14:00:03 #startmeeting tripleo 14:00:04 Meeting started Tue Jan 5 14:00:03 2021 UTC and is due to finish in 60 minutes. The chair is marios. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:05 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:07 The meeting name has been set to 'tripleo' 14:00:15 #topic agenda 14:00:15 * Review last minutes & action items 14:00:15 * One off agenda items 14:00:15 * Bugs & Blueprints 14:00:15 * Projects releases or stable backports 14:00:18 * Specs 14:00:20 * open discussion 14:00:23 Anyone can use the #link, #action and #info commands, not just the moderatorǃ 14:00:25 Hello folks, who is around today? o/ 14:00:34 o/ 14:00:35 «o/ 14:00:45 0/ 14:00:57 o/ 14:01:17 o/ 14:01:38 great hello folks, let's get going 14:01:48 #topic review last meeting logs & action items 14:01:59 #link http://eavesdrop.openstack.org/meetings/tripleo/2020/tripleo.2020-12-22-14.00.html 14:02:04 last logs there ^^^ 14:02:06 o> 14:02:13 checking for any action items 14:02:24 we had a request for victoria release i already have that on agenda 14:02:24 \o/ 14:02:41 anything else from last week folks want to mention/raise/revisit here? 14:03:28 k then moving on ... 14:03:34 #topic one off agenda items 14:03:36 #link https://etherpad.openstack.org/p/tripleo-meeting-items 14:03:44 we have a few there... lets take them in order 14:04:05 #info [slagle] Ephemeral Heat spec for review: https://review.opendev.org/c/openstack/tripleo-specs/+/765000 14:04:20 just wanted to point out the spec 14:04:23 slagle: o/ you want to say something about that? besides the spec is there a topic branch with something to look at? 14:04:36 we're moving forward with a prototype, so it's last call for spec comments 14:04:52 o/ 14:04:52 should have a branch up soon 14:05:03 slagle: ack. thanks 14:05:07 anything else on this topic? 14:05:27 #info [tengu] Tripleo dependency: the podman case http://lists.openstack.org/pipermail/openstack-discuss/2021-January/019625.html 14:05:38 ah, that's me :) 14:05:43 Tengu: sent a message to the list earlier today about this ... Tengu want to hilight something in particular? 14:06:06 well, just poking around in order to get inputs.... that case is pretty annoying and reccurring 14:06:26 as said in my last mail in the thread, we'd have the same case for the "virt" module on OSP, apparently. 14:06:53 soooo yeah.... something ensuring we're using the right stream, in a generic way, is kind of mandatory at some point. People don't read doc anyway. 14:07:00 Tengu: is there support for the tripleo-ansible case in the deployment framework team or haven't discussed yet? 14:07:15 i'd say drop the REquires in tripleoclient, and more generally, we should do that for a lot of things 14:07:18 (or are using automated tools and they don't really update them... /me hides) 14:07:33 marios: I've mostly discussed with Delivery at this point 14:07:45 there are too many things there that are required for the deployment, but not required by the actual python code 14:07:53 slagle: yah i hilighted that in the thread, i mean it impacts the ephemeral heat case since afaics right now the heat leanucher is (one of ) the main users right now for tha trequirement 14:08:02 slagle: so can we? remove that dependency? 14:08:07 slagle: how? use somehting not podman? 14:08:08 slagle: though a container is launched for heat from within the tripleoclient directly. So we'll need to do some ordering at some point 14:08:21 that's the beauty of this thing :) 14:08:51 anyway... please jump in the thread (or LP https://bugs.launchpad.net/tripleo/+bug/1910217) and let's find a clean solution 14:08:55 Launchpad bug 1910217 in tripleo "Package dependencies, podman, dnf modules" [Medium,Triaged] 14:08:56 the heat container isn't used by default 14:09:13 not sure it even works tbh 14:09:14 guess that's a nice 2021 topic: dependency cleanup and consolidation :) 14:09:28 slagle: k i just did a quick grep on the code when replying to the thread so maybe it isn't as hard a requirement as i thought 14:09:51 slagle: oh? really? I'd think it was used by default when deploying the undercloud/standalone in order to get the heat service. 14:10:07 if it's not used/working..... let's drop it for the sake of sanity :) 14:10:25 Tengu: thanks for bringing that topic and hilighting here and thanks slagle for jumping in 14:10:27 don't think so. pretty sure heat-all is still run uncontainerized 14:10:40 i'll reply on the thread 14:10:41 yes it's not containerized 14:10:43 slagle: since it's linked to your spec... :) 14:10:45 it's an option 14:10:48 but not the default 14:10:58 as discussed @ptg 14:11:04 mwhahaha: thanks for confirming that's what slagle is saying too 14:11:23 i thought it was default but i only did a quick grep to see what was calling podman 14:11:33 * Tengu prepars the kill-shot on the Requires podman 14:11:43 great, lets carry on in the list 14:11:47 moving on ... . 14:11:49 so we also have modules configurable in THT 14:11:53 there's some other options 14:12:35 #info [fultonj] Adding an SSH private key on overcloud node for cephadm https://review.opendev.org/c/openstack/tripleo-ansible/+/768365 https://specs.openstack.org/openstack/tripleo-specs/specs/wallaby/tripleo-ceph.html 14:12:37 Grzegorz Grasza proposed openstack/puppet-tripleo master: Add authtoken security options https://review.opendev.org/c/openstack/puppet-tripleo/+/769349 14:12:41 fultonj: o/ ^^ are you around for that one? 14:12:46 yes 14:13:02 basically it's a change for the cephadm work 14:13:14 Tengu: I didn't get the point in ML, why not to use Conflicts: podman < lower? 14:13:17 just asking for reviews 14:13:39 ack thanks fultonj 14:13:48 sorry, keep it for the open discussion 14:13:50 bogdando: Tengu: we can pickup again in the open session at the end? 14:13:52 bogdando: thanks 14:13:56 :) 14:14:09 #info [fultonj] Just need one more patch to make new ceph client migration complete https://review.opendev.org/c/openstack/tripleo-heat-templates/+/760915 https://specs.openstack.org/openstack/tripleo-specs/specs/wallaby/tripleo-ceph-client.html 14:14:16 brb 14:14:20 fultonj: o/ ^^ that one too 14:14:29 ack, one more reivew and new ceph client config work is complete 14:14:46 great work thanks fultonj 14:14:51 any comments or questions on those? 14:15:19 #info [fultonj] https://review.opendev.org/c/openstack/tripleo-specs/+/760536 merged 14:15:28 cool looks like the ports v2 spec merged zo/ 14:15:32 thanks hjensas|PTO 14:15:38 anything else on that one? 14:15:50 i know a few things depend on it like ephemeral heat spec 14:16:16 k moving on 14:16:23 #info (marios): promotion CI docs @ https://review.opendev.org/c/openstack/tripleo-docs/+/767375 (build@ https://f267574198a517c003d1-a40b9478ae1bd073dc32f331038fe6d7.ssl.cf5.rackcdn.com/767375/7/check/openstack-tox-docs/8252ffe/docs/ci/stages-overview.html ) 14:16:28 ports v2 spec depends on ephemeral heat? 14:16:38 fultonj: no other way around i believe 14:16:45 ok, that makes more sense 14:16:48 so we can get the network provision done before the heat stack create 14:16:57 marios: ++ 14:17:22 k so promotion docs you may have heard the ci team talk about a 'victoria promotion' or a master promotion etc that doc explains that if you are interested in our process 14:17:30 #info (marios): reduce content providers (upgrade jobs) https://review.opendev.org/q/topic:reduce-content-providers 14:17:47 this is some ongoign work that will reduce the number of content providers we have to run for upgrade jobs 14:18:04 the current situation is pretty bad, like standalone-upgrade-ussuri is running content providers for master, victoria, ussuri and train right now 14:18:15 #info (rfolco): https://review.opendev.org/c/openstack/tripleo-ci/+/767591 Enable html report for content provider jobs https://fa41a0ce81614565dae6-7ca08d7dc339532aeab9467cc90f03e8.ssl.cf1.rackcdn.com/767591/15/check/tripleo-ci-centos-8-content-provider/3f24737/logs/report.html 14:18:26 i added that for rfolco ^ anything you want to add there? 14:18:45 thats it 14:18:56 thanks rfolco that adds a nice report for all the container builds 14:18:57 self-explanatory 14:19:02 thx 14:19:02 will make our lives much easier debugging 14:19:05 rfolco:++ 14:19:10 #info made a stable/victoria release for repos at https://review.opendev.org/c/openstack/releases/+/768237 14:19:24 that was on request from ykarel last meeting it merged yesterday i believe ^^^ fyi 14:19:37 #info [release] Status: RED - pip resolver issue with publish-openstack-releasenotes-python3 http://lists.openstack.org/pipermail/openstack-discuss/2021-January/019611.html ** please help merge ** https://review.opendev.org/q/topic:%22fix-relmgt-pip-doc%22+(status:open%20OR%20status:merged 14:19:56 that came up yesterday related to the release - will need some help merging and fixing the ones that aren't working from the topic branch (tripleo repos) 14:20:05 that does look nice rfolco 14:20:08 sorry for rush trying to get through quickly 14:20:13 fiunally 14:20:20 #info reminder re [TripleO] moving stable/rocky for tripleo repos to unmaintained (+ then EOL) OK? http://lists.openstack.org/pipermail/openstack-discuss/2020-December/019338.html - we'll start moving that forward - possibly bump the thread a bit first to hilight. 14:20:40 so just bringing that up again, probably i should bump the thread but we will start making that so now since there were no objections so far 14:20:46 please see and reply if you have comments 14:21:06 we can revisit any of those topics in the open session at the end i am going to move on 14:21:13 thanks all for those topics 14:21:23 #topic Bugs and blueprints 14:21:30 #link https://bugs.launchpad.net/tripleo/ 14:21:30 #link https://storyboard.openstack.org/#!/project/openstack/tripleo-ansible 14:21:32 #link https://launchpad.net/tripleo/+milestone/wallaby-2 14:21:38 #link https://launchpad.net/tripleo/wallaby 14:21:58 as always a reminder to please check the wallaby-2 bugs and close out anything stale that belongs to you 14:22:25 any bugs folks want to hilight here? are the ruck/rover around anything worth mentioning weshay|ruck ? 14:23:09 just FYI.. to folks doing local deployments w/ centos ens3 is back as ens3-1 14:23:10 https://bugs.launchpad.net/tripleo/+bug/1910107 14:23:11 Launchpad bug 1910107 in tripleo "master/wallaby undercloud install failure unable to start network service" [Critical,Triaged] 14:23:28 https://review.opendev.org/c/openstack/tripleo-image-elements/+/769209 14:23:50 thank you weshay|ruck 14:24:05 moving on if there is nothing else from anyone 14:24:28 #topic Project releases or stable backports 14:24:30 #info tripleo wallaby repos https://releases.openstack.org/teams/tripleo.html#wallaby 14:24:51 anything here? any release requests? we should probably do a 'final' rocky release soon before we nuke it... 14:25:20 wallaby-3 is a way off apparently, mar 08-12 https://releases.openstack.org/wallaby/schedule.html 14:25:33 k if there is nothing here i will move on to specs 14:25:46 (I am using the agenda at https://wiki.openstack.org/wiki/Meetings/TripleO for these things by the way ) 14:25:49 for copy/pasta 14:26:01 #topic specs 14:26:07 #info https://review.opendev.org/q/project:openstack/tripleo-specs 14:26:07 #info https://opendev.org/openstack/tripleo-specs/src/branch/master/specs/wallaby 14:26:22 k so anything else here? we had a mention on the ephemeral heat already 14:26:24 the ports v2 merged 14:26:32 i think the frrouter can merge lemme fetch link 14:27:03 #link Spec for installing and configuring FRRouter for BGP https://review.opendev.org/c/openstack/tripleo-specs/+/758249 14:27:07 reviews please? ^^^ 14:27:39 anything else on this topic for today? 14:28:11 k... moving on 14:28:18 #topic open discussion 14:28:23 Anything else that folks want to bring up to the meeting? 14:28:43 bogdando: want to talk about the conflict thing? 14:28:45 Tengu: bogdando: mwhahaha: was there further discussion around the podman version issue 14:29:05 we're trying to solve it in tripleo-repos as well 14:29:19 i don't think we any rpm dep is really going to help 14:29:31 as we've already tried that and it's caused problems 14:30:01 so... while it seems nice to set upper/lower constraints for a given package, it might happen it's not that easy, and, as written in the ML, it will lead in a ton of LP/BZ about "unable to install, package conflict". I'm not the one wanting to sort them out just because ppl don't read doc, don't understand and other reasons ;). 14:30:40 without module deps in rpm specs we just have to document it and try and fixup the initial setup requirements 14:30:42 mwhahaha: right. I've broken teeth on this issue last year, and now... well, if we can get rid of it once for all and use a proper ansible thing in order to get the right version directly, I'm all for it. 14:31:03 though switching module stream when packages are installed might fail -.-' 14:31:13 without creating a meta package that'll do all the things (setup repos, install stuff) it's not something really solvable 14:31:48 we already have issues with versions in ci, which is why we were going to try and express all the versions in tripleo-repos 14:31:55 yeah..... maybe such a "meta-thing" is a better way than dropping the Requires and relying on tripleo-ansible roles? 14:32:10 https://review.opendev.org/c/openstack/tripleo-repos/+/767214 14:32:20 the other issue is the lack of an equivalent downstream for customers 14:32:30 yup 14:32:30 like there needs to be an initial package that does something like that 14:32:48 we tried to add Requirements on the osp-release (not sure about the name though) 14:32:49 that would be my recommendation is that the repo setup is automated via a cli tool 14:32:58 cloudnull, Tengu , https://review.opendev.org/c/openstack/tripleo-heat-templates/+/764988 please 14:32:58 so I was thinking about a simple way to fail the client package install if podman looks old... 14:33:02 and that be the 'official' way of doing thigns 14:33:08 bogdando: the issue isnt' too old, it's too new 14:33:22 bogdando: so we don't want container-tools:rhel8 for train 14:33:24 and conflicts: alongside requires: looked like a good fit. apparently not 14:33:25 cause it causes problems 14:33:40 bogdando: it's not a matter of "old" in some cases... container-tools:rhel8 provides podman 2.x, while we need 1.6.4 (at a critical sub-version) 14:33:44 we already tried version constraints and it leads to poor UX due to conflict messaging 14:33:59 also that -^ thanks mwhahaha 14:34:27 reducing the user's ability to miss a setup step is an improvement 14:34:43 having to specify module container-tools:2.0 is too much it appears 14:34:45 mwhahaha: having tripleo-repos setting everything fine is better, and ensuring via tripleo-ansible everything is actually GOOD is a nice to have 14:34:58 there's a THT change recent to ensure module versions 14:35:05 for upgrades iirc 14:35:07 for upgrades but could be used to ensure overcloud 14:35:11 and undercloud 14:35:12 Jose did it last year 14:35:16 it just landed 14:35:23 nice 14:35:39 (and there are validations.... I know, you don't like them, still....) 14:35:48 https://review.opendev.org/c/openstack/tripleo-heat-templates/+/767994+ 14:35:53 validations is too late in the game 14:36:14 not with the prep ones, but one has to run them. if they don't read the doc.... no way they run anything 14:36:32 Ade Lee proposed openstack/puppet-tripleo stable/victoria: Adding key_size option on the certmonger_certificate function https://review.opendev.org/c/openstack/puppet-tripleo/+/769195 14:36:36 the other goal is not to have to define these requirements in multiple palces 14:36:43 +42 14:36:51 so validations will require a duplicate update to something that may set it up 14:37:01 i would prefer we settle on one thing 14:37:06 I thought about the tripleo-ansible/tripleo_podman since it's, well, linked to podman.. 14:37:25 it's not what installs on the undercloud unfortunately which is where we traditional hit this issue 14:37:36 mwhahaha, Tengu , cloudnull https://review.opendev.org/c/openstack/tripleo-heat-templates/+/767933 please 14:37:36 but if we're going for a tripleo-repos modification, tripleo-ansible change might be useless. That is, as long as ppl are using this tripleo-repos, of course (they should be.) 14:37:44 i think tripleo-ansible is also too late 14:38:00 mwhahaha: depends - if we drop the Requirement from the spec... ? 14:38:29 mwhahaha, Tengu: ok, then that could be conflicts: podman > 1.9 14:38:41 Ade Lee proposed openstack/puppet-tripleo stable/ussuri: Adding key_size option on the certmonger_certificate function https://review.opendev.org/c/openstack/puppet-tripleo/+/769356 14:38:42 bogdando: conflicts not going to help, it'll just be the same confusing error 14:39:02 a fixed fence to never accept 2.0 for that release, and no need to poke it with newer 1.6.x 14:39:07 confusion leading to massive spam on BZ -.-' 14:39:17 ...or >= 2.0 14:39:27 1.6 and 1.9 are not compatible either 14:39:29 (seen that one) 14:40:38 what do you think folks? are we done on this one for now/here? 14:40:49 this is not something that can be solved simply, we'll need to talk more about the overal process 14:40:50 sounds like we are leaning towards the tripleo-repos review currently 14:40:58 but there can be more discussion upstream on the list too 14:40:58 guess so. Would love seeing your thoughts in the thread, mwhahaha :) 14:41:03 (or LP) 14:41:13 closed: invalid? got it 14:41:13 thank you for bringing that today Tengu 14:41:23 marios: yep - I'm all for the tripleo-repos change 14:41:25 * Tengu tortures mwhahaha with a scary Windows© ME user guide 14:41:29 any other topics someone wants to hilight here? 14:41:45 mwhahaha: I don't think installing with 2.0 podman instead of 1.6.x would improvie UX 14:41:55 so that would fix nothing from the UX pov... 14:42:08 bogdando: it will most likely crash at some point. 14:42:35 bogdando: container-tools:2.0 not podman 2.0 14:42:35 bogdando: also, beware, container-tools:2.0 doesn't provide podman-2.x -.- (yeah.... don't ask) 14:42:36 anwyay 14:43:50 thank you to everybody for participating and following along today 14:44:06 #info next meeting tuesday 19th Jan (2 weeks) 14:44:13 #endmeeting tripleo