| opendevreview | Verification of a change to openstack/ironic stable/2024.2 failed: Gracefully handle bad request exception https://review.opendev.org/c/openstack/ironic/+/933036 | 00:10 |
|---|---|---|
| opendevreview | Merged openstack/ironic stable/2024.2: Gracefully handle bad request exception https://review.opendev.org/c/openstack/ironic/+/933036 | 03:10 |
| rpittau | good morning ironic! o/ | 07:57 |
| opendevreview | Pierre Riteau proposed openstack/tenks stable/2.0: CI: Use 2024.1 requirements on stable/2.0 branch https://review.opendev.org/c/openstack/tenks/+/935567 | 08:29 |
| opendevreview | Pierre Riteau proposed openstack/tenks stable/2.0: CI: Use 2024.1 requirements for Python 3.8 job https://review.opendev.org/c/openstack/tenks/+/935567 | 08:50 |
| opendevreview | Verification of a change to openstack/tenks stable/2.0 failed: CI: Use 2024.1 requirements for Python 3.8 job https://review.opendev.org/c/openstack/tenks/+/935567 | 09:19 |
| opendevreview | Pierre Riteau proposed openstack/tenks stable/2.0: CI: Use 2024.1 requirements for Python 3.8 job https://review.opendev.org/c/openstack/tenks/+/935567 | 09:27 |
| opendevreview | Merged openstack/tenks stable/2.0: CI: Use 2024.1 requirements for Python 3.8 job https://review.opendev.org/c/openstack/tenks/+/935567 | 10:11 |
| opendevreview | OpenStack Release Bot proposed openstack/tenks stable/2.0: Update .gitreview for stable/2.0 https://review.opendev.org/c/openstack/tenks/+/927159 | 10:16 |
| opendevreview | Pierre Riteau proposed openstack/tenks stable/2.0: Remove deprecated use of include https://review.opendev.org/c/openstack/tenks/+/935624 | 10:40 |
| opendevreview | Riccardo Pittau proposed openstack/ironic master: Use OVN and OVS from OS packages in CI https://review.opendev.org/c/openstack/ironic/+/935628 | 11:06 |
| opendevreview | Merged openstack/ironic stable/2024.2: Actually ignore [inspector]power_off with fast track https://review.opendev.org/c/openstack/ironic/+/934424 | 11:31 |
| opendevreview | Merged openstack/ironic stable/2024.2: Fix policy checks added with runbooks https://review.opendev.org/c/openstack/ironic/+/935530 | 11:55 |
| opendevreview | cid proposed openstack/ironic-specs master: Add a Kea DHCP backend https://review.opendev.org/c/openstack/ironic-specs/+/931025 | 12:20 |
| opendevreview | cid proposed openstack/ironic master: Fix double JSON encoding of error message https://review.opendev.org/c/openstack/ironic/+/931795 | 12:21 |
| opendevreview | cid proposed openstack/ironic master: Use microversion handler https://review.opendev.org/c/openstack/ironic/+/935644 | 12:22 |
| opendevreview | cid proposed openstack/ironic master: [WIP] Agentless Kea DHCP backend. https://review.opendev.org/c/openstack/ironic/+/935645 | 12:22 |
| opendevreview | cid proposed openstack/ironic master: Fix double JSON encoding of error message https://review.opendev.org/c/openstack/ironic/+/931795 | 12:26 |
| cardoe | Morning ironic. | 12:59 |
| TheJulia | good morning | 14:01 |
| rpittau | good night! o/ | 16:51 |
| stephenfin | adamcarthur5: Sorry, I forgot to open Hexchat yesterday (and most of today, apparently). I have the WIP comment addressed since in nova. Let me respin with my attempt at a resolution and you can take it and rework as you see fit? | 17:19 |
| JayF | stephenfin: I think we were both confused about what that comment meant; it looked like e.g. that shards change does what needs to be done | 17:32 |
| JayF | so might be helpful to restate the issue in more specific detail if it's not 100% resolved or you want us to be able to tell what's going on | 17:32 |
| stephenfin | Yeah, that was poorly worded. I should have said "we don't use all the code yet" | 17:51 |
| JayF | jfyi our CI just got switched to noble, apparently | 18:16 |
| JayF | so be very aware when rechecking, I had a change get bombed just now :/ | 18:16 |
| opendevreview | Stephen Finucane proposed openstack/ironic master: api: Introduce new mechanism for API versioning https://review.opendev.org/c/openstack/ironic/+/928919 | 18:24 |
| opendevreview | Stephen Finucane proposed openstack/ironic master: api: Add schema validation framework https://review.opendev.org/c/openstack/ironic/+/928920 | 18:24 |
| opendevreview | Stephen Finucane proposed openstack/ironic master: WIP: api: Add schema for allocations API https://review.opendev.org/c/openstack/ironic/+/928921 | 18:24 |
| opendevreview | Stephen Finucane proposed openstack/ironic master: api: Don't use private attributes https://review.opendev.org/c/openstack/ironic/+/935697 | 18:24 |
| opendevreview | Jay Faulkner proposed openstack/ironic master: Use specific fix-commit from dnsmasq https://review.opendev.org/c/openstack/ironic/+/935699 | 18:39 |
| JayF | ^ is the new version of the dnsmasq workaround. We're kinda lucky it didn't land in the old version of it, bceause it would've blocked all CI when noble was switched to | 18:40 |
| JayF | although TBH I don't know if CI is happy right now or not :) was going to let that patch double as my clean test | 18:40 |
| TheJulia | I guess time will tell | 18:51 |
| JayF | I did email dnsmasq-discuss and update the ubuntu bug | 18:53 |
| JayF | asking for a release and a backport to noble | 18:53 |
| JayF | so hopefully we can get outta the dnsmasq building business one way or another, but at least now we can build a *shiny new version* instead of downgrading to a less-broken one | 18:53 |
| TheJulia | wheeeeee | 19:18 |
| TheJulia | JayF: hey, what was the feedback regarding out docs and networking details? I've got a downstream bug of sorts I'm looking at which basically points to upstream docs and I'm looking at them with a differen that on and thinking "wow, we did a bad job at setting the wrong context in the networking docs" | 20:17 |
| JayF | I don't remember specifically, it'd be in the full document -- but that sounds like part of the thread through everything: talking about what things are theoretically possible instead of giving opinionated use cases and how to implement them | 20:18 |
| JayF | I can search for the full doc | 20:18 |
| TheJulia | This seems like a lot of "focusing on how you can" and failing to build the base context | 20:19 |
| JayF | https://docs.google.com/document/d/1e9URuPHKNTx5QXdkCFAzsS0EAAxPJ-xQC4BEpDwESp4/edit?tab=t.0 my recollection is accurate | 20:19 |
| JayF | yep, exactly | 20:19 |
| TheJulia | basically, the question got raised "how do I do other non-provisioning" network port configuration | 20:20 |
| TheJulia | and some of that *is* defeinitely theory we seem to leave out | 20:20 |
| TheJulia | We're like... missing the "here is the 101 level class on how this works | 20:20 |
| JayF | extremely yes | 20:21 |
| TheJulia | And as much as I think every vendor would prefer customers just focusing on their downstream docs, people go look upstream and even downstream docs folks go look at upstream | 20:21 |
| JayF | yeah | 20:22 |
| JayF | similarly I have an ask from my (GR-OSS, not GR) downstream to help with like ... a reference architecture for running OpenStack + Kubernetes | 20:22 |
| JayF | not a "how do you arrange your k8s sandwich" doc, but literally a "if you don't know what to do; do it this way" kinda thing | 20:22 |
| JayF | and I think advanced (physical) networking is one of the big holes in the pieces that exist regardingt hat already | 20:23 |
| cardoe | OpenStack + Kubernetes in what sense? | 20:29 |
| cardoe | OpenStack on top of Kubernetes? | 20:29 |
| cardoe | Look at you guys caring about docs. Meanwhile I'm working on my fork of the keystone docs. | 20:34 |
| adamcarthur5 | stephenfin Thank you, much appreciated | 20:34 |
| opendevreview | Merged openstack/ironic-tempest-plugin master: Fix failure cases around test_ports https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/934387 | 20:36 |
| JayF | cardoe: the interesting thing is the answer can be either thing in either palce; which is part of why we want a good example | 20:44 |
| cardoe | https://github.com/rackerlabs/genestack is OpenStack on top of Kubernetes | 20:44 |
| cardoe | That's Kevin Carter's baby. | 20:45 |
| cardoe | https://review.opendev.org/c/openstack/tenks/+/907112 if you wanna +W that. That'll land Adam's chain. | 20:45 |
| cardoe | My baby is https://github.com/rackerlabs/understack which is very rough but it's Ironic (Nova, Neutron, Glance, Keystone) + Nautobot on top of Kubernetes. | 20:45 |
| JayF | The fact you, at a single comapny, just listed two examples is emblematical of the problem | 20:46 |
| JayF | when so many things are possible it leads to decision paralysis | 20:46 |
| JayF | (while asking for a review on a third) | 20:47 |
| JayF | cardoe: you didn't +2 907112, you only +1'd | 20:47 |
| cardoe | I can't +2 it | 20:47 |
| JayF | that is a bug in that ACL then, tenks is an ironic project and should be following the rest of the stuff | 20:48 |
| JayF | might wanna have a look at project-config, I'm not going to rabbithole on it now | 20:48 |
| cardoe | So different things for different folks AFAIK Kubernetes. So genestack uses kube-ovn so that the services are all plugged in together and magical with OVN/OVS Neutron or something. | 20:58 |
| cardoe | I dunno. I defer to jamesdenton on magical unicorn network bits. | 20:59 |
| cardoe | understack is baremetal kubernetes so that we can do DHCP and stuff. | 20:59 |
| JayF | Yeah, but I think you're nibbling around the entire point: there are a thousand knobs | 21:06 |
| cardoe | Correct | 21:07 |
| JayF | and when we are trying to pitch people on OpenStack, or more generally, and upstream-focused cloud stack including OpenStack/Ironic elements and Kubernetes, it'd be nice to have one that's relatively proscriptive for most people. | 21:07 |
| cardoe | 2^1000 possibilities | 21:07 |
| JayF | we need the sales engineer ppt for reference architecture :D | 21:07 |
| cardoe | Well that's the thing. That's not OpenStack. | 21:07 |
| cardoe | OpenStack is like the big box of multi colored Legos that's instruction books are a bunch of pictures of stuff you can build without the actual steps. | 21:08 |
| cardoe | Some projects *cough* keystone *cough* resent if you try to provide instructions. While others (Ironic) embrace having details in the docs. | 21:09 |
| JayF | I'm gonna be honest, I don't think it's productive to talk about other projects' flaws in Ironic's IRC channel | 21:09 |
| cardoe | I'm just grumpy right now. I'll stop. | 21:09 |
| JayF | I know you got issues with some of the stuff over there, but if we can keep positive and goal oriented we can get more done | 21:10 |
| cardoe | But ultimately OpenStack embraces the possibilities of what you can do. | 21:10 |
| cardoe | I dunno if you're going to come up with one answer. I know some folks are working on the OpenStack operators on top of OpenShift. There's OpenStack Helm. | 21:12 |
| JayF | well the thing is, especially for software like OpenStack -- and Ironic moreso given we even have more modes of operation (bifrost, metal3, standalone, integrated) -- that even though almost infinite things are possible; it's valuable for us to identify one or two use cases and have an opinion about how to use them | 21:13 |
| JayF | if nothing else to help people get started with their own architecture with something other than a blank sheet of paper | 21:13 |
| JayF | I think I just cracked the nut on getting CI working for Gentoo DIB again \o/ | 21:14 |
| JayF | which means I can pick up the ramdisk building | 21:14 |
| cardoe | So I do agree there and that's what my goal is. | 21:14 |
| cardoe | My downstream is a configurable deployment within my narrowly defined use case. | 21:15 |
| JayF | https://review.opendev.org/c/openstack/ironic/+/935699 is passing CI with the new version of the dnsmasq fix | 21:15 |
| JayF | if someone wants to land it I suspect it'll have a positive impact on CI randomness | 21:15 |
| cardoe | You also get a +1 from me on your Gentoo dev test by using make PREFIX=/usr install instead of the sed. | 21:17 |
| JayF | yeah tbh that original sed drove me a little bonkers but I'm wise enough not to try and shove a style fix through :) | 21:18 |
| JayF | I also tested to ensure it worked lol | 21:18 |
| JayF | #1 thing you learn working on Gentoo: never ever ever trust a build system | 21:18 |
| JayF | OMG cardoe I forgot | 21:18 |
| JayF | cardoe: I voluntell you to review this and help me maintain it: https://review.opendev.org/c/openstack/ironic/+/93569 | 21:19 |
| JayF | lol | 21:19 |
| JayF | (is it that easy?) | 21:19 |
| JayF | that link is wrong | 21:20 |
| cardoe | I already +'d it | 21:20 |
| JayF | https://review.opendev.org/c/openstack/diskimage-builder/+/923985 | 21:20 |
| JayF | I meant this one | 21:20 |
| cardoe | OH NO. | 21:20 |
| cardoe | Even before I clicked it I'm scared. | 21:20 |
| JayF | I've done all the hard work | 21:21 |
| cardoe | You know prometheanfire should also help review and maintain. | 21:22 |
| cardoe | I was his dev mentor. | 21:23 |
| JayF | he wrote the original stuff | 21:23 |
| JayF | basically we wanna fix this | 21:23 |
| JayF | move tinyipa ramdisk to a gentoo+dib ramdisk (since gentoo should let us slim it down dramatically) | 21:23 |
| JayF | then I wanna go over on gentoo wiki side and document this as a useful tool because it's the most gentoo thing I've ever seen outside of gentoo | 21:23 |
| JayF | I think Sam James from over there is interested in using DIB to build gentoo images once I get it working :D | 21:24 |
| cardoe | You're changes are actually definitely positive fixes and you've got good comments. | 21:24 |
| cardoe | That makes sense. | 21:25 |
| JayF | with a cached stage3 | 21:26 |
| JayF | using binary packages by default | 21:26 |
| JayF | the build is BLAZING FAST | 21:26 |
| cardoe | heh until you change a USE flag. | 21:29 |
| JayF | portage is pretty smart these days about handling that; basically if you hook up the binhost, it just works. You get binaries when the use flags match (and they are building multiple profiles; so you have more than one shot at matching even on a custom system) and you build em when they don't | 21:30 |
| JayF | we don't default to them because some of our test clouds are too old, but they even have x86-64-v3 binaries available | 21:30 |
| JayF | I think Eli also even has that build machine up to build random packages with random use flags as a test and as a bonus (you might win the lottery and have a version you need to upgrade to built for you) | 21:31 |
| cardoe | don't you talk like that about OSPC | 21:31 |
| JayF | OSPC? | 21:32 |
| cardoe | Your old thing at Rackspace. | 21:33 |
| JayF | ooh | 21:33 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!