Wednesday, 2026-03-25

*** mdfr8 is now known as mdfr00:09
opendevreviewIury Gregory Melo Ferreira proposed openstack/ironic master: Ensure NIC firmware version is cached  https://review.opendev.org/c/openstack/ironic/+/98203802:56
opendevreviewAbhishek Bongale proposed openstack/ironic-tempest-plugin master: Add conditional graphical console validation  https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/98207711:59
*** dking is now known as Guest585712:08
*** Guest5857 is now known as dking12:19
opendevreviewAbhishek Bongale proposed openstack/ironic-tempest-plugin master: Add conditional graphical console validation  https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/98207712:26
clifgm o/14:19
JayFo/14:23
cardoeJayF: so for the nova/ironic slowness... do they have a goal to land the fixes in nova?14:27
cardoeHow do folks feel about adjusting the UX of the CLI? Like https://review.opendev.org/c/openstack/python-ironicclient/+/981805 ?14:41
JayFcardoe: we have a PTG session to chat about it, my personal goal is to have a solid plan in place. Sean's changes, so far, are mainly vibe coded pointers to things that need to be fixed. They need review and real testing. I suspect they are only a small % of the overall improvement we can gain14:41
zigoHi team!14:41
zigoBuilding Ironic 35.0.0 in Debian Experimental, I get:14:41
zigohttps://paste.opendev.org/show/bjgJNSlqrX4cjZqUgrpi/14:41
zigoDoes this ring a bell to someone?14:42
JayFcardoe: *plan is to have a solid plan in place by the end of "H"14:42
JayFzigo: py hlapi or snmp version mismatch14:42
opendevreviewMahnoor Asghar proposed openstack/ironic master: Soft power off before ejecting virtual media regardless of CONF.inspector.power_off  https://review.opendev.org/c/openstack/ironic/+/98194214:42
JayFzigo: that code is unlikely to work even if imported, I'd suggest patching out the irmc+snmp driver14:42
JayFWe are removing it this cycle14:42
JayFwait, it's this cycle!!!!!!!14:43
* JayF gets out the code-machete14:43
zigoJayF: Ok, so I can just blacklist the 2 failing tests, then?14:44
JayFI would ensure the service starts up :) (I think in default config it would), but yes 14:44
JayFI think some of that keys on if a package is installed14:44
zigoThanks, that's enough for me! :P14:45
zigoFYI, we're on pysnmp 7.1.22 in Debian ATM.14:45
TheJuliacardoe: I could be good with that, I guess14:48
cardoeI've also wanted to alias "avail" to "provide" but I'm not necessarily as convinced.14:49
JayFcardoe: I am +1 to adding the "correct" ordering. I am -1 to deprecating the old way14:50
JayFit breaks scripts for .... no real good reason afaict?14:50
TheJuliaqueue "why not both" gif ?14:56
opendevreviewMerged openstack/python-ironicclient master: Add type annotations to `ironicclient/v1/portgroup.py`  https://review.opendev.org/c/openstack/python-ironicclient/+/98166114:58
opendevreviewMerged openstack/python-ironicclient master: Add type annotations to `ironicclient/v1/resource_fields.py`  https://review.opendev.org/c/openstack/python-ironicclient/+/98166214:58
fricklerJayF: oh, yes, please, that's essentially the only remaining u-c pin that we have, getting rid of that would be very cool https://review.opendev.org/c/openstack/requirements/+/97755115:16
JayFfrickler: yeah, I know, and I've felt bad about it that's why we're killing the driver15:17
JayFjust had to have the deprecation timer run out15:17
opendevreviewMahnoor Asghar proposed openstack/ironic master: Soft power off before ejecting virtual media regardless of CONF.inspector.power_off  https://review.opendev.org/c/openstack/ironic/+/98194215:45
JayFit's christmas in claude-code! Although a lot more red decorations than green ones :P 15:52
* JayF is using it to whack the pysnmp dep15:52
TheJuliaoh, my.15:55
opendevreviewMerged openstack/bifrost stable/2026.1: Update .gitreview for stable/2026.1  https://review.opendev.org/c/openstack/bifrost/+/98126815:56
opendevreviewMerged openstack/bifrost stable/2026.1: Update TOX_CONSTRAINTS_FILE for stable/2026.1  https://review.opendev.org/c/openstack/bifrost/+/98126915:56
JayFwe should mention deprecations in oif live, the big one and the ironic one 15:56
* JayF puts that here so it helps him remember15:57
TheJuliathe general session or the ironicy one?16:00
JayFyes16:00
TheJuliaack ack16:00
TheJuliaclif: w/r/t https://docs.google.com/presentation/d/1SDNGtEYIP45AkjMxAiyBfTjJDO4yQQcjaDMjosm17aM/edit?slide=id.g3d7b9bf4b0b_1_36#slide=id.g3d7b9bf4b0b_1_36 you might want to split in to 2-3 slides. Try and start with the basic challenge which exists so we can convey across what we're actually solving. Also, if you see slide 5, its a blank slide, I'd honestly try to keep a similar format just so there is a clean handoff and 16:02
TheJuliawe're also making the top level mental impression of "oh, there is a thing, and this person is going to speak about it.16:02
clifTheJulia: Sure, that works. I'll split it up and elaborate some more.16:03
cardoeJayF: gimme a better way to convey why the docs will be updated to the new way?16:05
JayF"This usage pattern is more consistent with the CLI; even though the old method is remaining for compatability"16:06
cardoeOkay. I didn’t intend for us to remove them. Just not use them in docs and examples16:07
cardoeI’ll change the message16:07
JayFdeprecated implies it'll be removed one day; I think if you just make a bigger comment it'd do the trick16:07
JayFI'm glad we agree just use different language16:08
TheJuliajay, stared a slide on the big deck16:09
JayFack16:09
JayFwdyt about the other ones already in place16:10
TheJuliaJayF: who is that directed to16:13
JayFyou, re: the big deck ironic stuff I put in16:13
JayFfor the original oi live16:13
TheJuliaI don't think I have that link16:18
gtfrans2reHello, I've task #3 on bifrost installation and usage. Anyone facing issues with it?16:19
gtfrans2reOr anythhing else I can help with?16:20
JayitaPramanik[m]gtfrans2re: what issue you are facing?16:36
JayitaPramanik[m]gtfrans2re: I also started exploring the documentation, what are the issues you are facing?16:38
TheJuliaThat is the critical question to ask and information to exchange. Maintainers and regular condtributors might not ever see the kind of issue someone who is coming into a project with a "fresh" perspective, so if you can paint the picture for us to understand, its a lot easier for us to help or provide insight.16:42
JayitaPramanik[m]Hi Ironic mentors!... (full message at <https://matrix.org/oftc/media/v1/media/download/AfOg5mm4OQMx6TFoJXC4Wl1M8cWGFf71VtaNI6JM4rdh-lY0GpQTvXcBk8NPxreT8YwrFtaGVAKZ74lxqwNfHZpCedbgc68gAG1hdHJpeC5vcmcvek1YUWtKVENRdHN4SXBpYWVPSGNsQW5m>)16:57
JayitaPramanik[m]* Hi Ironic mentors!... (full message at <https://matrix.org/oftc/media/v1/media/download/AW8rzs3W17sisinyalLLbm615tUU1ziFbCwUay7PEdC2uZkjC9DLNbKrhUK21OTvr1LHkYPOfHSnRMrp5PNa0FtCedbgxM0AAG1hdHJpeC5vcmcvSXdYdUpDWFpWZXNpaHpHeGtEdVNZVERi>)17:03
LizzyHi everyone I am Lizzy, I have recently started working with Bifrost and learning more about OpenStack Ironic. Looking forward to learning from you all and contributing where I can.17:12
Priestess[m]Lizzy: Welcome, Lizzy17:20
LizzyThank you for the warm welcome @Priestess17:22
cardoeI've passed https://bugs.launchpad.net/ironic/+bug/2076265 back over to Nova because before we can change anything, they need to change. I've proposed a fix on their side but have gotten zero feedback.17:33
opendevreviewJay Faulkner proposed openstack/ironic master: Remove the iRMC hardware type and all interfaces  https://review.opendev.org/c/openstack/ironic/+/98214519:17
opendevreviewJay Faulkner proposed openstack/ironic master: Remove the SNMP hardware type and power interface  https://review.opendev.org/c/openstack/ironic/+/98214619:17
opendevreviewJay Faulkner proposed openstack/ironic master: Remove pysnmp and related dependencies  https://review.opendev.org/c/openstack/ironic/+/98214719:17
JayFdon't review those19:17
TheJuliatoooooo late ;)19:20
TheJuliaOne thing stuck out19:21
JayFI know unit tests are failing on the first commit19:21
JayFbut I can't figure out why so I wanted to look in gerrit lol19:21
JayFit's like I broke driver loading somehow19:21
TheJuliaschweet19:46
clifin launchpad is it possible to add me as an ironic member? I guess this goes for any other related project as well? or do I have to be core in order to have that level of access?19:58
TheJuliait should totally be doable20:02
TheJuliauhhhh20:02
TheJuliahttps://launchpad.net/~ironic-bugs right ?20:03
clifI suppose?20:03
clifI joined the team20:04
TheJuliaokay :)20:06
TheJuliaif that doesn't give you what you need, just let me know20:06
clifrgr20:06
cardoeclif: you know the initiation rule right? You need to personally fix 10 bugs in 7 days.20:07
TheJuliacardoe: wow, your evil.20:07
TheJulia;)20:07
clifez, just merge 10 bugs and then fix them right after20:07
TheJulialol20:08
cardoeIt's been one of those days.20:08
cardoeclif: this is the way20:08
TheJuliacardoe: it is what it is some days20:09
JayFyou know, clif brings up a reasonable question20:12
clifcardoe: are you actively working this issue? https://bugs.launchpad.net/ironic/+bug/2146355 would you assign yourself if so20:12
JayFgiven we no longer have a dashboard that properly identifies review quantities20:12
JayFhow do we manage the core group?20:12
JayFlfx insights is not useful for managing review quantities and it's all we've got20:12
cardoeclif: done.20:13
cardoeAre folks okay with my proposed fix?20:13
cardoeAnd can I term this as a backportable fix so that I can have it in 2026.1?20:13
JayFso right now I'm asking clif to review enough to be a -reviewer and ... I can't even judge if he's on the right track with either a manager *or* DPL hat on20:13
clifIt sounds reasonable to me cardoe, but I don't have a lot of context on the redfish driver20:13
cardoeclif: well if the BMC clock is too far off SSL stuff doesn't work.20:14
TheJuliaoh wow, I'm a jerk according to: https://www.stackalytics.io/?release=all&module=ironic-group20:14
clifcardoe: yea that makes sense20:14
clifI'm kind of surprised there's no way to tell it to contact a time server20:15
TheJuliaJayF: and that dashboard has data from today.20:15
JayFTheJulia: can I trust that data? It's unofficial and it's unclear to me how to get affiliation updates merged there. 20:15
JayFTheJulia: last I checked, that was crediting some nontrivial % of GR work to others20:16
JayFYeah, CID is not tracked as GR in that dash.20:16
TheJuliais there an entry in the stackalytics affiliation config for CID?20:17
JayFWhere does that live now? I've updated it in 3 different places over my time here, and I directed him to update it as well20:17
TheJuliagreat questions20:18
JayFI'm going to email the list20:18
JayFhonestly still leftover-annoyed from just everyone being broken on short notice20:18
JayFand nobody giving a damn that it's still broken months later20:18
ciderm,, nice we have a new dashboard. 20:22
cidAnd I'm not on it :facepalm: 20:23
TheJuliahttps://opendev.org/x/stackalytics/src/branch/master/etc/default_data.json20:23
cidJayF, this is what we were configuring for the other time?20:23
cidMight I have missed something?20:23
JayFcid: this is the unofficial dashboard everyone is using now because the official one is more broken than the unofficial one20:23
cid:D20:24
JayFcid: if you can make sure that file linked by julia is up to date for you, it'd be nice. I think you already proposed a change once, it might have not been merged?20:24
cidLet me check20:24
cardoeclif: you can tell the BMC to contact a time server... but you gotta be able to connect to the BMC to tell it to contact the timeserver... the only command Ironic ignores the SSL errors on is to set the BMC clock.20:56
cardoeI want to have a runbook for our operators to clear BMC jobs and set the clock back. And I wanna have that runbook ultimately be part of the automated cleaning via runbook stuff.20:58
cardoeRight now clear BMC jobs and set clock is "stuff done outside of Ironic"20:58
cardoeAt least the way I've envisioned the runbooks API (and others can correct me if I'm wrong) is that it's a set of operations that are packaged up that's safe-ish for a DC tech todo or potentially a lessee to do on the box.20:59
TheJuliaClearly, presentation AI art is needed to remind ourselves that "bmcs are evil"21:00
cardoeSo just thinking of all the out-of-ironic operations we do and seeing how we can implement them as runbooks.21:00
TheJulia+++++21:00
TheJuliacardoe: Hi, is there an online whiskey store gift card I can get you for that?21:01
cardoeheh21:01
TheJuliaSeriously, I mentioned that people "do physical fleet management activities with ironic" on a call today and a bunch of people just looked at me like I had grown an extra set of eyes and more arms21:01
TheJulia(in other words, more examples == better)21:04
cardoeWe just got several cabs of hardware and my instructions to the person pulling them in was to write down every time you have to do something outside of the baremetal CLI command. And every time something seems awkward.21:04
cardoeShe had some pretty interesting feedback as a hot take.21:05
cardoeThe "baremetal node add trait" being different than other commands was one.21:05
cardoe"baremetal node provide" grammatically doesn't match the other verbs was another.21:06
cidre'stacklytics... took a bit of time trying to understand the ordering, figured it's by launchpad id.21:06
cid  I have patch up now at https://review.opendev.org/c/x/stackalytics/+/98216721:07
TheJuliacardoe: so her "not wrong" column I take it now has at least two checks ?21:07
cardoeYeah I mean she's not wrong.21:07
cardoeAll the other commands are "baremetal node <thing> add" except for trait.21:08
cardoeAnd stuff like "baremetal node manage -> manageable" or "baremetal node rescue -> rescue" or "baremetal node enroll -> enroll" but then "baremetal node provide -> available"21:09
JayFcardoe: the idea there is verb->noun21:39
JayFcardoe: the act of making something available is to provide is21:40
JayF*it21:40
JayFcardoe: I'm not saying it's good; I'm saying it follows a pattern. rescue/enroll are cleaner because they are both noun/verb depending on usage21:40
cardoeyeah makes sense.21:41
opendevreviewJulia Kreger proposed openstack/ironic-python-agent master: Update auto-tls cert signing to SHA-512  https://review.opendev.org/c/openstack/ironic-python-agent/+/98217322:04
opendevreviewJulia Kreger proposed openstack/ironic-python-agent master: WIP: Enforce TLS 1.2 minimum across all connections  https://review.opendev.org/c/openstack/ironic-python-agent/+/98217422:04
opendevreviewJulia Kreger proposed openstack/ironic-python-agent master: WIP: Upgrade TLS certificates to P-384 curve with configurable support  https://review.opendev.org/c/openstack/ironic-python-agent/+/98217522:04
* TheJulia gets out the CI popcorn22:04
JayFare we really spending time on this post-quantum stuff?22:04
TheJuliaeh, we've got some bad defaults in the agent, so makes sense to do some changes and provide some knobs22:05
JayFI only bristle because of the push at last PTG by a security person at redhat who had no interest in helping with active-on-fire security stuff22:05
JayFit feels like rewarding bad behavior to go along with that push (not that I would ever -1 or -2 out of spite)22:06
opendevreviewJay Faulkner proposed openstack/ironic master: Remove the SNMP hardware type and power interface  https://review.opendev.org/c/openstack/ironic/+/98214622:08
opendevreviewJay Faulkner proposed openstack/ironic master: Remove pysnmp and related dependencies  https://review.opendev.org/c/openstack/ironic/+/98214722:08
opendevreviewJay Faulkner proposed openstack/ironic master: Remove the iRMC hardware type and all interfaces  https://review.opendev.org/c/openstack/ironic/+/98217622:08
TheJuliaI'll put it this way, we have an underlying mandate to improve the state on my team, so on the ironic side, we likely want to add a tls enforcement threshold knob as well. The person who your speaking of is a product manager, not an engineer and also not the person to make the $other_security_stuff business cases, albeit they should at least try and I've had some recent discussions with them trying to guide them in the 22:09
TheJuliaright direction.22:09
TheJuliaHence, also, why two of those are WIPs for now, I want to see how horrible it will be ;)22:09
TheJulia(and, have this discussion)22:10
JayFProject manager or not, it's difficult to not paint a picture with all the data even if that individual isn't responsible for it all :)22:11
opendevreviewJay Faulkner proposed openstack/ironic master: Remove pysnmp and related dependencies  https://review.opendev.org/c/openstack/ironic/+/98214722:11
opendevreviewJay Faulkner proposed openstack/ironic master: Remove pysnmp and related dependencies  https://review.opendev.org/c/openstack/ironic/+/98214722:12
JayFthis stack should be happier now22:13
JayFit would've taken me a damn day to track that unit test failure w/o claude22:13
JayFtl;dr we were updating (in a test) `redfish.management` in a place that `redfish.management` existing relied on the iRMC driver  to have already been imported22:13
JayFI'll h/t it once CI is proven happy22:13

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!