| JayF | iurygregory: I'll try to remember and fetch it for you when I'm on a work PC tomorrow. I have a query I use for GR-OSS reporting that we can just remove the GR filter from :D | 00:53 |
|---|---|---|
| JayF | https://openstack.biterg.io/app/dashboards#/view/Overview is the front door if you wanna figure it out yourself :D | 00:54 |
| iurygregory | JayF, tyvm! | 01:19 |
| zigo | What's the status of proliantutils ? Is it deprecated / abandonned ? (FYI: https://review.opendev.org/c/x/proliantutils/+/962644 ) | 08:01 |
| zigo | It needs further work for it to be pysnmp 7 compatible though, as I got unit test discovery failing for Ironic if I put it as build-depends. | 08:13 |
| rpittau | good morning ironic! o/ | 08:20 |
| rpittau | zigo: ironic is not pysnmp 7 ready itself, we use pysnmp-lextudio 6.x | 08:23 |
| rpittau | proliantutils is an HP maintained library, but I think it's in the same situation | 08:23 |
| rpittau | zigo https://opendev.org/x/proliantutils/src/branch/master/requirements.txt#L9 | 08:24 |
| zigo | rpittau: I know, I'm writing the compat ! :) | 08:24 |
| zigo | I had moderate success so far. | 08:25 |
| rpittau | yeah, my point is that migrating proliantutils won't really work if ironic is not compatible, they will clash for dependencies, so some kind of retro compatibility is needed, I guess | 08:27 |
| rpittau | also there are multiple tentatives to migrate to pysnmp 7, not very successfull :/ | 08:27 |
| zigo | I've so far made is so it detects if there's pysnmp.hlapi.v3arch.asyncio and act accordingly. | 08:29 |
| zigo | Stuff like: | 08:30 |
| zigo | if importlib.util.find_spec("pysnmp.hlapi.v3arch.asyncio") is not None: | 08:30 |
| zigo | I wonder why they decided to break everyone with that new path though... :/ | 08:31 |
| zigo | Hum... it's harder than I thought because of mocking of pysnmp.hlapi that needs to become pysnmp.hlapi.v3arch.asyncio.cmdgen :/ | 08:56 |
| zigo | Is it possible to make conditional @mock.object.patch ? | 08:56 |
| zigo | Saw it is. | 09:55 |
| *** tkajinam_ is now known as tkajinam | 10:54 | |
| iurygregory | good morning ironic | 10:57 |
| opendevreview | Verification of a change to openstack/ironic stable/2025.1 failed: Fix cache image hard link between different file systems https://review.opendev.org/c/openstack/ironic/+/962328 | 11:21 |
| dtantsur | Folks, our docs say: is_root_volume - Set to true if this is the root volume. At most one logical disk can have this set to true; the other logical disks must have this set to false. The root device hint will be saved, if the RAID interface is capable of retrieving it. | 12:23 |
| dtantsur | Are there any RAID implementation that actually do the root device hint part? It does not look like hardware RAID does it. | 12:23 |
| kubajj | dtantsur: nope, it does not do anything. Somebody was looking into it a couple of weeks ago | 12:24 |
| dtantsur | Not even for software RAID? | 12:24 |
| dtantsur | If not, I'll remove it from examples in our docs since people get confused | 12:25 |
| kubajj | It was drannou, on the August 13th (was scrolling in the IRC log | 12:28 |
| kubajj | dtantsur: yep, not even for software RAID | 12:29 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: Docs: play down the role of is_root_volume in RAID https://review.opendev.org/c/openstack/ironic/+/962670 | 12:34 |
| dtantsur | kubajj: looks about right ^^? | 12:34 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: Docs: play down the role of is_root_volume in RAID https://review.opendev.org/c/openstack/ironic/+/962670 | 12:35 |
| opendevreview | Mathieu Parent proposed openstack/ironic-python-agent-builder master: Force ifupdown on Debian when simple-init is used https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/962672 | 12:40 |
| kubajj | dtantsur: I actually never looked into this on the Ironic side. Will have a look deeper. Is it possible that this configures where the EFI partition lands? | 12:45 |
| kubajj | mostepha[m]: this might be interesting to investigate | 12:45 |
| opendevreview | Mathieu Parent proposed openstack/ironic-python-agent-builder master: Wait up to 30 seconds for config drive https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/962673 | 12:45 |
| dtantsur | kubajj: I don't see any mentions of is_root_volume in the entire IPA code base, and the docs say that software RAID does not support it | 12:45 |
| dtantsur | I think the root device is always determined through root device hints (and that's where the EFI partition goes) | 12:46 |
| kubajj | dtantsur: yep, sounds right. But can you have a hint pointing to an array? I don't think that's possible now | 12:47 |
| TheJulia | so I think you can, we've had customers set such hints, albeit to the device which is not exactly stable | 12:54 |
| * dtantsur nods | 12:55 | |
| dtantsur | It's a known pain point | 12:55 |
| TheJulia | yeah | 12:58 |
| opendevreview | Allain Legacy proposed openstack/bifrost master: Add support for OVS as a virtual switch https://review.opendev.org/c/openstack/bifrost/+/962038 | 12:59 |
| opendevreview | Allain Legacy proposed openstack/bifrost master: Add support for standalone ironic networking https://review.opendev.org/c/openstack/bifrost/+/962394 | 12:59 |
| *** darmach47 is now known as darmach | 13:01 | |
| TheJulia | Good morning | 13:02 |
| dtantsur | morning TheJulia | 13:02 |
| iurygregory | TheJulia dtantsur rpittau JayF cardoe seems like we have almost 18hrs for topics (considering 30min for the ones without time)... | 13:17 |
| iurygregory | in the past we booked 12hrs only =) | 13:17 |
| TheJulia | So those times are all ballpark estimates | 13:19 |
| TheJulia | At least from my point of view, It might help for someone else to look at them and take a guess as well ? | 13:20 |
| rpittau | if memory serves, we usually overestimate :) | 13:20 |
| TheJulia | Yeah, async ironic at 1hr is likely a bit much as well. :) | 13:20 |
| TheJulia | vxlan, I suspect will need at least 45 if we can get a cross-project session and traction | 13:21 |
| TheJulia | I could also see graphical console dialing back to like a third of what the time I put on there | 13:22 |
| opendevreview | Jacob Anders proposed openstack/ironic master: [WIP] Make cache_firmware_components more resilient during upgrades https://review.opendev.org/c/openstack/ironic/+/960711 | 13:27 |
| iurygregory | yeah, I do remember we normally overestimate :D | 13:28 |
| iurygregory | for cross project I was thinking in 1h, this would totally depend on the number of folks that would have input in the topic... | 13:29 |
| iurygregory | I will book the normal 12hrs first, and check again the topics | 13:29 |
| iurygregory | we will be using meetpad right? | 13:30 |
| opendevreview | Jacob Anders proposed openstack/ironic master: [WIP] Make cache_firmware_components more resilient during upgrades https://review.opendev.org/c/openstack/ironic/+/960711 | 13:32 |
| kubajj | dtantsur: TheJulia: ah, makes sense. We might have a look into implementing is_root_volume and some nicer hints for RAID (but probably only after the summit) | 13:33 |
| opendevreview | Ettore Simone proposed openstack/ironic master: Allow multiple HTTP redirects for image source https://review.opendev.org/c/openstack/ironic/+/962676 | 13:35 |
| TheJulia | Meetpad makes sense to me | 13:42 |
| opendevreview | Julia Kreger proposed openstack/ironic master: fix idrac9 version handling in redfish vmedia boot interface https://review.opendev.org/c/openstack/ironic/+/962678 | 13:46 |
| opendevreview | nidhi proposed openstack/sushy master: Add comprehensive PCIe function support to Sushy https://review.opendev.org/c/openstack/sushy/+/962680 | 13:50 |
| dtantsur | kubajj: that would be great! I'd still prefer to update the docs for now: I started this conversation after meeting a confused operator. | 13:52 |
| kubajj | dtantsur: agree, left a comment but otherwise looks good | 13:53 |
| dtantsur | ah, oh. Fixing now | 13:56 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: Docs: play down the role of is_root_volume in RAID https://review.opendev.org/c/openstack/ironic/+/962670 | 13:57 |
| opendevreview | Mathieu Parent proposed openstack/ironic-python-agent-builder master: Force ifupdown on Debian when simple-init is used https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/962672 | 14:18 |
| opendevreview | Verification of a change to openstack/ironic stable/2025.1 failed: Fix cache image hard link between different file systems https://review.opendev.org/c/openstack/ironic/+/962328 | 14:23 |
| TheJulia | alegacy_: https://opendev.org/openstack/requirements | 14:34 |
| TheJulia | dtantsur: you around and have 10 minutes? | 14:37 |
| dtantsur | TheJulia: in meetings until at least 1530 UTC | 14:38 |
| opendevreview | Mathieu Parent proposed openstack/ironic-python-agent-builder master: Wait up to 30 seconds for config drive https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/962673 | 14:40 |
| TheJulia | dtantsur: so, talking with alegacy_, seems like he is going to move dynamic switch config reloading to a separate patch because the code presently adds a new python dependency | 14:43 |
| TheJulia | so, a middle ground and we can figure it out as we go | 14:45 |
| dtantsur | TheJulia: alegacy_, in Metal3, we do some ugly things with inotify to reload certificates. With the new ironic operator though, it will restart the pods itself. | 14:46 |
| dtantsur | So if the question is "am I fine with delaying this aspect", I totally am | 14:47 |
| rpittau | dtantsur: we switched to ugly watchmedo :D | 14:47 |
| dtantsur | it's probably still inotify as a mechanism? | 14:48 |
| TheJulia | yeah, folder watching | 14:48 |
| rpittau | oh yeah | 14:48 |
| opendevreview | Takashi Kajinami proposed openstack/ironic master: Drop explicit description about default values https://review.opendev.org/c/openstack/ironic/+/962690 | 15:21 |
| alegacy_ | TheJulia: FYI... global-requirements change: https://review.opendev.org/c/openstack/requirements/+/962691 | 15:24 |
| opendevreview | Mathieu Parent proposed openstack/ironic-python-agent-builder master: Wait up to 30 seconds for config drive https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/962673 | 15:32 |
| opendevreview | Verification of a change to openstack/ironic stable/2025.1 failed: Fix cache image hard link between different file systems https://review.opendev.org/c/openstack/ironic/+/962328 | 16:07 |
| iurygregory | TheJulia, JayF cardoe any of you have some info regarding this 3 working items https://paste.opendev.org/show/bHv5a2tubKKgvpzGCo8H/ I'm working on the slides for the OpenInfra Live tomorrow https://docs.google.com/presentation/d/1cPcNJNgltj2jx-ddje37IXUaktomGIR0UKc7JR8zLY4/edit?usp=sharing so I wanted to have some details about it | 16:08 |
| JayF | iurygregory: eventlet, plus billboard retirements of inspector + metalsmith | 16:09 |
| iurygregory | JayF, did you find the link from https://openstack.biterg.io/app/dashboards#/view/Overview ? I was only able to retrieve number of commits and number of contributors.. | 16:09 |
| JayF | iurygregory: I'm not sure I'd report on the future looking stuff | 16:09 |
| JayF | iurygregory: what specifically do you want | 16:09 |
| iurygregory | # Lines of code and # bugs fixed | 16:10 |
| iurygregory | oh right, we should mention inspector retirement and metalsmith | 16:10 |
| iurygregory | ironic-lib also was this cycle if I recall | 16:10 |
| JayF | LOC and number of bugs is not in bitergio | 16:16 |
| JayF | you'd have to query launchpad and/or git | 16:16 |
| iurygregory | ohhhh | 16:16 |
| JayF | https://openstack.biterg.io/app/dashboards#/view/95487340-6762-11e9-a198-67126215b112?_g=(filters:!(),refreshInterval:(pause:!t,value:0),time:(from:now-6M,to:now))&_a=(description:'Gerrit%20Approvals%20panel%20by%20Bitergia',filters:!(('$state':(store:appState),meta:(alias:Bots,disabled:!f,index:gerrit,key:author_bot,negate:!t,params:(query:!t),type:phrase),query:(match:(author_bot:(query:!t,type:phrase)))),('$state':(store:ap | 16:17 |
| JayF | pState),meta:(alias:!n,disabled:!f,index:gerrit,key:author_org_name,negate:!f,params:(query:G-Research),type:phrase),query:(match_phrase:(author_org_name:G-Research)))),fullScreenMode:!f,options:(darkTheme:!f,hidePanelTitles:!f,useMargins:!t),query:(language:lucene,query:''),timeRestore:!f,title:'Gerrit%20Approvals',viewMode:view) | 16:17 |
| TheJulia | iurygregory: that deck says cinder under your name :) | 16:17 |
| JayF | ugh I wanted it to let me make it a .txt | 16:17 |
| iurygregory | TheJulia, what?! :O which slide | 16:18 |
| iurygregory | Slide 19 has Ironic and my name o.o | 16:19 |
| dtantsur | Flamingo has flown, folks, time for https://review.opendev.org/c/openstack/bifrost/+/887934? :) | 16:19 |
| JayF | iurygregory: https://tinyurl.com/ironic-gerrit-stats is one I use for gerrit-oriented stats, that's the only intersting thing that's not obvious from the initial dash imo | 16:19 |
| TheJulia | iurygregory: slide 1 :) | 16:19 |
| iurygregory | OH JESUS | 16:20 |
| * iurygregory asks Alisson to update | 16:20 | |
| JayF | iurygregory: some other things maybe to mention: automated cleaning via runbook, redfish clock validation/setting, api-call inspection rule, recording steps executed in node history | 16:25 |
| iurygregory | automated cleaning I've added | 16:25 |
| JayF | I really really would steer clear of the forward-looking networking stuff except maybe to invite folks to PTG to help discuss it | 16:26 |
| iurygregory | steps in node history is also present | 16:26 |
| TheJulia | I think the steps bit is already covered, but its framed as "Document..." | 16:26 |
| TheJulia | That is likely the wrong word leading in | 16:26 |
| TheJulia | I'd *totally* enumerate some high level topics which will be at the PTG, creating the "hook" | 16:27 |
| iurygregory | I was using the info from the work items :D | 16:27 |
| JayF | maybe OSSA-2025-001 should be mentioned too | 16:27 |
| JayF | (file:// url handling cleanups) | 16:27 |
| iurygregory | https://security.openstack.org/ossa/OSSA-2025-001.html noted | 16:27 |
| JayF | I'm glad you asked this iurygregory, reading the release notes made me feel a bit better about achievements. I think my memory wiped during my sabbatical lol | 16:28 |
| iurygregory | I think I will remove the Slide with work Items, and have the info in Highlights | 16:30 |
| jizaymes | hello Team. First thanks, and congrats on the Flamingo release. | 16:31 |
| iurygregory | JayF, api-call inspection rule would fit in Enhanced inspection in Redfish right? | 16:31 |
| TheJulia | iurygregory: mind if I revise the networking bits? | 16:32 |
| JayF | yeah probably so | 16:32 |
| jizaymes | I see release notes about ironic-inspector being removed, but I also see other chatter and documentation about how ipa is the future, so I'd like to understand more clearly what is the 'best' path for inspection these days | 16:32 |
| iurygregory | TheJulia, I do need some help on that part :D | 16:32 |
| iurygregory | feel free | 16:32 |
| JayF | jizaymes: so you're sorta talking about apples and oranges, let me try and clear things up | 16:35 |
| JayF | jizaymes: IPA is Ironic Python Agent. Almost every function in-band on a node uses this to do it. That would include deployment, cleaning, and (in-band/agent) inspection | 16:35 |
| JayF | jizaymes: there used to be ironic-inspector, an external API service to handle inspection only. Now, you only need Ironic's API and it is (more or less) intercompatible | 16:35 |
| jizaymes | I see. so the change is more functional on using ironic, and not so much the mechanism for which inspection is performed against a host? | 16:36 |
| JayF | jizaymes: so the ideal case for a former inspector user AIUI is to use inspect_interface of agent | 16:36 |
| JayF | jizaymes: ideally, the end result is you run one less service (ironic-inspector) and get the same functionality. Integrating inspector into Ironic also gives us more opportunities to do interesting things, e.g. with inspector rules | 16:37 |
| jizaymes | thank you. I will use the inspect_interface of agent. | 16:37 |
| jizaymes | next question! how long should inspection typically take on a modern and empty server with a single drive and no existing OS. It seems like it should be minutes, and I keep getting timeouts, and I'm not sure if its because of lack of being able to hit the https callback endpoint, or it not actually doing any inspection and just sitting there ;) | 16:38 |
| iurygregory | JayF, would be ok using https://qrcodecreator.com/ to generate ? | 16:39 |
| jizaymes | it seems Ironic is the one of the final bosses on my openstack journey | 16:39 |
| JayF | probably iurygregory :) | 16:43 |
| TheJulia | iurygregory: some revisions made, w/r/t mentioning OpenAPI, I suspect we should just drop it, its not 100% completed yet | 16:44 |
| TheJulia | iurygregory: I'm thinking "concise" and "specific" is the key | 16:45 |
| iurygregory | TheJulia, I was planing to mention we did some good progress | 16:47 |
| TheJulia | I guess I'm trying to get us to avoid ambiguity because we really have a short window to convey nuanced context | 16:48 |
| TheJulia | *shrug* | 16:49 |
| JayF | OpenAPI is also not operator-facing | 16:50 |
| JayF | even when Ironic is "done" it's barely operator facing | 16:51 |
| iurygregory | ack | 16:51 |
| TheJulia | ++ | 16:57 |
| TheJulia | yeah, Operator and end Impact for the Operator/Users of Ironic are key | 16:57 |
| opendevreview | Merged openstack/ironic stable/2025.1: Fix cache image hard link between different file systems https://review.opendev.org/c/openstack/ironic/+/962328 | 17:25 |
| opendevreview | Allain Legacy proposed openstack/ironic-specs master: Update standalone networking initial conditions https://review.opendev.org/c/openstack/ironic-specs/+/960550 | 17:48 |
| alegacy_ | TheJulia: dtantsur: JayF: After a discussion with TheJulia I've changed my approach to how to restore the inspection vlan which i'm now referring to the "idle" vlan. spec has been updated ^^^ | 17:57 |
| alegacy_ | dtantsur: this will remove the need to mess with the delete and cleanup actions. | 17:57 |
| alegacy_ | so i'll delete that change request. | 17:57 |
| opendevreview | cid proposed openstack/ironic master: WIP: Asynio prototype for PTG https://review.opendev.org/c/openstack/ironic/+/962708 | 18:21 |
| JayF | What's the best way to test a change like ^ that to the sensor data loop | 18:26 |
| TheJulia | alegacy_: Thanks | 18:27 |
| JayF | My suggestion was "do the power status loop first because it's easier" but that's a really crummy answer | 18:27 |
| TheJulia | ... mmmmm hmmmm | 18:29 |
| TheJulia | I mean, we could wire it up with sushy-tools to query for stats, I'm not sure how much surface it will generate *but* it would at least generate $something | 18:32 |
| JayF | I think the short term plan to get any data at all is to hook the patches up, force default to "on and do it async" for sensor data sending | 18:35 |
| JayF | and see if it goes kaboom | 18:35 |
| JayF | we can't verify it works this way, but we can potentially solicit angry feedback from ci | 18:35 |
| TheJulia | cardoe: did you ever create a bug for the iqn -> ??? change in nova or cinder just self correcting connector data? | 18:44 |
| TheJulia | Yeah, that is doable in theory from redfish | 18:44 |
| TheJulia | I don't think vbmc and all will return any data | 18:44 |
| TheJulia | stevebaker[m]: I replied back to https://review.opendev.org/c/openstack/ironic/+/961434.. I checked choices and realized it would be a super bad idea | 19:27 |
| iurygregory | JayF, wouldn't we need sushy supporting async first? | 20:20 |
| JayF | cid has a patch for that he's going to line up to depend-on, but I think he was cleaning up unit tests first | 20:20 |
| JayF | before pushing it there | 20:20 |
| JayF | (he pushed the ironic one for our 1:1 so we could review) | 20:20 |
| iurygregory | I can probably help with it, this is one of my top priorities for the Gazpacho cycle =) | 20:21 |
| iurygregory | I'm going to the gym now, but I will try to provide feedback | 20:22 |
| JayF | he might be EOD already, I know his tz isn't identical to mine, but coordinate with him :) | 20:22 |
| iurygregory | sure | 20:23 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/2025.1: Increase memory allocation for ironic-base https://review.opendev.org/c/openstack/ironic/+/962723 | 20:37 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/2024.2: Increase memory allocation for ironic-base https://review.opendev.org/c/openstack/ironic/+/962724 | 20:37 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/2024.1: Increase memory allocation for ironic-base https://review.opendev.org/c/openstack/ironic/+/962725 | 20:38 |
| TheJulia | I think ^ should help with other backport jobs failing in CI | 20:38 |
| TheJulia | We may want to consider ipa-b backports as well | 20:40 |
| TheJulia | https://storage.bhs.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_c7b/openstack/c7b3e39aa0754ccabead69be4fe6f9c1/controller/logs/ironic-bm-logs/node-0_console_2025-09-02-17%3A31%3A50_log.txt | 20:42 |
| cardoe | TheJulia: err ummm errrrr look at that crazy person behind the curtain and certainly not over here! | 21:45 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!