/neutron/+/821190 | 09:39 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: Mock the OVN events providing a "priority" value https://review.opendev.org/c/openstack/neutron/+/820911 | 09:43 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: Provide a "priority" value to the OVN events https://review.opendev.org/c/openstack/neutron/+/820911 | 09:44 |
mdbooth | Hi, just trying to debug a CI failure in a configuration which normally works and hasn't been changed (2 node devstack on stable victoria). Server failed to come up on the worker with PortBindingFailed. I see the following 2 lines earlier in the neutron logs on the controller: | 10:03 |
mdbooth | Dec 08 23:25:22 [controller-devstack.log] capo-e2e-controller neutron-server[104782]: DEBUG neutron.plugins.ml2.drivers.mech_agent [req-4edf64dc-5644-409a-bbf6-2b06c9b5e63c req-06eb50f4-4241-4a3d-99db-038ddfad2cd4 service neutron] Port e3517cda-7198-4483-b868-921e83cec8ea on network 1a4f49b3-95d6-4590-a3e3-84924ce4235e not bound, no agent of type | 10:03 |
mdbooth | Linux bridge agent registered on host capo-e2e-worker {{(pid=104782) bind_port /opt/stack/neutron/neutron/plugins/ml2/drivers/mech_agent.py:97}} | 10:03 |
mdbooth | Dec 08 23:25:22 [controller-devstack.log] capo-e2e-controller neutron-server[104782]: ERROR neutron.plugins.ml2.managers [req-4edf64dc-5644-409a-bbf6-2b06c9b5e63c req-06eb50f4-4241-4a3d-99db-038ddfad2cd4 service neutron] Failed to bind port e3517cda-7198-4483-b868-921e83cec8ea on host capo-e2e-worker for vnic_type normal using segments [{'id': | 10:03 |
mdbooth | '6e9a9521-a90f-476b-b836-2888ec1385ea', 'network_type': 'vxlan', 'physical_network': None, 'segmentation_id': 1707, 'network_id': '1a4f49b3-95d6-4590-a3e3-84924ce4235e'}] | 10:03 |
mdbooth | Is that interesting or a 'normal failure'? Just re-running to see if it reproduces, but posting in case it rings a bell. | 10:03 |
ignaziocassano | Hello, is there any way to integrate neutron vxlan with vmware nsx ? | 10:21 |
ignaziocassano | In other words nsx could be used as a ovn controller ? | 10:22 |
*** redrobot6 is now known as redrobot | 10:23 | |
ralonsoh | ignaziocassano, what does it means to use nsx as ovn-controller? You'll need to run OVS and ovn-controller service in a node | 10:28 |
ralonsoh | mdbooth, did you configure the ml2.type_drivers? | 10:28 |
ralonsoh | type_drivers = flat,vlan,vxlan,geneve | 10:28 |
ralonsoh | for example | 10:28 |
ignaziocassano | ralonsoh, I never tried. Mi collegues want to use nsx-t but I yhink it is very expensive | 10:31 |
ignaziocassano | I do not know what can I do. At this thime for connectting vxlan between vcenter and openstack, we used a vlan a routing. | 10:32 |
ralonsoh | the point is that nsx provides a SDN by itself. I don't know if this is compatible somehow with OVN as backend | 10:33 |
ignaziocassano | I do not know what can I do. At this thime for connectting vxlan between vcenter and openstack, we used a vlan and routing. | 10:33 |
ralonsoh | ah, you mean connect openstack | 10:33 |
ignaziocassano | yes | 10:33 |
ralonsoh | you can use vxlan networks | 10:33 |
ralonsoh | and now vlan (with a static GW) | 10:33 |
ralonsoh | so yes, you can use both types | 10:34 |
ignaziocassano | I want to transport vxlan between vcenter and openstack | 10:34 |
ralonsoh | I see no problem, you can have vxlan provider networks with OVN | 10:34 |
ignaziocassano | ralonsoh: so I mus use ovn . I never used ovn. In this case the controller will be the nsx ? Is ther any detailed configuration ? | 10:35 |
ralonsoh | what controller? there are two levels | 10:36 |
ralonsoh | neutron controller or ovn controller | 10:36 |
ignaziocassano | ovn, I think | 10:37 |
ralonsoh | the ovn-controller is a service in each node to connect the SB DB with the local OVS instance | 10:37 |
ralonsoh | you can create a chassis (that will hold an ovn-controller) in a node | 10:38 |
ignaziocassano | ralonsoh: so I will not have a local openbswitch db but I centalized db ? | 10:38 |
ralonsoh | no | 10:38 |
ralonsoh | the NB and SB DB are centralized | 10:39 |
ralonsoh | the local OVS DB is local | 10:39 |
ralonsoh | the ovn-controller translates the SB registers into the local OVS DB | 10:39 |
ralonsoh | NB -> SB -> (local) OVS DB | 10:39 |
ignaziocassano | ralonsh: And how can I use nsx in this architecture ? | 10:40 |
ralonsoh | I have no idea, sorry | 10:40 |
ralonsoh | I don't know if you deploy another SDN on top of OVN | 10:40 |
ignaziocassano | ralansoh: no problem Thanks anycase | 10:41 |
opendevreview | Oleg Bondarev proposed openstack/neutron master: Add Local IP fullstack test cases https://review.opendev.org/c/openstack/neutron/+/816327 | 10:55 |
ignaziocassano | ralonsoh: l2gw and hardware vtep ? Vmware nsx supports hw vtep.If l2 gw works, we could connect vxlan without L3 but using L2 ? | 10:59 |
ralonsoh | I don;t know if OVN supports HW VTEP | 11:05 |
ralonsoh | it does, in theory, but I don't know how to configure it in Neutron (if we support it) | 11:06 |
ignaziocassano | ralonsoh: thanks again | 11:07 |
sean-k-mooney | ralonsoh: i think you woudl have to use the l2/l3gateway proejcts to configure external vteps | 11:34 |
sean-k-mooney | the core neutron api does not have the ablity to model them | 11:34 |
ralonsoh | no, we can't | 11:34 |
ralonsoh | and in OVN this is part of the OVN core | 11:34 |
ralonsoh | so in Neutron we can't model the VTEP connectivity | 11:35 |
ralonsoh | I know we can, in OVS, configure manually the HW VTEP | 11:35 |
ralonsoh | but not in OVN | 11:35 |
sean-k-mooney | well with ovn you can model them in the nortdb | 11:35 |
sean-k-mooney | but we woudl need an api extion to do so or use l* gateway proejct and enhacne them to support ovn | 11:36 |
sean-k-mooney | somethign bascially woudl have to expose ovns supprot for external top of rack vtps | 11:36 |
sean-k-mooney | and that at least for now feels like soemthing that is outside the scope of core neutron although it could be added if there was enough interest i guess | 11:37 |
sean-k-mooney | its really an admin only capablitity however | 11:37 |
sean-k-mooney | for most admins i suspect that flat/vlans provider networks possible with bgp integration will suffice | 11:38 |
sean-k-mooney | and as you said a second sdn can alway sbe ued to configure vteps at the TOR switch if they really need that out of band or neutron | 11:39 |
ralonsoh | sean-k-mooney, checking https://www.openvswitch.org/support/dist-docs-2.5/ovn-architecture.7.txt, this could be a bit more complex | 11:40 |
ralonsoh | we need to deploy another service, ovn-controller-vtep | 11:40 |
sean-k-mooney | yes you would | 11:40 |
ralonsoh | (I said nothing, we already deploy it) | 11:41 |
sean-k-mooney | proably not in all installers | 11:41 |
ralonsoh | right, I'm checking devstack | 11:41 |
ralonsoh | along with the schema creation, etc. | 11:41 |
sean-k-mooney | ya devstack proably does i suspect ooo does not | 11:42 |
sean-k-mooney | it might but if we dont test/use it its proably not there | 11:42 |
ralonsoh | right | 11:42 |
sean-k-mooney | anywya it shoudl eb therotically possible to deploy it and then add a neutron api exteion to allow defining the vteps if someone put in the work | 11:43 |
sean-k-mooney | im just not aware of anyone working on it now | 11:43 |
ralonsoh | I'll check that internally | 11:43 |
ralonsoh | that could be a good feature (not present in OVS ml2, if I'm not wrong) | 11:44 |
sean-k-mooney | its aviable via the l2 gateway and l3 gateway project | 11:44 |
sean-k-mooney | but its not in core neutorn | 11:44 |
sean-k-mooney | https://opendev.org/x/networking-l2gw | 11:45 |
ralonsoh | ah yes, right | 11:45 |
sean-k-mooney | https://opendev.org/x/networking-l2gw/src/commit/b20b90d7694f8c47e2275bc17995f4854eedf6ae/specs/newton/l2-border-gateway-implementation.rst | 11:45 |
sean-k-mooney | like networking bgp this might be a case wehre the api from that extion might want to be resued and the fucntioanlity delegated to ovn | 11:46 |
sean-k-mooney | but i agree it would be a nice feature to have | 11:47 |
sean-k-mooney | just not sure its an urgent one | 11:47 |
ralonsoh | right, that could be easier in OVN | 11:48 |
ralonsoh | and we can reuse the API | 11:48 |
opendevreview | Maor Blaustein proposed openstack/neutron-tempest-plugin master: Add test_create_port_with_dns_name https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/820456 | 12:23 |
ykarel | ralonsoh, slaweq when u get chance please have a look at backports https://review.opendev.org/q/Ia8d202f8a38362272e9519c1cbd9d6ba9359e0a1 | 12:48 |
ralonsoh | ykarel, but we didn't detect this problem in older versions | 12:49 |
ykarel | seeing those failures in stable/branches jobs | 12:49 |
ralonsoh | ykarel, do you have a link? | 12:49 |
ykarel | ralonsoh, yes http://logstash.openstack.org/#dashboard/file/logstash.json?query=message%3A%5C%22os_ken.app.ofctl.exception.InvalidDatapath%3A%20Datapath%20Invalid%5C%22%20AND%20build_branch%3A%5C%22stable%2Fwallaby%5C%22 | 12:50 |
ykarel | also seen in neutron-tempest-plugin where branched jobs run | 12:51 |
ykarel | similarly can see for victoria too | 12:51 |
ralonsoh | ykarel, then what you need to backport is https://review.opendev.org/c/openstack/neutron/+/815459 | 12:53 |
ykarel | ralonsoh, ahh you mean only ^ or both are needed? | 12:54 |
ralonsoh | last one is enough | 12:55 |
opendevreview | Slawek Kaplonski proposed openstack/neutron master: WIP Update secure RBAC policies accrodingly to the new guidelines https://review.opendev.org/c/openstack/neutron/+/821208 | 12:58 |
ykarel | ralonsoh, ohkk strange as 815459 is not in xena | 13:00 |
ykarel | and i don't see much failure in xena, only failures are in grenade jobs, which upgrades from wallaby to xena | 13:01 |
ykarel | Ia8d202f8a38362272e9519c1cbd9d6ba9359e0a1 is in xena so i thought that will also clear victoria/wallaby too so proposed backports | 13:03 |
ykarel | but likely you mean https://review.opendev.org/c/openstack/neutron/+/815459 is better workaround, right? | 13:04 |
ralonsoh | yes | 13:04 |
ykarel | okk will propose backport for that then | 13:04 |
ralonsoh | in any case, the problem should be also solved in os-ken | 13:04 |
ralonsoh | but we can't bump library versions | 13:04 |
ykarel | yes okk | 13:04 |
ykarel | ahh it conflicts during cherry-pick to xena itself | 13:06 |
ykarel | will check | 13:06 |
opendevreview | yatin proposed openstack/neutron stable/xena: [OVS] Workaround when OpenFlow controller restarts https://review.opendev.org/c/openstack/neutron/+/821211 | 13:17 |
opendevreview | yatin proposed openstack/neutron stable/wallaby: [OVS] Workaround when OpenFlow controller restarts https://review.opendev.org/c/openstack/neutron/+/821182 | 13:17 |
opendevreview | yatin proposed openstack/neutron stable/victoria: [OVS] Workaround when OpenFlow controller restarts https://review.opendev.org/c/openstack/neutron/+/821183 | 13:18 |
ykarel | ralonsoh, done ^ | 13:18 |
ykarel | will abandon the other one in favor of this | 13:18 |
*** ykarel is now known as ykarel|away | 13:21 | |
opendevreview | Merged openstack/neutron stable/ussuri: [OVN] Fix gateway_mtu option should not always be set https://review.opendev.org/c/openstack/neutron/+/819581 | 14:07 |
opendevreview | Tamas Gergely proposed openstack/neutron master: Check whether vxlan group and local addresses are IPv4 or IPv6 https://review.opendev.org/c/openstack/neutron/+/820376 | 15:11 |
opendevreview | Bence Romsics proposed openstack/neutron master: WIP Make the dead vlan actually dead https://review.opendev.org/c/openstack/neutron/+/820897 | 15:38 |
opendevreview | Slawek Kaplonski proposed openstack/neutron master: Update secure RBAC policies accrodingly to the new guidelines https://review.opendev.org/c/openstack/neutron/+/821208 | 15:57 |
opendevreview | Merged openstack/neutron master: Provide a "priority" value to the OVN events https://review.opendev.org/c/openstack/neutron/+/820911 | 16:57 |
*** weechat1 is now known as amorin | 17:54 | |
*** weechat1 is now known as amorin | 18:00 | |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: [WIP] When creating a VXLAN interface, a device is mandatory https://review.opendev.org/c/openstack/neutron/+/821271 | 18:13 |
slaweq | gmann lbragstad_ hi, I pushed today patch to change neutron's secure rbac according to what was agreed in the community goal (at least I think it's like it should be now), if You will have some time, please take a look at https://review.opendev.org/c/openstack/neutron/+/821208 | 21:14 |
slaweq | thx in advance | 21:14 |
opendevreview | Ade Lee proposed openstack/neutron master: Add FIPS enabled scenario jobs https://review.opendev.org/c/openstack/neutron/+/797537 | 22:06 |
gmann | slaweq: sure. I will be checking that next week early | 22:33 |
slaweq | thx gmann | 22:34 |
*** tbachman_ is now known as tbachman | 22:54 | |
*** tbachman_ is now known as tbachman | 23:04 | |
opendevreview | Miguel Lavalle proposed openstack/neutron master: Test commit https://review.opendev.org/c/openstack/neutron/+/821334 | 23:23 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!