| opendevreview | Merged openstack/neutron-fwaas-dashboard master: Replace deprecated ugettext_lazy and ungettext_lazy https://review.opendev.org/c/openstack/neutron-fwaas-dashboard/+/849259 | 00:02 |
|---|---|---|
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ovs: improve log message when ofport is not configured https://review.opendev.org/c/openstack/neutron/+/850147 | 06:23 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ovs: add fdb_entries details to the logs https://review.opendev.org/c/openstack/neutron/+/850226 | 06:23 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ovs: make vlanmanager to handle more vlan mapping per network https://review.opendev.org/c/openstack/neutron/+/839479 | 06:23 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: do not merge https://review.opendev.org/c/openstack/neutron/+/850582 | 06:23 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ovs: use a local vlan per network/segmentation https://review.opendev.org/c/openstack/neutron/+/840415 | 06:23 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: dhcp/rpc: retrieve network details with segments https://review.opendev.org/c/openstack/neutron/+/840416 | 06:23 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: utils: add option to generate dhcp device id per segmentation https://review.opendev.org/c/openstack/neutron/+/840417 | 06:23 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ml2/mech: bind port on segment that subnet belongs https://review.opendev.org/c/openstack/neutron/+/840418 | 06:23 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: dhcp: make device manager to clean only not used devices https://review.opendev.org/c/openstack/neutron/+/840419 | 06:23 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: dhcp: add/use cleanup stale devices API https://review.opendev.org/c/openstack/neutron/+/840420 | 06:23 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: dhcp: support multiple segmentations per network https://review.opendev.org/c/openstack/neutron/+/840421 | 06:23 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: db: add segment_index to the unique constraint https://review.opendev.org/c/openstack/neutron/+/841459 | 06:23 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: segment: enable multisegments support for host https://review.opendev.org/c/openstack/neutron/+/841838 | 06:23 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ovs: fix potential value error on undefined variable https://review.opendev.org/c/openstack/neutron/+/850609 | 06:23 |
| sahid | Morning all, I'm completly suck trying to fix CI job failing for https://review.opendev.org/c/openstack/neutron/+/839479/21 - there are one test failing https://zuul.opendev.org/t/openstack/build/5d9d1ddb53164a26b5adb7e23b82c5f3 | 06:26 |
| sahid | any chance to get some help? :-) | 06:27 |
| ralonsoh | let me check | 06:30 |
| lajoskatona | sahid: is this job and the test failing always? | 06:31 |
| ralonsoh | and this is the 4th patch in the series | 06:32 |
| sahid | yes always the same | 06:32 |
| ralonsoh | the error could be in the previous patches | 06:32 |
| sahid | the other one are just log messages that I have added to help me debug | 06:32 |
| ralonsoh | sahid, in the last execution, the test failing is not the same | 06:34 |
| sahid | could you point me that? | 06:35 |
| ralonsoh | https://storage.bhs.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_66b/839479/20/check/neutron-ovs-grenade-dvr-multinode/66bfbf7/testr_results.html | 06:35 |
| sahid | oh yes, i was focus on the exception raised | 06:37 |
| sahid | what is your feeling? | 06:37 |
| ralonsoh | that the problem is not related to the patch | 06:38 |
| ralonsoh | but a problem with rabbitmq and the RPC | 06:38 |
| ralonsoh | but I would need to check it | 06:38 |
| sahid | it seems the previous are alwyas passing it's why i considered that patch the problem | 06:39 |
| sahid | ralonsoh: is it possible that you check that or at least give me pointers so i can have a look? | 06:40 |
| ralonsoh | just checking the compute OVS agent | 06:40 |
| ralonsoh | and the rabbitmq service | 06:40 |
| ralonsoh | if there is something wrong there | 06:40 |
| ralonsoh | but I would need to check the patch first | 06:41 |
| sahid | there are some sort of issue related to rpc timeout on ovs agent, but i also noticed this error on other ci job passing, so i did not have considered that as an issue | 06:42 |
| sahid | and for rabbit i have never checked, i will have a look thank you! | 06:42 |
| sahid | s/i will have/looking | 06:43 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ovs: make vlanmanager to handle more vlan mapping per network https://review.opendev.org/c/openstack/neutron/+/839479 | 07:14 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ovs: use a local vlan per network/segmentation https://review.opendev.org/c/openstack/neutron/+/840415 | 07:14 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: dhcp/rpc: retrieve network details with segments https://review.opendev.org/c/openstack/neutron/+/840416 | 07:14 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: utils: add option to generate dhcp device id per segmentation https://review.opendev.org/c/openstack/neutron/+/840417 | 07:14 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ml2/mech: bind port on segment that subnet belongs https://review.opendev.org/c/openstack/neutron/+/840418 | 07:14 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: dhcp: make device manager to clean only not used devices https://review.opendev.org/c/openstack/neutron/+/840419 | 07:14 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: dhcp: add/use cleanup stale devices API https://review.opendev.org/c/openstack/neutron/+/840420 | 07:14 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: dhcp: support multiple segmentations per network https://review.opendev.org/c/openstack/neutron/+/840421 | 07:14 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: db: add segment_index to the unique constraint https://review.opendev.org/c/openstack/neutron/+/841459 | 07:14 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: segment: enable multisegments support for host https://review.opendev.org/c/openstack/neutron/+/841838 | 07:14 |
| lajoskatona | ralonsoh, slaweq, bcafarel: Hi, https://bugs.launchpad.net/neutron/+bug/1982206 | 07:54 |
| ralonsoh | let me check | 07:54 |
| lajoskatona | ralonsoh, slaweq, bcafarel: With OS_TEST_TIMEOUT:180 , (see https://review.opendev.org/c/openstack/neutron/+/850544 and https://review.opendev.org/c/openstack/neutron/+/850391 ) | 07:54 |
| lajoskatona | this code part seems to be failing: https://opendev.org/openstack/neutron/src/branch/stable/victoria/neutron/agent/linux/iptables_firewall.py#L104 | 07:55 |
| lajoskatona | Victoria example: https://44478d02c804c0b9df5e-a6bfe3158a83ff259d3753105258a398.ssl.cf5.rackcdn.com/850544/1/check/openstack-tox-py38/9aa8800/job-output.txt | 07:55 |
| lajoskatona | Ussuri: https://7f8143495509f2da98f9-0ee3eeb68aa256c74f1e35f60c262d61.ssl.cf2.rackcdn.com/850391/2/check/openstack-tox-py36/519b9f5/testr_results.html | 07:55 |
| lajoskatona | to be precise: not failing but taking long time | 07:55 |
| lajoskatona | there was some ubuntu pkg changes that started (see the last comment in lp) so not sure what should be next step | 07:57 |
| bcafarel | hmm I wonder if we can blacklist some system updates (we had a similar issue some time ago, though it was fixed with another update) | 08:00 |
| bcafarel | if we can track down which one is causing it and hand over a not too large reproducer in ubuntu launchpad | 08:01 |
| lajoskatona | bcafarel: this is the pkg diff between the working and red runs: https://paste.opendev.org/show/b8JwAKL4wMqdxCGOZpp8/ (linked to the lp bug also) | 08:03 |
| opendevreview | liuyulong proposed openstack/neutron-specs master: Spec for Router Advertisement on Openvswitch Agent https://review.opendev.org/c/openstack/neutron-specs/+/832660 | 08:21 |
| ralonsoh | bcafarel, instead of launching this command, I'll try using "lsmod | grep br_netfilter" to check if br_netfilter is loaded | 08:40 |
| ralonsoh | lajoskatona, ^^ | 08:40 |
| ralonsoh | I'll push a DNM patch, just for testing | 08:40 |
| ralonsoh | the error in W I don't think is related | 08:40 |
| ralonsoh | this is affecting bionic (4.15 kernel, not 5.x) | 08:41 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron stable/victoria: [DNM] Check if "br_netfilter" is loaded reading "lsmod" output https://review.opendev.org/c/openstack/neutron/+/850612 | 08:53 |
| ralonsoh | bcafarel, lajoskatona ^^ | 08:53 |
| slaweq | ralonsoh but this is UT, it shouldn't really interact with OS | 08:54 |
| slaweq | maybe we are missing some mock there | 08:54 |
| ralonsoh | you are right on this | 08:54 |
| ralonsoh | ok | 08:54 |
| ralonsoh | if that patch doesn't work, I'll mock it | 08:54 |
| slaweq | ++ | 08:54 |
| ralonsoh | in any case, that should not fail, there is something in this kernel | 08:55 |
| slaweq | yes, that's true | 08:55 |
| slaweq | maybe if we can reproduce same issue without Neutron, we can report bug for the Ubuntu kernel then | 08:55 |
| ralonsoh | I've tried manually with a bionic cloud image | 08:55 |
| ralonsoh | but I can't reproduce this timeout | 08:56 |
| lajoskatona | mocking sounds good | 08:57 |
| opendevreview | Lajos Katona proposed openstack/neutron stable/victoria: DNM: check victoria https://review.opendev.org/c/openstack/neutron/+/850544 | 09:27 |
| ykarel | ralonsoh, you see the same kernel in bionic vm you used? | 09:37 |
| ykarel | 4.15.0-189-generic #200 | 09:37 |
| ralonsoh | yes | 09:38 |
| ykarel | okk | 09:38 |
| ralonsoh | ubuntu@server1:~$ uname -a | 09:38 |
| ralonsoh | Linux server1 4.15.0-189-generic #200-Ubuntu SMP Wed Jun 22 19:53:37 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux | 09:38 |
| ykarel | okk may be we can try to hold CI node and check | 09:39 |
| opendevreview | Lajos Katona proposed openstack/neutron stable/ussuri: DNM: check ussuri https://review.opendev.org/c/openstack/neutron/+/850391 | 09:47 |
| lajoskatona | ykarel, ralnsoh: by putting a self.fail() to here: https://opendev.org/openstack/neutron/src/branch/stable/ussuri/neutron/agent/linux/iptables_firewall.py#L102 something simlar can be achieved like we see on zuul jobs :P | 09:48 |
| sahid | ralonsoh, lajoskatona I think I have found the issue: https://review.opendev.org/c/openstack/neutron/+/839479/21..22/neutron/plugins/ml2/drivers/openvswitch/agent/ovs_neutron_agent.py#1955 | 10:06 |
| opendevreview | Slawek Kaplonski proposed openstack/neutron master: Port update will trigger less notifications to the DHCP agents https://review.opendev.org/c/openstack/neutron/+/850529 | 10:12 |
| sahid | the break was stopping the first loop, but i hqve introduced q new one since multisegement, which was not stopped, the iteration was continuing to the next one, which would have make the process the clean the port where it should have not | 10:13 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ovs: fix potential value error on undefined variable https://review.opendev.org/c/openstack/neutron/+/850609 | 10:16 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ovs: add fdb_entries details to the logs https://review.opendev.org/c/openstack/neutron/+/850226 | 10:16 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ovs: make vlanmanager to handle more vlan mapping per network https://review.opendev.org/c/openstack/neutron/+/839479 | 10:16 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ovs: use a local vlan per network/segmentation https://review.opendev.org/c/openstack/neutron/+/840415 | 10:16 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: dhcp/rpc: retrieve network details with segments https://review.opendev.org/c/openstack/neutron/+/840416 | 10:16 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: utils: add option to generate dhcp device id per segmentation https://review.opendev.org/c/openstack/neutron/+/840417 | 10:16 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ml2/mech: bind port on segment that subnet belongs https://review.opendev.org/c/openstack/neutron/+/840418 | 10:16 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: dhcp: make device manager to clean only not used devices https://review.opendev.org/c/openstack/neutron/+/840419 | 10:16 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: dhcp: add/use cleanup stale devices API https://review.opendev.org/c/openstack/neutron/+/840420 | 10:16 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: dhcp: support multiple segmentations per network https://review.opendev.org/c/openstack/neutron/+/840421 | 10:17 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: db: add segment_index to the unique constraint https://review.opendev.org/c/openstack/neutron/+/841459 | 10:17 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: segment: enable multisegments support for host https://review.opendev.org/c/openstack/neutron/+/841838 | 10:17 |
| opendevreview | Merged openstack/neutron master: ovs: improve log message when ofport is not configured https://review.opendev.org/c/openstack/neutron/+/850147 | 10:23 |
| lajoskatona | sahid: cool, checking | 10:31 |
| opendevreview | Merged openstack/networking-bagpipe master: Py3: Import urllib properly https://review.opendev.org/c/openstack/networking-bagpipe/+/845271 | 10:42 |
| opendevreview | Rodolfo Alonso proposed openstack/networking-bagpipe master: py310: Add required-projects list to py310 job https://review.opendev.org/c/openstack/networking-bagpipe/+/841917 | 10:44 |
| opendevreview | Rodolfo Alonso proposed openstack/networking-bagpipe master: Remove "distutils" library https://review.opendev.org/c/openstack/networking-bagpipe/+/842369 | 10:45 |
| opendevreview | Rodolfo Alonso proposed openstack/os-ken master: [DNM][Testing] Remove "nose" library https://review.opendev.org/c/openstack/os-ken/+/850631 | 10:48 |
| lajoskatona | ralonsoh, ykarel, slaweq: for mocking sysctl call see these: https://review.opendev.org/c/openstack/neutron/+/850391 & https://review.opendev.org/c/openstack/neutron/+/850544 | 11:45 |
| lajoskatona | ralonsoh, ykarel, slaweq: perhaps would be better to mock it on master and backport to avoid such issues on later branches also, what do you think? | 11:45 |
| *** amoralej is now known as amoralej|lunch | 12:20 | |
| *** haleyb_ is now known as haleyb | 12:52 | |
| *** dasm|off is now known as dasm|ruck | 13:05 | |
| sahid | it finally passed that CI job :-) https://review.opendev.org/c/openstack/neutron/+/839479/23 | 13:28 |
| *** amoralej|lunch is now known as amoralej | 13:28 | |
| slaweq | lajoskatona: yes, I think we should fix it in master too, in UT we should have mocked all such things | 13:43 |
| ykarel | lajoskatona yes +1 to fix master and backports | 14:00 |
| ykarel | results looks good on the patches | 14:02 |
| mnasiadka | good afternoon | 14:51 |
| lajoskatona | mnasiadka: good afternoon | 14:53 |
| mnasiadka | OVN and E/W routing (something like https://bugzilla.redhat.com/show_bug.cgi?id=1826364) - is anybody working on overcoming that? I have an environment with a lot of baremetal ports and E/W routing would be nice | 14:53 |
| mnasiadka | And currently active chassis from ha_chassis_group AND active gateway in a router that has external gateway need to be on the same chassis for the routing to work | 14:56 |
| lajoskatona | mnasiadka: I don't know about such an upstream RFE | 14:57 |
| mnasiadka | lajoskatona: thanks | 14:57 |
| mnasiadka | If anybody has an idea how to ,,fix'' that on Neutron side, I can work on a patch. | 14:58 |
| mnasiadka | I'll try to bug ralonsoh_ later or tomorrow | 14:58 |
| lajoskatona | mnasiadka: open an RFE, and I can add it to tomorrow's drivers meeting agenda for quick discussion about it | 15:03 |
| ralonsoh_ | mnasiadka, this RFE is still in our plate but I don't know if this is being assigned | 15:04 |
| ralonsoh_ | so far the assignee of this BZ is jlibosva | 15:04 |
| * jlibosva looks | 15:04 | |
| ralonsoh_ | but there is not target release yet | 15:04 |
| ralonsoh_ | BZ https://bugzilla.redhat.com/show_bug.cgi?id=1826364 | 15:04 |
| ralonsoh_ | I don't now if https://review.opendev.org/c/openstack/neutron/+/826912 affects to external (sriov) ports | 15:05 |
| mnasiadka | if there's an RFE in launchpad about this - and somebody is willing to give me some guidance if there's any magic in OVN that allows to tie one to the other (I doubt Neutron chasing this relationship is a good idea) | 15:06 |
| mnasiadka | ralonsoh_: I'm using that (I authored the original patch - https://review.opendev.org/c/openstack/neutron/+/782250) | 15:06 |
| ralonsoh_ | mnasiadka, but this is for FIP in VLAn | 15:07 |
| ralonsoh_ | not external ports, right? | 15:07 |
| mnasiadka | Yes, right. | 15:07 |
| ralonsoh_ | so you can copy/paste the BZ description in Launchpad and start working on this | 15:07 |
| ralonsoh_ | you are more than welcome | 15:07 |
| ralonsoh_ | and I'll review those patches for sure | 15:07 |
| ralonsoh_ | do we need core OVN support? | 15:08 |
| mnasiadka | Well, probably we do - as mentioned Neutron should not be chasing that ,,relationship'' and updating it to get data flowing. | 15:08 |
| mnasiadka | I'll consult with OVN core folks first. | 15:09 |
| ralonsoh_ | mnasiadka, please, feel free to create a LP bug | 15:09 |
| jlibosva | ralonsoh_: mnasiadka this is not about SR-IOV | 15:09 |
| jlibosva | mnasiadka: do you use ovn-chassis-mac-mappings on the nodes? | 15:10 |
| mnasiadka | jlibosva: yes I do | 15:10 |
| mnasiadka | only on computes (not on the controllers==network nodes) | 15:10 |
| jlibosva | mnasiadka: and the external ports are on the network nodes? | 15:12 |
| mnasiadka | jlibosva: yes | 15:12 |
| ralonsoh_ | (how is that possible? shouldn't sriov ports be on the computes?) | 15:13 |
| jlibosva | mnasiadka: then I think the mac mapping should be there because that's where the routing happens | 15:14 |
| mnasiadka | ralonsoh_: problem is I'm not talking about SRIOV ports - in my case those are Ironic baremetal ports. | 15:14 |
| mnasiadka | jlibosva: let me try | 15:14 |
| jlibosva | mnasiadka: I think it's the same, there is an external port bound somewhere that is used to talk to the OVN overlay networks | 15:15 |
| jlibosva | basically in theory both baremetal and SR-IOV ports are physical ports on the provider network | 15:15 |
| mnasiadka | nope, setting ovn-chassis-mac-mappings does not help | 15:20 |
| mnasiadka | works from an instance (OVS normal port), doesn't work from the external port (when active chassis for external ports from ha_chassis_group is not the same as active gateway chassis for that particular router) | 15:21 |
| mnasiadka | I'll consult the OVN folks and raise a bug to work on it. | 15:22 |
| mnasiadka | Thanks for your time ralonsoh_ and jlibosva :) | 15:22 |
| ralonsoh_ | yw | 15:23 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: dhcp: support multiple segmentations per network https://review.opendev.org/c/openstack/neutron/+/840421 | 15:26 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: db: add segment_index to the unique constraint https://review.opendev.org/c/openstack/neutron/+/841459 | 15:26 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: segment: enable multisegments support for host https://review.opendev.org/c/openstack/neutron/+/841838 | 15:26 |
| jlibosva | mnasiadka: thank you! | 15:28 |
| opendevreview | Lajos Katona proposed openstack/neutron master: Test: mock out _check_netfilter_for_bridges in unit tests https://review.opendev.org/c/openstack/neutron/+/850670 | 15:32 |
| opendevreview | Merged openstack/ovn-octavia-provider master: Ensure members without subnet belong to VIP subnet or fail https://review.opendev.org/c/openstack/ovn-octavia-provider/+/849691 | 15:44 |
| *** amoralej is now known as amoralej|off | 15:57 | |
| opendevreview | Sharon Koech proposed openstack/neutron-lib master: Network Cascade Deletion Extension https://review.opendev.org/c/openstack/neutron-lib/+/849046 | 16:59 |
| opendevreview | Rodolfo Alonso proposed openstack/os-ken master: [DNM][Testing] Remove "nose" library https://review.opendev.org/c/openstack/os-ken/+/850631 | 17:11 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: dhcp: support multiple segmentations per network https://review.opendev.org/c/openstack/neutron/+/840421 | 19:34 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: db: add segment_index to the unique constraint https://review.opendev.org/c/openstack/neutron/+/841459 | 19:34 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: segment: enable multisegments support for host https://review.opendev.org/c/openstack/neutron/+/841838 | 19:34 |
| *** dasm|ruck is now known as dasm|off | 21:04 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!