| opendevreview | Takashi Kajinami proposed openstack/networking-baremetal master: Declare Python 3.14 support https://review.opendev.org/c/openstack/networking-baremetal/+/1007989 | 00:48 |
|---|---|---|
| opendevreview | Takashi Kajinami proposed openstack/networking-generic-switch master: Declare Python 3.14 support https://review.opendev.org/c/openstack/networking-generic-switch/+/1007990 | 00:50 |
| opendevreview | Verification of a change to openstack/ironic master failed: Use official service type name for cinder https://review.opendev.org/c/openstack/ironic/+/1007590 | 00:59 |
| opendevreview | Bernd Kuespert proposed openstack/python-ironicclient master: Fix filecache on Windows: use threading lock instead of fcntl https://review.opendev.org/c/openstack/python-ironicclient/+/1008017 | 08:28 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic-specs master: BMC Enrollment Steps https://review.opendev.org/c/openstack/ironic-specs/+/1006205 | 10:03 |
| dtantsur | TheJulia, cardoe, a more generalized version ^^ | 10:03 |
| dtantsur | I'm still not convinced we can make it fully generic given Redfish realities | 10:03 |
| Uggla | Hello Ironic team, do we need a PTG X session Ironic/Nova ? | 12:32 |
| opendevreview | Jakub Jelinek proposed openstack/ironic-specs master: Partial fast track with neutron networks https://review.opendev.org/c/openstack/ironic-specs/+/1008056 | 12:55 |
| cardoe | Uggla: not sure. I would love to have a convo about maintainership of the ironic driver and how to increase cadence there. I understand the current core's are not familiar and it doesn't have the priority. The community has been working on performance and other features. | 13:10 |
| cardoe | dtantsur: I’ll give it a glance. | 13:12 |
| TheJulia | Uggla: JayF *might* have a desire, but he should be awake and interacting in the next hour. | 13:54 |
| Uggla | TheJulia, ok thx. We can wait for JayF there is no hurry. | 13:55 |
| TheJulia | cardoe: by chance did you manage to get a patch into nova to make it friendlier to data tyles for volumes, or is it more a matter we need to get ourselves to find the nvme data acceptable itself first? I'm thinking the latter is the blocking reality. | 13:55 |
| TheJulia | err, s/tyles/types/ | 13:55 |
| TheJulia | FWIW, my focus as of recent has been more to set the stage for vendors to reach the same page quickly on the BMC side of things which was the quick needful and possible in advance of the timberland-sig landing edk2 support for vendors to pickup. | 13:57 |
| TheJulia | (recognizing, this is going to take a long time) | 13:57 |
| TheJulia | dtantsur: I'll try to take a look today, I've got a huge document (like 40 pages) that I need to sift through downstream today | 14:00 |
| dtantsur | I have it's a very exciting and completely not AI-written document :) | 14:01 |
| TheJulia | \o/ | 14:01 |
| TheJulia | speaking of not AI, Random question for another thread of discussion: If you were interested in a topic which someone gave a talk on, would you find a high level summary useful and help motivate to watch the video? | 14:02 |
| dtantsur | TheJulia: 100%. In fact, I only watch videos if I want to clarify some details from the summary. If it doesn't have a summary, the bar is very high to interest me. | 14:03 |
| TheJulia | On the topic of Centos, I asked spotz w/r/t centos 9-stream cloud images. Looks like they are still lacking checksum files for everything but azure. She indicated it was known, but noted "should be okay to continue to consume sans the checksum file issue... for now". I'm going to see if I can find the issue in their bug tracking here in a little bit but its starting to feel like we need to move to 10 across the board on builds on | 14:05 |
| TheJulia | master and possibly consider pushing that onto stable branches | 14:05 |
| opendevreview | Riccardo Pittau proposed openstack/bifrost stable/2026.1: [stable only] CI: Fall back to HTTPS when cirros 0.5.3 is not cached https://review.opendev.org/c/openstack/bifrost/+/1008072 | 14:29 |
| JayF | Uggla: We have phase two of the ironic nova performance improvement stuff, but I don't know if there's anything new to talk about there. | 14:30 |
| JayF | Uggla: If anything, I'm more curious and invested in the rollout of the expanded reviewer teams over there | 14:30 |
| TheJulia | w/r/t 9-stream, For those of you not on the -dib channel, I've added a comment to https://redhat.atlassian.net/browse/CS-3532 which appears to report the same issue but for vagrant box images. | 14:31 |
| Uggla | JayF ok. I'm fine not to have a dedicated X session. Regarding nova-reviewer group topic, I'm currently scheduling it. (probably Monday) | 14:33 |
| TheJulia | Uggla: I hope trust can be extended. Going back to what Jay raised, the former still being an issue is more a long running tail of the side-effects related to the latter. | 14:35 |
| TheJulia | (not trying to beat anyone up of course, more so just highlighting its a good example of side effects) | 14:36 |
| TheJulia | (with a 1 line proximity in chat) | 14:36 |
| Uggla | TheJulia at least we will have a discussion about it and hopefully an outcome that will help everyone. | 14:40 |
| TheJulia | \o/ | 14:40 |
| TheJulia | And that is always the first step :) | 14:40 |
| opendevreview | Merged openstack/bifrost stable/2026.2: Update TOX_CONSTRAINTS_FILE for stable/2026.2 https://review.opendev.org/c/openstack/bifrost/+/1006454 | 14:41 |
| cardoe | TheJulia: no I wasn't able to get something to land | 15:13 |
| TheJulia | I guess on a technical side, we need to land something in ironic first, I think we've both proposed changes to that effect somewhere. | 15:19 |
| opendevreview | Jed Preist proposed openstack/bifrost master: Add support for authentication when downloading deployment images https://review.opendev.org/c/openstack/bifrost/+/1008093 | 16:28 |
| opendevreview | Iury Gregory Melo Ferreira proposed openstack/ironic master: [WIP] Aync Sensor Data Collection https://review.opendev.org/c/openstack/ironic/+/1007649 | 17:21 |
| adamcarthur5 | Hey folks, just wanted to check-in and see if there was any planning for times during the PTG we want to have the ironic meetings? | 17:26 |
| TheJulia | adamcarthur5: that is a great question, iury.... I do't see iurygregory | 17:31 |
| iurygregory | o/ | 17:39 |
| TheJulia | iury came back! | 17:39 |
| iurygregory | hello ironic | 17:39 |
| iurygregory | my bad, trying to get used to using two laptops for work =X | 17:40 |
| * TheJulia attempts to resist the urge to channel "plucky duck" | 17:40 | |
| TheJulia | No worries, the question has been raised about ptg planning and time slotting, I guess we need to get our stuff penciled in and begin building our schedule | 17:41 |
| iurygregory | I think we can at least book the hours it used to work for us ? | 17:43 |
| iurygregory | and check the etherpad to schedule move the topics for each day | 17:44 |
| TheJulia | I think that seems like a solid plan | 17:44 |
| iurygregory | but this time we lost Mitaka room | 17:44 |
| iurygregory | so I will be going with Diablo <evil laugh> | 17:45 |
| iurygregory | diablo is booked on other days, so I will go with Liberty =) | 17:54 |
| iurygregory | On Hibiscus PTG did we have topics for all 5 days?! =O | 17:55 |
| iurygregory | I've booked all slots based on the previous PTG o/ | 17:56 |
| opendevreview | cid proposed openstack/ironic master: Extend slash notation support to plugin-data paths https://review.opendev.org/c/openstack/ironic/+/1008109 | 18:09 |
| TheJulia | Evil, for good++ | 18:35 |
| TheJulia | iurygregory: we did, and we’ve got over 300 lines on the etherpad already | 18:36 |
| iurygregory | ack | 18:41 |
| iurygregory | tomorrow I will check the etherpad to see the topics | 18:41 |
| iurygregory | cardoe, we need can you +1 the topics you are interested (so I can avoid scheduling them when there is an overlap with TC slots) | 18:42 |
| cardoe | will do | 18:42 |
| JayF | make sure to please not schedule things, generally, over the TC+community time slots | 18:43 |
| iurygregory | I think it's basically impossible, every PTG we have overlap, I think we handled like in the past (trying to not have discussion on topics that folks that are attending TC meetings want to participate) | 18:56 |
| JayF | there are *TC* sessions, then there are *TC+Community* sessions | 19:04 |
| JayF | I'm not saying to keep the whole TC slot clear; I'm suggesting we specifically ensure that our whole community can participate in the larger sessions | 19:05 |
| TheJulia | When it comes to slotting, I think jay is calling out a need to be mindful | 20:00 |
| JayF | we are, by some metrics, the largest openstack service project in contribution | 20:28 |
| JayF | that means if we schedule over the community time, we're setting the example for the whole project and basically opting outta community-wide stuff, that's no bueno | 20:29 |
| TheJulia | That is a super good point | 20:33 |
| JayF | cardoe: there are no ironic slides and we have the presentation in ... ~16 hours? | 20:47 |
| TheJulia | Eeek | 20:51 |
| TheJulia | Going back to the ptg, would it make sense to do a swag time estimate and build a rough schedule focused on Tuesday/Wednesday/Thursday ? | 20:58 |
| TheJulia | I can likely do that tomorrow once I have regained my spoons stolen by documents today | 21:00 |
| JayF | I think if we got a couple of folks in a call, we could probably knock it out in a hurry. Just need more than one perspective to have confidence about how contentious a discussion might be. | 21:08 |
| cardoe | JayF: yeah I had some personal stuff come up and got taken away from the computer I’ll work on them this evening. | 21:09 |
| JayF | I can draft some stuff up, please make sure to polish them before the morning though | 21:10 |
| TheJulia | I’m all for a quick call to collab to knock them out | 21:11 |
| opendevreview | Steve Baker proposed openstack/ironic-python-agent-builder master: ironic-python-agent-ramdisk-base don't depend on epel https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/1007322 | 21:49 |
| opendevreview | cid proposed openstack/ironic master: Extend slash notation support to plugin-data paths https://review.opendev.org/c/openstack/ironic/+/1008109 | 22:09 |
| opendevreview | cid proposed openstack/ironic master: Normalize inspection rule args without mutating them https://review.opendev.org/c/openstack/ironic/+/1008149 | 22:09 |
| opendevreview | cid proposed openstack/ironic master: Catch inspection rule argument validation failures https://review.opendev.org/c/openstack/ironic/+/1008150 | 22:09 |
| opendevreview | cid proposed openstack/ironic master: Compare the whole value in the matches operator https://review.opendev.org/c/openstack/ironic/+/1008151 | 22:09 |
| opendevreview | cid proposed openstack/ironic master: Report the first iteration's result for multiple: first https://review.opendev.org/c/openstack/ironic/+/1008152 | 22:09 |
| opendevreview | cid proposed openstack/ironic master: Always remove the key unset-plugin-data was given https://review.opendev.org/c/openstack/ironic/+/1008153 | 22:09 |
| cardoe | JayF: I'll go eat some dinner since I just got back tot he desk. | 22:34 |
| cardoe | JayF: I was gonna mention safer security defaults after TLS | 22:34 |
| JayF | just make sure to be specific since we already say spooky security stuff in the by the numbers slide | 22:34 |
| JayF | and I'll have already done (by this point in OIF) the VMT section talking about the overall patterns | 22:35 |
| JayF | I mainly wanted to highlight Ironic OSSAs for (1) credit that we did things and (2) the call to action for operators to read the OSSNs | 22:35 |
| cardoe | The IPMI send_raw step is off, IPA no longer installs bootloaders, agent container registries verify TLS again, and file URL allowed paths are tighter. The API also rejects oversized or deeply nested JSON, and the unauthenticated ramdisk endpoints can be turned off for each API instance. | 22:35 |
| cardoe | Those are my verbal notes. | 22:35 |
| JayF | A suggestion? | 22:35 |
| JayF | Say those when talking about OSSAs/OSSNs | 22:35 |
| JayF | all those are attached to an OSSA/OSSN | 22:35 |
| JayF | we can even restore the security slide if you wanna have text about those on it | 22:36 |
| cardoe | that works. I'll post my full notes then as well. | 22:36 |
| JayF | if you go that route, mention the per-step RBAC + stricter validation of config items (like kernel params) | 22:36 |
| JayF | I'll restore the "separate security slide" model and find some more stats to pad that slide | 22:36 |
| JayF | so you can review/put in your voice / add security stuff as a quick pass | 22:36 |
| cardoe | I wrote talking points while I was out and about today on the phone. | 22:39 |
| cardoe | yeah per-step RBAC is in there along with runbooks. | 22:39 |
| JayF | cardoe: aight, bowing outta the doc. Good luck :) if you want me to look at a thing ping here and I may remember to check :D | 23:00 |
| cardoe | Sounds good | 23:01 |
| cardoe | JayF: so my lead in question is going to be new AI hardware is appearing and how is that shaping user asks of Ironic so I'm gonna change it up a little | 23:24 |
| JayF | I hate that framing | 23:38 |
| JayF | You do you, I just wanted to make sure we had something in place. I'm just glad they aren't asking me to field that question because they might not like the answer. | 23:38 |
| JayF | Hardware is hardware is hardware | 23:39 |
| JayF | We've been provisioning it ever since GPT was more of a SmarterChild | 23:40 |
| cardoe | I'm just the messenger here. | 23:46 |
| TheJulia | heh | 23:56 |
| TheJulia | If it is any conciliation, earlier today I was noting that the desire to make everything the same is not exactly applicable to AI related workloads because of the variety of things people are now doing. We can't put everyone in the same exact box and for them to be happy. Weird/odd/shiny new hardware doesn't help. | 23:58 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!