| opendevreview | Verification of a change to openstack/kayobe master failed: ntp: Remove removal of chrony container https://review.opendev.org/c/openstack/kayobe/+/827487 | 02:04 |
|---|---|---|
| opendevreview | wangxiyuan proposed openstack/kolla-ansible master: [WIP]Add openEuler Distro support https://review.opendev.org/c/openstack/kolla-ansible/+/830115 | 02:32 |
| opendevreview | wangxiyuan proposed openstack/kolla-ansible master: [WIP]Add openEuler Distro support https://review.opendev.org/c/openstack/kolla-ansible/+/830115 | 07:02 |
| opendevreview | Radosław Piliszek proposed openstack/kolla-ansible master: Add missing services to logrotate https://review.opendev.org/c/openstack/kolla-ansible/+/830433 | 07:50 |
| mnasiadka | morning | 07:59 |
| sri_ | morning | 08:01 |
| opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: Add openvswitch and prometheus to logrotate https://review.opendev.org/c/openstack/kolla-ansible/+/830433 | 08:03 |
| mnasiadka | yoctozepto: I hate useless commit messages ;-) | 08:03 |
| frickler | mnasiadka: +1 | 08:08 |
| mnasiadka | frickler: do you know if Rocky builds work now? I noticed the EPEL patch to DIB has been merged. | 08:19 |
| yoctozepto | mnasiadka: "useless" is a bit harsh but yours is better | 08:19 |
| SvenKieske | morning! is the above logrotate fix going to be backported? seems fairly simple | 08:20 |
| mnasiadka | yes, just added "should backport" flag | 08:22 |
| mnasiadka | frickler: I'll reply to myself, now it's this to fix: https://review.opendev.org/c/opendev/glean/+/830539 | 08:26 |
| wxy-xiyuan_ | Hi team, I'm facing a problem that the VM created from nova can't be ping and ssh. https://zuul.opendev.org/t/openstack/build/7987ce78edfd4ed792b53bda220eee6a It looks that DHCP maybe hits the problem. Any idea about the issue? Big thanks | 08:29 |
| mnasiadka | check openvswitch log | 08:38 |
| SvenKieske | what do you think, is it worthwhile backporting this? https://review.opendev.org/c/openstack/kolla-ansible/+/758576 I will need to backport it internally anyway, so I thought it might be cool to get it upstream as well. I still need some review btw, if you don't mind. :) | 08:47 |
| mnasiadka | First let's merge it in master :) You can propose a backport then - since it's a bugfix | 08:55 |
| SvenKieske | cool :) | 08:58 |
| *** hrw is now known as Guest369 | 09:00 | |
| *** hrww is now known as hrw | 09:00 | |
| *** Guest369 is now known as hrw | 09:00 | |
| opendevreview | Sven Kieske proposed openstack/kolla-ansible master: re-add rabbitmq config for interfaces https://review.opendev.org/c/openstack/kolla-ansible/+/758576 | 09:02 |
| *** hrw is now known as Guest371 | 09:16 | |
| *** hrww is now known as hrw | 09:16 | |
| wxy-xiyuan_ | neutron-openvswitch-agent doesn't refresh iptable rules, but it seems there is no erorr raised in the agent log. | 09:36 |
| opendevreview | Merged openstack/kolla-ansible master: Add openvswitch and prometheus to logrotate https://review.opendev.org/c/openstack/kolla-ansible/+/830433 | 10:37 |
| opendevreview | Merged openstack/kolla-ansible stable/xena: CI: Bump Ceph to Pacific https://review.opendev.org/c/openstack/kolla-ansible/+/828757 | 10:37 |
| opendevreview | Radosław Piliszek proposed openstack/kolla-ansible stable/xena: Add openvswitch and prometheus to logrotate https://review.opendev.org/c/openstack/kolla-ansible/+/830712 | 11:15 |
| opendevreview | Radosław Piliszek proposed openstack/kolla-ansible stable/wallaby: Add openvswitch and prometheus to logrotate https://review.opendev.org/c/openstack/kolla-ansible/+/830713 | 11:15 |
| opendevreview | Radosław Piliszek proposed openstack/kolla-ansible stable/victoria: Add openvswitch and prometheus to logrotate https://review.opendev.org/c/openstack/kolla-ansible/+/830714 | 11:16 |
| opendevreview | Radosław Piliszek proposed openstack/kolla-ansible stable/victoria: Add openvswitch and prometheus to logrotate https://review.opendev.org/c/openstack/kolla-ansible/+/830714 | 11:17 |
| opendevreview | Merged openstack/kayobe stable/xena: Sync enable flag defaults with kolla ansible https://review.opendev.org/c/openstack/kayobe/+/830578 | 11:27 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe master: Remove stale interfaces https://review.opendev.org/c/openstack/kayobe/+/830222 | 11:35 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe master: Remove stale interfaces https://review.opendev.org/c/openstack/kayobe/+/830222 | 11:55 |
| jingvar | Hi! | 12:26 |
| jingvar | How can I update certs on a working cloud | 12:27 |
| yoctozepto | jingvar: hi, update them in the deployment config and rerun the deploy | 13:01 |
| jingvar | the main question what happens with cloud, recrteating enpoints etc? | 13:21 |
| priteau | mnasiadka: Are you planning to submit Kolla cycle highlights this week so they can be integrated in marketing material? | 13:26 |
| mnasiadka | priteau: it's still in the plan, although I had some unplanned support work ;) | 13:36 |
| mnasiadka | priteau: if you have any proposals for highlights, feel free to pass them to me :) | 13:40 |
| priteau | mnasiadka: for kayobe, I think the support for building multiple disk images through direct use of diskimage-builder is worth mentioning | 13:43 |
| priteau | for kolla, libvirt-exporter | 13:43 |
| mnasiadka | priteau, yoctozepto, mgoddard: https://review.opendev.org/c/openstack/releases/+/830821 - feel free to update if there's anything missing. | 14:04 |
| yoctozepto | mnasiadka: extra whitespace | 14:05 |
| yoctozepto | mnasiadka: I wonder if we can use highlights to mention deprecation of binary | 14:06 |
| yoctozepto | as it's quite a highlight | 14:06 |
| yoctozepto | jingvar: haproxy will bounce | 14:06 |
| mnasiadka | yoctozepto: so add it there, I feel like in a middle of warfare today :) | 14:07 |
| yoctozepto | so existing connections will be terminated | 14:07 |
| yoctozepto | mnasiadka: I'm not sure if it's appropriate, waiting for opinions of others | 14:07 |
| mnasiadka | yoctozepto: mgoddard is off today, so I guess we can ask priteau and frickler for opinions :) | 14:08 |
| yoctozepto | ++ | 14:08 |
| priteau | yoctozepto: I fixed the whitespace | 14:08 |
| yoctozepto | priteau ++ | 14:08 |
| yoctozepto | though that was not disputable :D | 14:08 |
| priteau | Cycle highlights give a high-level, user-focused summary of what has changed in the latest release. This is not necessarily the most technically complex work you accomplished in the release, but is the work that will have the largest impact on users. | 14:09 |
| priteau | Definitely binary should be mentioned, according to these criteria | 14:09 |
| yoctozepto | ok, so I felt right | 14:09 |
| priteau | Feel free to add it on top | 14:09 |
| yoctozepto | added | 14:10 |
| yoctozepto | I wonder if lines need to broken | 14:11 |
| yoctozepto | we will see | 14:11 |
| mnasiadka | yoctozepto: https://ethercalc.openstack.org/kolla-ovs-versions - look at the bug in the bottom (I did a proper facepalm today) :) | 14:36 |
| yoctozepto | mnasiadka: oh no! | 15:04 |
| mnasiadka | yoctozepto: currently I solved it downstream, but question if we want to solve it upstream | 15:05 |
| yoctozepto | yeah, it looks like relying on upper constraints does not make sense? | 15:06 |
| yoctozepto | what makes sense then? | 15:06 |
| yoctozepto | surely the testing is faulty if it missed it | 15:06 |
| yoctozepto | meh | 15:06 |
| mnasiadka | I don't think we're testing Octavia with OVN | 15:09 |
| yoctozepto | mnasiadka: we are | 15:10 |
| yoctozepto | but I meant for neutron/octavia | 15:10 |
| mnasiadka | Well, they test OVS version they use in CI, and I guess it's not 2.16/2.17 | 15:10 |
| mnasiadka | And the commit I located in openvswitch repo - probably that's the only case of breakage when you use OVS 2.16 | 15:11 |
| yoctozepto | mnasiadka: hmm | 15:11 |
| yoctozepto | canonical used to bump ovs in devstack | 15:11 |
| mnasiadka | based on the bug report - we should be using python ovs package version the same as installed openvswitch | 15:11 |
| yoctozepto | maybe they forgot, eh | 15:11 |
| yoctozepto | hmm | 15:12 |
| yoctozepto | we can override that one thing | 15:12 |
| mnasiadka | and what is even more bizarre - this bug is visible in Octavia, but not in Neutron (which is doing all operations in the background and properly retrying the connection until they don't hit this bug) | 15:12 |
| yoctozepto | odd | 15:20 |
| mnasiadka | well, not really odd, for some reason ovn-octavia-provider is doing the tasks and responding to user API call that they are done (instead of doing them in background) | 15:24 |
| yoctozepto | mhm | 15:32 |
| mnasiadka | but anyway, it's just a workaround to retry and retry the OVSDB transaction and wait for proper seq_no ;) | 15:43 |
| yoctozepto | yeah, kinda dumb | 15:43 |
| yoctozepto | I would go with overriding the ovs python package version | 15:43 |
| yoctozepto | just need to know the version (-: | 15:44 |
| mnasiadka | well, I guess newest should work on older OVS, just not the other way around | 15:44 |
| yoctozepto | *maybe* | 15:45 |
| yoctozepto | though I would not risk going assuming 2 versions back are still supported | 15:46 |
| yoctozepto | as it's the usual contract boundary | 15:46 |
| mnasiadka | Other option is just to pin to some version | 15:46 |
| mnasiadka | Just wondering if removing upper-constraints entry and using distro provided python3-openvswitch would not work here. | 15:47 |
| mnasiadka | (this way we would get consistent ovs pip package with installed openvswitch) | 15:48 |
| mnasiadka | And I stumbled also across this - https://review.opendev.org/c/openstack/kolla/+/701880 - wonder if it's needed anymore - because we're carrying this :) | 15:49 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe stable/victoria: Use stream8 images with molecule https://review.opendev.org/c/openstack/kayobe/+/830725 | 15:50 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe stable/victoria: Set correct gateway for the bifrost provision network https://review.opendev.org/c/openstack/kayobe/+/827745 | 15:51 |
| yoctozepto | mnasiadka: no idea, probably yes, who knows | 15:51 |
| yoctozepto | anyway, it makes sense to me to use distro-provided version | 15:51 |
| yoctozepto | same like libvirt | 15:51 |
| yoctozepto | (the reason why I prefer containerised libvirt is that we control both sides) | 15:52 |
| mnasiadka | Ok, I'll try the distro approach and let's see what breaks. | 15:53 |
| yoctozepto | ++ | 15:54 |
| yoctozepto | do you recall any other similar deps? | 15:54 |
| yoctozepto | I can't | 15:54 |
| jingvar | <yoctozepto: for some reasons openstack.kolla/centos-source-prometheus-openstack-exporter:xena "dumb-init --single-…" 9 days ago Up 20 minutes | 15:55 |
| yoctozepto | I don't understand why I'm called | 15:55 |
| jingvar | ok, thanks | 15:56 |
| opendevreview | Michal Nasiadka proposed openstack/kolla master: WIP: Use python3-openvswitch from distro https://review.opendev.org/c/openstack/kolla/+/830844 | 16:01 |
| opendevreview | Michal Nasiadka proposed openstack/kolla master: Revert "neutron: fix build in stable branches" https://review.opendev.org/c/openstack/kolla/+/830858 | 16:01 |
| mnasiadka | let's see if that is still required | 16:02 |
| mnasiadka | because looks at least funny | 16:02 |
| mnasiadka | and neutron itself is not in upper-constraints anymore | 16:03 |
| yoctozepto | mnasiadka: neutron ovs agent does not need it? | 16:05 |
| mnasiadka | https://review.opendev.org/c/openstack/requirements/+/553030 | 16:05 |
| yoctozepto | mnasiadka: I mean ovs | 16:05 |
| mnasiadka | yes, it uses ovsdbapp, which is a ,,middleware'' for talking to ovs library - but we install python3-openvswitch in neutron-base I think | 16:06 |
| mnasiadka | yes we do, in neutron-base | 16:06 |
| mnasiadka | even in source | 16:06 |
| mnasiadka | I think dropping binary in Zed will be fun, finding all these additional packages which we don't need in -source | 16:07 |
| yoctozepto | interesting | 16:08 |
| yoctozepto | so it was feeding neutron already | 16:08 |
| yoctozepto | it should be in ovs agent though | 16:08 |
| yoctozepto | no need for server to talk to ovs | 16:09 |
| yoctozepto | or linuxbridge | 16:09 |
| mnasiadka | server talks with ovs - in ovn case | 16:09 |
| mnasiadka | ovn dbs are ovsdb :) | 16:09 |
| yoctozepto | and it's the same library then, got it | 16:09 |
| yoctozepto | and l3 and dhcp also talk to it | 16:09 |
| yoctozepto | makes sense | 16:09 |
| yoctozepto | ok | 16:09 |
| yoctozepto | so we should be good with your patch | 16:09 |
| mnasiadka | but I think it was downgraded due to upper-constraints.txt entry | 16:09 |
| yoctozepto | yeah, it should have got reinstalled | 16:10 |
| yoctozepto | but you could check | 16:10 |
| yoctozepto | for sanity | 16:10 |
| mnasiadka | yeah, waiting for CI to build images and will check the logs | 16:10 |
| yoctozepto | I mean before the change | 16:11 |
| mnasiadka | ah, yeah - let's see | 16:11 |
| mnasiadka | yes, it is being downgraded to 2.13 | 16:12 |
| yoctozepto | ok, then we have our sanity | 16:14 |
| mnasiadka | hmm, not really downgraded, installed freshly in the venv under /var/lib/kolla/venv | 16:19 |
| mnasiadka | ah right, now I get it, we're creating the venv in openstack-base | 16:19 |
| mnasiadka | so python3-openvswitch would need to be installed there | 16:19 |
| yoctozepto | is not the venv including system packages? | 16:22 |
| yoctozepto | it doesn't matter when you install then | 16:22 |
| mnasiadka | it is | 16:23 |
| mnasiadka | ok then, so maybe due to upper-constraints.txt it was installing older in the venv | 16:23 |
| yoctozepto | yes | 16:23 |
| yoctozepto | it should stop now | 16:23 |
| yoctozepto | if you filtered right | 16:23 |
| yoctozepto | we'll see | 16:23 |
| yoctozepto | I've also sent "check experimental" to test a bit more | 16:24 |
| mnasiadka | https://9a0240ca9f61a595b570-86672578d4e6ceb498f2d932b0da6815.ssl.cf1.rackcdn.com/periodic/opendev.org/openstack/kolla/master/kolla-publish-centos8s-source-quay/772ff82/kolla/build/neutron-base.log | 16:24 |
| mnasiadka | yep, tried to uninstall and installed older in the venv path | 16:25 |
| opendevreview | Merged openstack/kayobe master: ntp: Remove removal of chrony container https://review.opendev.org/c/openstack/kayobe/+/827487 | 16:30 |
| mnasiadka | INFO:kolla.common.utils.neutron-base:Requirement already satisfied: ovs>=2.10.0 in /usr/lib64/python3.6/site-packages (from neutron==19.1.0.dev536) (2.15.4) | 16:56 |
| mnasiadka | looks ok | 16:56 |
| mnasiadka | Ubuntu feels the same | 16:57 |
| mnasiadka | INFO:kolla.common.utils.neutron-base:Requirement already satisfied: ovs>=2.10.0 in /usr/lib/python3/dist-packages (from neutron==19.1.0.dev536) (2.16.90) | 16:57 |
| opendevreview | Michal Nasiadka proposed openstack/kolla master: Use python3-openvswitch from distro https://review.opendev.org/c/openstack/kolla/+/830844 | 17:00 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe stable/victoria: CentOS Stream: update DNF repo file names https://review.opendev.org/c/openstack/kayobe/+/830873 | 17:06 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe stable/victoria: Use stream8 images with molecule https://review.opendev.org/c/openstack/kayobe/+/830725 | 17:07 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe stable/victoria: CentOS Stream: update DNF repo file names https://review.opendev.org/c/openstack/kayobe/+/830873 | 17:10 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe stable/victoria: Use stream8 images with molecule https://review.opendev.org/c/openstack/kayobe/+/830725 | 17:11 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe stable/victoria: Set correct gateway for the bifrost provision network https://review.opendev.org/c/openstack/kayobe/+/827745 | 17:11 |
| frickler | should we mention ansible-kolla-collection in the highlights or that something for the next cycle? | 17:18 |
| opendevreview | Franco Mariotti proposed openstack/kolla-ansible master: exposes Prometheuss exporters on kolla external interface https://review.opendev.org/c/openstack/kolla-ansible/+/830877 | 17:25 |
| mnasiadka | frickler: I don't know if that's any feature for the users, we might add something later when that's merged and working in Kayobe :) | 17:31 |
| opendevreview | Michal Nasiadka proposed openstack/kolla master: Use python3-openvswitch from distro https://review.opendev.org/c/openstack/kolla/+/830844 | 17:38 |
| opendevreview | Franco Mariotti proposed openstack/kolla-ansible master: exposes Prometheuss exporters on kolla external interface https://review.opendev.org/c/openstack/kolla-ansible/+/830877 | 17:45 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe stable/victoria: CI: set kayobe-*-upgrade-centos8s jobs non-voting https://review.opendev.org/c/openstack/kayobe/+/830886 | 18:41 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe stable/victoria: CI: set kayobe-*-upgrade-centos8s jobs non-voting https://review.opendev.org/c/openstack/kayobe/+/830886 | 19:11 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe stable/victoria: Set correct gateway for the bifrost provision network https://review.opendev.org/c/openstack/kayobe/+/827745 | 19:12 |
| opendevreview | Pierre Riteau proposed openstack/kayobe stable/victoria: CI: drop failing jobs on stable/victoria https://review.opendev.org/c/openstack/kayobe/+/830907 | 21:59 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe stable/victoria: CI: temporary disable voting of some jobs https://review.opendev.org/c/openstack/kayobe/+/830886 | 22:27 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe stable/victoria: CentOS Stream: update DNF repo file names https://review.opendev.org/c/openstack/kayobe/+/830873 | 22:29 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe stable/victoria: Use stream8 images with molecule https://review.opendev.org/c/openstack/kayobe/+/830725 | 22:29 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe stable/victoria: Set correct gateway for the bifrost provision network https://review.opendev.org/c/openstack/kayobe/+/827745 | 22:31 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!