| opendevreview | Victor Morales proposed openstack/neutron master: Use more pythonic approach https://review.opendev.org/c/openstack/neutron/+/806065 | 01:04 |
|---|---|---|
| *** iurygregory_ is now known as iurygregory | 06:28 | |
| opendevreview | Merged openstack/neutron stable/wallaby: Replace assertItemsEqual with assertCountEqual https://review.opendev.org/c/openstack/neutron/+/805791 | 06:55 |
| *** rpittau|afk is now known as rpittau | 07:24 | |
| opendevreview | Lajos Katona proposed openstack/networking-odl master: DNM: Check master status https://review.opendev.org/c/openstack/networking-odl/+/806095 | 07:27 |
| opendevreview | Lajos Katona proposed openstack/networking-bgpvpn master: DNM: check master https://review.opendev.org/c/openstack/networking-bgpvpn/+/806096 | 07:30 |
| opendevreview | Lajos Katona proposed openstack/networking-bagpipe master: DNM: check master https://review.opendev.org/c/openstack/networking-bagpipe/+/806097 | 07:30 |
| opendevreview | yatin proposed openstack/networking-bgpvpn master: Use assertCountEqual instead of assertItemsEqual https://review.opendev.org/c/openstack/networking-bgpvpn/+/806101 | 08:17 |
| opendevreview | Lajos Katona proposed openstack/neutron-lib master: DNM: test muster https://review.opendev.org/c/openstack/neutron-lib/+/806102 | 08:23 |
| opendevreview | Bernard Cafarelli proposed openstack/neutron stable/train: [DNM] Check train CI status https://review.opendev.org/c/openstack/neutron/+/806108 | 08:34 |
| bcafarel | ^ ralonsoh more fun with that new advanced image flavor https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_bcc/804720/2/check/neutron-tempest-plugin-scenario-linuxbridge-rocky/bccc908/job-output.txt | 08:35 |
| bcafarel | EM branches use job definition from master n-t-p, but older tagged checkout :) | 08:35 |
| ralonsoh | bcafarel, but the devstack plugin is the same | 08:36 |
| ralonsoh | it should work too, at least with your patch | 08:36 |
| ralonsoh | right? | 08:37 |
| bcafarel | we checkout neutron-tempest-plugin at older tag so it does not have the create_flavor_for_advance_image call (looks so from the line numbers) | 08:37 |
| opendevreview | Lajos Katona proposed openstack/neutron-lib master: Cap oslo.messaging to be lower than 12.9.0 https://review.opendev.org/c/openstack/neutron-lib/+/806113 | 08:37 |
| bcafarel | but zuul jobs we get from master and it tries ADVANCED_INSTANCE_TYPE: ntp_image_384M | 08:38 |
| ralonsoh | bcafarel, so is using the new image but not creating it | 08:38 |
| ralonsoh | I don't understand | 08:38 |
| ralonsoh | the zuul file was modified in the same patch where the method was implemented | 08:38 |
| bcafarel | ralonsoh: yes so it works fine for current branches, but EM we override checkout https://opendev.org/openstack/neutron-tempest-plugin/src/branch/master/zuul.d/train_jobs.yaml#L9 | 08:39 |
| bcafarel | so the code in devstack plugin is not there for these branches | 08:39 |
| ralonsoh | ahhhhhh | 08:40 |
| ralonsoh | bcafarel, ok ok | 08:40 |
| bcafarel | (took me some head scractching to figure out this morning :) ) | 08:40 |
| ralonsoh | so, what I'm going to do is to override the default value in EM branches | 08:40 |
| ralonsoh | and use the default one | 08:40 |
| ralonsoh | EM branches == up to queens?? | 08:40 |
| ralonsoh | actually this problem was found in newer versions | 08:41 |
| ralonsoh | I think we can use the default image up to Train | 08:41 |
| bcafarel | ah good so if we can still use default flavor in EM it sounds easier then | 08:41 |
| ralonsoh | I'll push a patch right now | 08:41 |
| bcafarel | and up to Train now (time flies...) | 08:41 |
| bcafarel | ralonsoh: thanks :) | 08:42 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron-tempest-plugin master: Use default "ds512M" advance image in older stable branches https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/806119 | 08:59 |
| ralonsoh | bcafarel, ^^ | 08:59 |
| lajoskatona | Hi, neutron-lib unittests are broken, due to the release of oslo.messaging | 09:37 |
| lajoskatona | https://dcdfbd7cdeab11842c91-86f17d3268054037bb0d094a3c69f66d.ssl.cf5.rackcdn.com/806102/1/check/openstack-tox-py38/dbc7db8/testr_results.html | 09:38 |
| ralonsoh | lajoskatona, will you work on this? I can do it if you don't have time | 09:39 |
| ralonsoh | do you have a bug? | 09:39 |
| lajoskatona | ralonsoh: I tried, but lost in mock.... | 09:39 |
| ralonsoh | lajoskatona, I'll try too | 09:39 |
| ralonsoh | first I'll open a bug | 09:40 |
| lajoskatona | I pushed a patch to cap it: https://review.opendev.org/c/openstack/neutron-lib/+/806113 | 09:40 |
| lajoskatona | but hit some dependency hell issue with it | 09:40 |
| lajoskatona | good idea | 09:40 |
| lajoskatona | I hoped that it will easy few minutes thing.... | 09:41 |
| ralonsoh | https://bugs.launchpad.net/neutron/+bug/1941725 | 09:42 |
| ralonsoh | lajoskatona, https://github.com/openstack/oslo.messaging/blob/master/oslo_messaging/_metrics/client.py#L72 | 09:49 |
| ralonsoh | we mock this variable instead of providing a value | 09:49 |
| ralonsoh | I'll push a patch | 09:49 |
| lajoskatona | the issue is that even if metrics_enabled = False (as it is the default) the code goes to the new _metrics.client, which is strange.... | 09:56 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron-lib master: Provide int value for oslo.messaging "metrics_buffer_size" https://review.opendev.org/c/openstack/neutron-lib/+/806129 | 09:59 |
| ralonsoh | lajoskatona, ^ | 09:59 |
| ralonsoh | ok, still not working | 10:01 |
| lajoskatona | ralonsoh: for me if I put it to TimeoutTestCase setup it fixes a lot | 10:04 |
| ralonsoh | I still need to set those overrides in the other class | 10:04 |
| lajoskatona | I think I left out oslo_messaging_metrics .... | 10:04 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron-lib master: Provide int value for oslo.messaging "metrics_buffer_size" https://review.opendev.org/c/openstack/neutron-lib/+/806129 | 10:05 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron-lib master: Provide int value for oslo.messaging "metrics_buffer_size" https://review.opendev.org/c/openstack/neutron-lib/+/806129 | 10:09 |
| lajoskatona | ralonsoh: for me metrics_thread_stop_timeout = 2 worked :-) | 10:13 |
| ralonsoh | and metrics_thread_stop_timeout=0 better | 10:13 |
| ralonsoh | because that will prevent from trying to send any metric | 10:13 |
| ralonsoh | and we'll not see any warning message in the UTs logs | 10:13 |
| ralonsoh | see last patch | 10:14 |
| lajoskatona | ok , I see now, but with rpc.TRANSPORT.conf.oslo_messaging_metrics.metrics_enabled = False it works as well | 10:16 |
| opendevreview | Lajos Katona proposed openstack/neutron-lib master: Fix hostroute validation with BFD https://review.opendev.org/c/openstack/neutron-lib/+/805834 | 10:21 |
| ralonsoh | hi folks, can I draw your attention to https://review.opendev.org/c/openstack/neutron/+/804218? | 10:40 |
| ralonsoh | I think the patch and the bug are easy to review | 10:41 |
| ralonsoh | and I have a customer hitting it | 10:41 |
| *** ksambor__ is now known as ksambor | 10:55 | |
| opendevreview | Rodolfo Alonso proposed openstack/neutron-lib master: Disable "oslo.messaging" metrics https://review.opendev.org/c/openstack/neutron-lib/+/806129 | 10:57 |
| opendevreview | Merged openstack/neutron stable/wallaby: Use elevated context to get default SG from database https://review.opendev.org/c/openstack/neutron/+/803439 | 12:11 |
| opendevreview | Balazs Gibizer proposed openstack/neutron master: Enable QoS min bw tempest in neutron-ovs-tempest-multinode-full https://review.opendev.org/c/openstack/neutron/+/806170 | 12:45 |
| opendevreview | Mamatisa Nurmatov proposed openstack/neutron master: use payload for OVSDB_RESOURCE https://review.opendev.org/c/openstack/neutron/+/805397 | 12:53 |
| opendevreview | Lajos Katona proposed openstack/neutron-dynamic-routing master: DNM: test master https://review.opendev.org/c/openstack/neutron-dynamic-routing/+/806177 | 13:15 |
| opendevreview | Merged openstack/neutron-lib master: Disable "oslo.messaging" metrics https://review.opendev.org/c/openstack/neutron-lib/+/806129 | 14:15 |
| spatel | Folk, I have question related SRIOV compute node with OVN Reading: https://docs.openstack.org/neutron/latest/admin/ovn/sriov.html | 14:26 |
| spatel | what is this going to be in my setup? ovn-bridge-mappings="public:br-ex" | 14:26 |
| spatel | i have vlan:br-provider at present for normal compute nodes but not sure what i should set this for SR-IOV compute | 14:27 |
| opendevreview | Oleg Bondarev proposed openstack/neutron-lib master: Add Local IP API def https://review.opendev.org/c/openstack/neutron-lib/+/803051 | 14:35 |
| ralonsoh | lajoskatona, https://review.opendev.org/c/openstack/neutron-lib/+/805834 | 14:40 |
| ralonsoh | can you rebase it? | 14:40 |
| lajoskatona | ralonsoh: thanks, I do | 14:41 |
| opendevreview | Lajos Katona proposed openstack/neutron master: Doc: follow-up for recent job renames https://review.opendev.org/c/openstack/neutron/+/806200 | 14:41 |
| ralonsoh | spatel, did you check https://docs.openstack.org/neutron/latest/admin/ovn/sriov.html? | 14:42 |
| spatel | Yes i am reading that ralonsoh | 14:42 |
| ralonsoh | the SRIOV config is the same as it was before | 14:42 |
| ralonsoh | the SRIOV config is the same as it was before | 14:43 |
| ralonsoh | then you need to enable a chassis as GW | 14:43 |
| spatel | I am running VLAN base provider and my router is my physical datacenter gateway | 14:43 |
| opendevreview | Lajos Katona proposed openstack/neutron-lib master: Fix hostroute validation with BFD https://review.opendev.org/c/openstack/neutron-lib/+/805834 | 14:43 |
| ralonsoh | ah, and you remember to deploy metadata on GW nodes (local metadata won't work with sriov ports) | 14:43 |
| spatel | ralonsoh ah! | 14:44 |
| spatel | currently i have normal compute which is default running metadata on local compute. | 14:44 |
| ralonsoh | that will work woth OVN ports | 14:44 |
| ralonsoh | but SRIOV port will receive the metadata from the GW metadata service | 14:45 |
| spatel | I have configure DVR so every compute node is my GW node | 14:45 |
| ralonsoh | spatel, BTW, check "known limitation", point 2 | 14:45 |
| spatel | ralonsoh interesting so in my case it won't work right, i am planning to use vlan provider | 14:48 |
| ralonsoh | check the bug, I didn't track it | 14:48 |
| spatel | This is still open bug | 14:49 |
| lajoskatona | slaweq: shall we include https://review.opendev.org/c/openstack/neutron-lib/+/805834 in neutron-lib ? | 14:49 |
| spatel | ralonsoh regardless that bug if i want to run OVN with SRIOV then i should need to set /etc/neutron/plugins/ml2/sriov_nic_agent.ini file and daemon as we used to do earlier right? | 14:50 |
| ralonsoh | spatel, yes, SRIOV agent is needed in the compute node | 14:51 |
| ralonsoh | and the config file (same as in Nova) | 14:51 |
| ralonsoh | I mean, Nova config is also needed for SRIOV | 14:51 |
| spatel | Let me give it a try and see what issue i can face. according that i will update that BUG report | 14:52 |
| spatel | Assuming OVN will automatically create external port on my chassis | 14:52 |
| opendevreview | mitya-eremeev-2 proposed openstack/neutron master: Added remote securiy group proccessing step. https://review.opendev.org/c/openstack/neutron/+/806203 | 14:52 |
| ralonsoh | lajoskatona, I think we should, BFD validation is already released | 14:53 |
| ralonsoh | and we need this fix | 14:53 |
| spatel | ralonsoh is this looks correct mechanism_drivers = ovn,sriovnicswitch | 14:53 |
| ralonsoh | yes | 14:54 |
| spatel | thanks! | 14:54 |
| lajoskatona | ralonsoh: could you wf+1 it, please ;) | 14:55 |
| ralonsoh | lajoskatona, sure | 14:55 |
| lajoskatona | ralonsoh: thanks | 14:58 |
| spatel | ralonsoh assuming i also need openvswitch_agent.ini file on compute node right? | 15:21 |
| ralonsoh | spatel, no you don't, no one should be using OVS agent | 15:27 |
| spatel | oh! okay | 15:27 |
| spatel | ralonsoh just curious because i have seen many example people using it, is this the case with OVN to not use it? | 15:28 |
| ralonsoh | there are no OVS agents, you don't need it | 15:29 |
| ralonsoh | all ml2 config can be in ml2_plugin.ini | 15:29 |
| ralonsoh | it is wrong to name a file like this is you don't have OVS agents | 15:29 |
| spatel | hmm | 15:30 |
| spatel | ralonsoh so i have setup all sriov config and created sriov port and when try to attach to vm i got this error on neutron-server logs - https://paste.opendev.org/show/808354/ | 15:57 |
| spatel | can you see what went wrong here? or did i missing something | 15:57 |
| spatel | I can see HA_Chassis_Group then why neutron not able to find that? | 15:58 |
| ralonsoh | you didn't set one of the controller as GW | 15:58 |
| ralonsoh | --> https://docs.openstack.org/neutron/latest/admin/ovn/sriov.html#ovn-database-information | 15:58 |
| ralonsoh | for SRIOV, you need to declare a chassis GW | 15:59 |
| spatel | hmm i do have this - ovn-cms-options=enable-chassis-as-gw | 16:00 |
| spatel | did i mis-understand anything here? | 16:02 |
| *** rpittau is now known as rpittau|afk | 16:02 | |
| ralonsoh | and do you have a HA_Chassis_Group? | 16:02 |
| ralonsoh | named "default_ha_chassis_group" | 16:02 |
| spatel | look at this output - https://paste.opendev.org/show/808354/ | 16:02 |
| spatel | Yes i do have HA_Chassis_Group | 16:02 |
| spatel | https://paste.opendev.org/show/808356/ | 16:03 |
| ralonsoh | execute | 16:04 |
| ralonsoh | ovn-nbctl list HA_Chassis 8f072dda-070a-4ff3-805c-bb0f40b99348 c4c08be8-5cdc-4cd7-99ed-a22bb787ab55 | 16:04 |
| spatel | here you go - https://paste.opendev.org/show/808357/ | 16:05 |
| ralonsoh | spatel, ok now I'm checking https://paste.opendev.org/show/808354/ again | 16:06 |
| ralonsoh | none of those chassis have "ovn-cms-options=enable-chassis-as-gw" in external_ids | 16:06 |
| ralonsoh | ah yes | 16:06 |
| ralonsoh | at the end | 16:06 |
| spatel | are you talking about this flag = external_ids : {} | 16:07 |
| ralonsoh | L4 and L13 | 16:07 |
| spatel | i do have that for both compute nodes "ovn-cms-options=enable-chassis-as-gw" | 16:08 |
| ralonsoh | and command | 16:09 |
| ralonsoh | ovn-nbctl find Logical_Switch_Port type=external | 16:09 |
| ralonsoh | let me check | 16:09 |
| opendevreview | Merged openstack/neutron master: Use more pythonic approach https://review.opendev.org/c/openstack/neutron/+/806065 | 16:09 |
| opendevreview | Bernard Cafarelli proposed openstack/neutron-vpnaas stable/train: Pin isort to 4.3.21 https://review.opendev.org/c/openstack/neutron-vpnaas/+/805969 | 16:11 |
| spatel | ralonsoh sure, standby | 16:13 |
| ralonsoh | spatel, can you report a launchpad bug with this information? | 16:18 |
| spatel | ralonsoh sure, i can do that | 16:18 |
| ralonsoh | I'll ping lucasgomes | 16:18 |
| spatel | now i have removed all my normal compute nodes and just kept sriov compute node enabled and try to spin up VM and getting this error on neutron-server - https://paste.opendev.org/show/808358/ | 16:19 |
| ralonsoh | spatel, do you have https://review.opendev.org/c/openstack/neutron/+/795260 | 16:20 |
| ralonsoh | check https://bugs.launchpad.net/neutron/+bug/1931220 | 16:20 |
| spatel | checking | 16:20 |
| spatel | ralonsoh i don't have that patch | 16:23 |
| ralonsoh | apply it | 16:23 |
| spatel | ok | 16:23 |
| spatel | i am comparing that file line by line and look like i have lots of things are different in that file.. i may running little older version of neutron | 16:25 |
| spatel | i can move my branch to master if that help here | 16:25 |
| ralonsoh | it should | 16:28 |
| spatel | ralonsoh let me upgrade my neutron to master branch, i don't think direct applying patch will work | 16:28 |
| spatel | i will get back to you after upgrade.. | 16:28 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [WIP] Write FW OF rules belonging to a port in a single operation https://review.opendev.org/c/openstack/neutron/+/806246 | 16:55 |
| spatel | ralonsoh i am little confused here, i am looking at master branch and that patch isn't part of latest master - https://opendev.org/openstack/neutron/src/branch/master/neutron/plugins/ml2/ovo_rpc.py | 17:07 |
| ralonsoh | spatel, it is but another patch updated it | 17:08 |
| spatel | so what should i use that patch specific SHA or latest master | 17:09 |
| ralonsoh | no, you should use master for testing | 17:10 |
| spatel | let me use master then | 17:11 |
| spatel | ralonsoh same error :( - https://paste.opendev.org/show/808360/ | 17:21 |
| ralonsoh | attach to the bug all logs, not only this line | 17:22 |
| spatel | how do ovn know i have eno49 is SRIOV and map to that VF ? | 17:22 |
| ralonsoh | I can't figure out what is happening there | 17:22 |
| ralonsoh | reading the neutron port | 17:23 |
| spatel | anyway let me open BUG report and see | 17:24 |
| spatel | ralonsoh here its - https://bugs.launchpad.net/neutron/+bug/1941784 | 17:44 |
| ralonsoh | there should be more information in the neutron server logs | 17:45 |
| ralonsoh | not only this line | 17:45 |
| ralonsoh | there should be an exception | 17:45 |
| ralonsoh | or something | 17:45 |
| spatel | hmm let me enable debug and put more info | 17:48 |
| spatel | one more thing i am seeing i don't have any neutron-sriov-nic-agent.service service running on compute node | 17:48 |
| ralonsoh | so no one is binding this port | 17:48 |
| spatel | i do have sriov_nic_agent.ini file | 17:48 |
| ralonsoh | if you dont' have a sriov agent, no one will bind this port | 17:49 |
| ralonsoh | this is not an OVN problem right now | 17:49 |
| spatel | checking my systemd service | 17:49 |
| spatel | re-running my playbook, something went wrong here let me fix it | 17:52 |
| ralonsoh | spatel, I'm leaving now, it's almost 8PM for me | 17:53 |
| spatel | ralonsoh thank you for your help, i will update that ticket with more information | 17:53 |
| ralonsoh | cool | 17:53 |
| spatel | will talk to you tomorrow if you around, have a good night | 17:53 |
| spatel | ralonsoh it works after fixing sriov-nic-agent service | 18:04 |
| spatel | Now everything works, i can ping outside world from my VM and it got IP address from DHCP also | 18:16 |
| spatel | going to delete that bug | 18:16 |
| opendevreview | Slawek Kaplonski proposed openstack/neutron master: [DVR] Set arp entries only for single IPs given as allowed addr pair https://review.opendev.org/c/openstack/neutron/+/800059 | 19:09 |
| *** tbachman is now known as Guest5500 | 19:14 | |
| opendevreview | Merged openstack/neutron-lib master: Fix hostroute validation with BFD https://review.opendev.org/c/openstack/neutron-lib/+/805834 | 19:34 |
| opendevreview | mitya-eremeev-2 proposed openstack/neutron master: Added remote securiy group proccessing step. https://review.opendev.org/c/openstack/neutron/+/806203 | 20:24 |
| *** tbachman is now known as Guest5504 | 20:26 | |
| *** tbachman is now known as Guest5507 | 20:50 | |
| opendevreview | mitya-eremeev-2 proposed openstack/neutron master: Added remote security group proccessing step. https://review.opendev.org/c/openstack/neutron/+/806203 | 20:54 |
| opendevreview | Lance Bragstad proposed openstack/neutron master: Fix typos in agent policy names https://review.opendev.org/c/openstack/neutron/+/806277 | 21:13 |
| *** tbachman is now known as Guest5510 | 21:18 | |
| opendevreview | Merged openstack/neutron-tempest-plugin master: Add new scenario jobs for distributed dhcp https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/803470 | 21:57 |
| *** tbachman is now known as Guest5516 | 23:03 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!