| opendevreview | Jacob Anders proposed openstack/sushy master: Revert "Skipping UsbCd workaround on Supermicro ARS-111GL-NHR" https://review.opendev.org/c/openstack/sushy/+/981089 | 00:10 |
|---|---|---|
| cardoe | Anyone on that's got some boxes they'd be happy to share a "openstack baremetal node -f yaml $NODE_ID > node.yaml" and "openstack baremetal node inventory save -o inventory.json $NODE_ID". Feel free to sanitize first. | 02:39 |
| opendevreview | Doug Goldstein proposed openstack/ironic master: Add Zuul job to test inspection rules tester https://review.opendev.org/c/openstack/ironic/+/981423 | 03:04 |
| opendevreview | Doug Goldstein proposed openstack/ironic master: add GitHub Action for inspection rules testing https://review.opendev.org/c/openstack/ironic/+/981424 | 03:04 |
| cardoe | Glimmr_: I just ported our GitHub Action to the Ironic repo. And made a Zuul test for running it in Gerrit. That should give you an idea how you can use that. If we do in fact have a bug, let's add some test data with an example rule to exercise the bug. | 03:07 |
| cardoe | One of the things we're still lacking is one more thing of "expected" so like the rule can succeed against the node but need to check that it mutated the data correctly. | 03:08 |
| cardoe | A number of the hooks can probably be re-implemented as rules in the engine and we can validate this stuff ahead of time. | 03:09 |
| cardoe | If we're cool with it I can toss a couple of machines into that. | 03:21 |
| cardoe | We're not actually using rules yet but I think it's got a lot of promise. | 03:25 |
| rpittau | good morning ironic! happy friday ! o/ | 06:51 |
| abongale | good morning ironic 0/ | 07:37 |
| frickler | that ESP copying idea is really nice indeed | 08:25 |
| opendevreview | Victor proposed openstack/ironic master: Record VIF attach/detach failures in node history https://review.opendev.org/c/openstack/ironic/+/981459 | 08:43 |
| victor_obi_ | Good day! I am looking at bug #2100707 (Node history API alignment). Is this a suitable task for me to attempt as a contribution right now, or is there a specific roadmap for this issue that I should be aware of?. Thank you. | 09:05 |
| opendevreview | Victor proposed openstack/ironic master: Log node history on vif_attach and vif_detach failures https://review.opendev.org/c/openstack/ironic/+/981474 | 10:22 |
| victor_obi_ | Hi, I've submitted my second patch for bug #2131973 (vif_attach/vif_detach should log failures to node history): https://review.opendev.org/c/openstack/ironic/+/981474 — took a bit longer due to sorting out some git branching issues but it's clean now. All feedbacks are appreciated. | 10:26 |
| opendevreview | Harald Jensås proposed openstack/networking-baremetal master: Refactor to use guard clauses and early returns https://review.opendev.org/c/openstack/networking-baremetal/+/981479 | 10:48 |
| opendevreview | Harald Jensås proposed openstack/networking-baremetal master: Add segment_id to subport binding profiles https://review.opendev.org/c/openstack/networking-baremetal/+/981480 | 11:13 |
| victor_obi_ | Hi, I've now submitted two patches (#981399 and #981474). Are there any other node history related issues I could help with? | 11:25 |
| Mahnoor | Hello o/ I will be out on Monday, so I've left the bug deputy updates in the meeting agenda (just 2 new bugs) | 11:51 |
| rpittau | thanks Mahnoor :) | 12:13 |
| opendevreview | Harald Jensås proposed openstack/networking-generic-switch master: Use segment_id for L2VNI cleanup https://review.opendev.org/c/openstack/networking-generic-switch/+/981509 | 12:44 |
| opendevreview | Harald Jensås proposed openstack/networking-baremetal master: Refactor to use guard clauses and early returns https://review.opendev.org/c/openstack/networking-baremetal/+/981479 | 13:03 |
| opendevreview | Harald Jensås proposed openstack/networking-baremetal master: Simplify conditional logging for link information https://review.opendev.org/c/openstack/networking-baremetal/+/981514 | 13:20 |
| cardoe | frickler: ESP copying? | 13:46 |
| opendevreview | Harald Jensås proposed openstack/networking-baremetal master: Fix ha_chassis_group list access in trunk manager https://review.opendev.org/c/openstack/networking-baremetal/+/981515 | 13:46 |
| TheJulia | cardoe: software raid + ironic + uefi | 13:47 |
| cardoe | Ah | 13:47 |
| frickler | https://review.opendev.org/c/openstack/ironic-python-agent/+/981403 to be more specific | 13:59 |
| TheJulia | I really love the lockout/tagout spec cid, a couple items commented in-line. I feel like there is a huge question regarding do we support multiple concurrent lockouts, or singular. I guess the underlying challenge of the unknown is how we shape the rbac model and the interaction. It feels a little weird that we have the other table if its entirely singular. | 14:10 |
| Glimmr_ | cardoe: I have looked into the tester.py you mentioned & the yaml next to it. Even with that notation it seems to not work for me, as the last error on the nodes details show(>: not supported between instances of dict and int). | 14:28 |
| Glimmr_ | I now start doubting it could be due to my deployment method here(kayobe), as using the example from the repo got registered on the openstack cli with {'inventory.memory.physical': None} | 14:28 |
| cid | TheJulia, thanks I will take a look at your comments. About handling concurrency requests, it will likely depend on how we write the API | 14:31 |
| cid | My understanding is that the bulk actions API will be right endpoint for multiple lockout requests | 14:32 |
| TheJulia | Are you aware of how lockout/tagout might work at a physical data center? | 14:32 |
| cid | I have tried to imagine it, so I do have a working picture in mind :D | 14:33 |
| cardoe | TheJulia: We plop a person in front of the rack and they glare at anyone that tries to touch the box? | 14:33 |
| TheJulia | I'd encourage us to imagine "someone got started" and "someone else went to go join in the fight and physically put their own lock on the tag" | 14:33 |
| TheJulia | cardoe: Are they a blocker and are you playing "DC Lemmings" ? | 14:34 |
| TheJulia | ... It would not be hard to update pingus to swap blocks to servers.... | 14:34 |
| cardoe | I need to read the spec cause this does sound like something interesting for me as well. Something that's currently managed outside of Ironic entirely. | 14:34 |
| cid | Huh, so, multiple locks, one tag. | 14:36 |
| cardoe | Glimmr_: pastebin / share whatever you can to expose the bug. I'm interested in inspection rules but not yet using them. I'm trying to come up with the tester and other validation stuff so that it doesn't feel like sticking your hand into a bag of razor blades with your eyes closed. | 14:36 |
| TheJulia | cid: yup, In one of my comments I put an example tag link :) | 14:36 |
| TheJulia | cid: and yes, its physical locks. | 14:37 |
| TheJulia | :) | 14:37 |
| cid | In the real world, multiple locks on one door mostly spell trouble | 14:38 |
| TheJulia | yes, it sometimes can. | 14:39 |
| cid | Heh. that's new information to how I imagined it would work. So thanks. | 14:40 |
| TheJulia | We're in a world with no locks, just a badge reader and some sort of biometric device next to a door of sorts. | 14:40 |
| TheJulia | Which might scream "trouble" even more, depending on one's point of view | 14:41 |
| cid | TheJulia, do you imagine there's a "someone" who can remove the tag and suddently all locks are gone? A master key of sort! | 14:41 |
| cid | Via RBAC | 14:41 |
| TheJulia | I don't know, typically its everyone must individuallly remove their locks | 14:42 |
| TheJulia | That is sort of like "These people added this tag for the wrong reasons" and without interfacing with the lock appliers, its sort of taking a grinder and cutting the tag off | 14:42 |
| TheJulia | The implications could be disasterous | 14:42 |
| TheJulia | Granted, we're toggling power at BMC level, so one could always lockout/tagout the BMC and then once everything is down actually cut physical power overall | 14:43 |
| TheJulia | (but these are super rare things, having worked in DC ops its like the last resort nightmare option) | 14:44 |
| TheJulia | The kind of nightmare where sometimes additional staff called in would bring booze | 14:44 |
| cid | :D, because lots of work involed. | 14:45 |
| cid | Makes sense actually. No overrides | 14:45 |
| TheJulia | More so, thermal properties and overall start survivability of some devices | 14:46 |
| TheJulia | because if they didn't restart/recover, it would become a very long day/night | 14:46 |
| TheJulia | a concept which might also work is to allow owner-admin or system-admin to remove all locks as well, in that case they have to accept the responsibility | 14:49 |
| TheJulia | but, I don't knwo if we want to make it that complex overall | 14:49 |
| TheJulia | (akin to a grinder being taken to the tag | 14:50 |
| JayF | I can't wait for the horror story from the data center tech who gets nicked by a fan from some power drunk owner admin clearing his lock 😥😥😥😥😥 | 14:51 |
| TheJulia | JayF: but wait, there is more! | 14:53 |
| TheJulia | that bump on the fan blade damanged the balance, and eventually the fan explodes 33.915 hours later. | 14:53 |
| TheJulia | causing reduced airflow and damaged circuits! Suddenly, the mobo catches fire and FLAMES are exiting the rear of the rack, oh noes! | 14:54 |
| TheJulia | (silly combustible cooling fluids) | 14:55 |
| * TheJulia may have had too much coffee already | 14:56 | |
| dtantsur | We have bootc support nowadays, right? and it works automagically through the oci support? if so, are there any public images I could try? | 15:12 |
| TheJulia | oci://quay.io/podman/machine-os:5.4 ? | 15:14 |
| JayF | It's even better than you. Think: a bootc image specified via oci URL on a node with the auto detect deploy interface will deploy via bootc | 15:14 |
| TheJulia | (I just typed that from memory, so it is likely wrong) | 15:14 |
| dtantsur | JayF: ah, so it needs the autodetect interface? noted | 15:15 |
| dtantsur | TheJulia: isn't it a normal OCI image? or did they switch to bootc? | 15:16 |
| JayF | No, we have a specific bootc interface as well | 15:16 |
| JayF | I'm just telling you it's hooked into even more cool magic | 15:16 |
| dtantsur | Yep, cool. That does mean that I cannot just try it in metal3 :) | 15:16 |
| TheJulia | dtantsur: aiui, oci + able to be bootc'ed because the layering is present for a full container | 15:16 |
| TheJulia | which can be bootc'ed | 15:16 |
| dtantsur | I see, thanks | 15:17 |
| TheJulia | they have artifacts *and* a bootc container layered under there | 15:17 |
| * dtantsur is updating metal3 docs with recent additions | 15:17 | |
| JayF | dtantsur: Auto detect deploy interface supports falling back to direct as fallback method | 15:17 |
| dtantsur | Indeed, but we'll need some massaging around API versions | 15:17 |
| TheJulia | I dunno which would get chosen by default since they also have mastered disk images | 15:17 |
| JayF | dtantsur: The entire point of that interface is to allow you to use different types of deploys without having to switch your interface, so if you have a use case that's not covered, we should talk about it | 15:17 |
| TheJulia | worth trying | 15:17 |
| dtantsur | or autodiscover what is available | 15:17 |
| dtantsur | (baremetal-operator is not tied to one Ironic version) | 15:17 |
| * TheJulia takes the corgi for a walk | 15:18 | |
| JayF | (I actually took a day off today and I'm going to jump off IRC now) o/ | 15:18 |
| dtantsur | enjoy! | 15:18 |
| cardoe | "tools/vnc-container/bin/start-x11vnc.sh:33:1: E004 File did not end with a newline" | 15:23 |
| cardoe | wut? I didn't touch that | 15:23 |
| cardoe | I also cannot get that to reproduce locally. | 15:28 |
| cardoe | Do we have a way in IPA to skip dhcp-all-interfaces? Specifically skipping https://opendev.org/openstack/diskimage-builder/src/commit/6b42cb23eef0ca921ca933f354238f83a45229d8/diskimage_builder/elements/dhcp-all-interfaces/install.d/dhcp-all-interfaces.sh#L176 | 15:30 |
| TheJulia | cid: worth posting a patch if CI is kicking it on that error, I bet if you add -r to your tox command line it would list it | 15:38 |
| cid | TheJulia, are you referring to the pickle change? | 15:42 |
| TheJulia | well, your saying it appears to have suddenly appeared | 15:44 |
| TheJulia | so... | 15:44 |
| TheJulia | That is where I'm coming from | 15:44 |
| cardoe | okay. re-running with -r | 16:02 |
| cardoe | But I wanted to get thoughts/feels on https://review.opendev.org/c/openstack/ironic/+/981423 | 16:02 |
| cardoe | I've been having Nidhi work on inspection rules stuff internally with the goal of using it at some point after we upgrade to Ironic 35 | 16:03 |
| cardoe | I want to have confidence in them before deploying them. | 16:04 |
| cardoe | So what I've had her do is dump a node and inventory save and then setup a CI loop across those with the rules. | 16:04 |
| cardoe | So that patch brings that into Zuul. I'd contribute the same node she's used but others can add nodes over time. | 16:05 |
| cardoe | Like Glimmr_'s issue above he could contribute that and we could have testing to ensure that the rules work consistently. | 16:05 |
| cardoe | pep8: OK (203.95=setup[63.17]+cmd[140.78] seconds) congratulations :) (204.41 seconds) | 16:06 |
| cardoe | So nope... tox -r -e pep8 didn't reproduce locally. | 16:06 |
| cardoe | https://zuul.opendev.org/t/openstack/build/c71e0fabb4314949b5107e2de44a9131 | 16:06 |
| cardoe | The file ends with a 0x0a | 16:10 |
| TheJulia | interesting | 16:24 |
| FreemanBoss[m] | Hi everyone. I'm Habeeb Babasulaiman (also FreemanBoss). I am currently an outreachy applicant looking forward to contribute to openstack - ironic on the project: Extend Ironic deployment history api. I'm looking forward to collaborating with and learning from everyone in the community. | 16:38 |
| FreemanBoss[m] | rpittau: hroy: looking forward to your guidance my mentors. Thank you | 16:38 |
| TheJulia | cardoe: we'd likely have to build an overriding element if we wnat to change/influence the behavior with ipa | 17:01 |
| denver_ba | Hi everyone, I'm Denver Baraka, (Outreachy Applicant.) | 17:22 |
| dtantsur | denver_ba: hi, please email the mentors, they're not online right now :) | 17:27 |
| dtantsur | Speaking of which, folks, I'll be at KubeCon EU the entire next week. Expect spotty availability. | 17:28 |
| denver_ba | okay | 17:28 |
| TheJulia | dtantsur: have fun! | 17:36 |
| TheJulia | hmm, another person I need to talk to is at kubecon next week. Wheeeee! | 18:00 |
| lesliem | Hey guys, I'm Leslie Mwendwa. Happy to start contributing | 18:54 |
| TheJulia | Welcome! | 19:18 |
| TheJulia | err | 19:18 |
| * TheJulia sighs | 19:18 | |
| Pointbr8ker | Hi all — I'm David Nwosu, an Outreachy applicant for the Ironic project. I actually applied last year (project wasn't funded then), but I'm back and excited to see it funded this time. Working through the initial tasks now — looking forward to contributing! | 19:25 |
| TheJulia | Welcome! | 19:27 |
| Olamiposi | Hello Everyone, my name is Olamiposi and I'm an outreachy contributor. I hope to make meaningful changes that will help drive the project forward! | 19:50 |
| TheJulia | Welcome! As a note, we're heading into the weekend, most contributors are away until monday :) | 19:55 |
| Olamiposi | Oh okay no problem | 20:12 |
| emmanuel801234 | Hi everyone, I’m Emmanuel, an Outreachy applicant interested in OpenStack Ironic. Happy to be here! | 21:19 |
| kelvinasante[m] | Hello guys, My name Kelvin Asante from Ghana. I am looking forward to contributing to the ironic project | 22:08 |
| chinnisreeaddagalla | Hi everyone! I'm Chinni Sree, an Outreachy applicant for the May 2026 . I'm interested in the "Extend Ironic node deployment history API" project. I have experience with Python and I'm excited to start contributing! I have completed the initial article about Ironic. | 23:08 |
| opendevreview | David Nwosu proposed openstack/ironic master: Separate enrollment documentation by content type https://review.opendev.org/c/openstack/ironic/+/975600 | 23:18 |
| opendevreview | cid proposed openstack/ironic stable/2025.1: Make RPCService picklable for oslo.service https://review.opendev.org/c/openstack/ironic/+/977261 | 23:26 |
| *** emmanuel801234 is now known as e3plexdev | 23:31 | |
| *** e3plexdev is now known as _3plexdev | 23:31 | |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!