Amarachi_O | Good Morning Ironic, I wish you all a great week! | 06:45 |
---|---|---|
rpittau | good morning ironic! o/ | 06:47 |
opendevreview | Adam McArthur proposed openstack/ironic master: api: Correct types for validation decorators https://review.opendev.org/c/openstack/ironic/+/945216 | 07:11 |
opendevreview | Adam McArthur proposed openstack/ironic master: api: Add schema for node firmware API https://review.opendev.org/c/openstack/ironic/+/945943 | 07:12 |
opendevreview | Adam McArthur proposed openstack/ironic-tempest-plugin master: Testing all microversion tests on CI https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/943086 | 07:16 |
opendevreview | Adam McArthur proposed openstack/ironic-tempest-plugin master: Adding better error messages to microversion tests https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/945945 | 07:16 |
opendevreview | Adam McArthur proposed openstack/ironic master: DNM: Testing ITP tests https://review.opendev.org/c/openstack/ironic/+/945742 | 07:17 |
freemanboss[m] | good morning ironic! o/ | 08:13 |
opendevreview | Adam McArthur proposed openstack/ironic master: api: Add schema for node firmware API https://review.opendev.org/c/openstack/ironic/+/945943 | 08:37 |
iurygregory | good morning ironic | 11:16 |
queensly[m] | Good morning :) | 12:02 |
opendevreview | Lennart Jern proposed openstack/sushy-tools master: Add config option SUSHY_EMULATOR_STORAGE_POOL https://review.opendev.org/c/openstack/sushy-tools/+/945959 | 12:46 |
freemanboss[m] | rpittau: please can the stand alone method also be run in a testenv? | 13:08 |
opendevreview | Vasyl Saienko proposed openstack/ironic master: Fix passing vtep fields to neutron https://review.opendev.org/c/openstack/ironic/+/945964 | 13:22 |
frickler | double-checking the meeting time after the DST change, I'm confused. all the recent meetings have been at 15 UTC, but this says 16? https://opendev.org/opendev/irc-meetings/src/branch/master/meetings/ironic-bare-metal-team-meeting.yaml | 13:27 |
dtantsur | frickler: I expect it to be in 90 minutes | 13:27 |
opendevreview | Julia Kreger proposed openstack/ironic master: devstack: network simulator support for sonic https://review.opendev.org/c/openstack/ironic/+/945726 | 13:27 |
frickler | dtantsur: that would be 15 UTC, which matches what the whiteboard says. so I guess I can do a patch to fix the calendar | 13:29 |
vsaienko_ | TheJulia, JayF: fyi https://docs.openstack.org/ironic/latest/admin/ovn-networking.html#vtep-switch-support this feature is not working. It has missing code parts in neutron that will create logical switch and bind ports (this step https://github.com/ovn-org/ovn/blob/f68cc0ce3163911dde38080cac6a5b38b097e254/ovn-architecture.7.xml#L2276). Besides of that OVN DHCP is not configured for vtep ports, and it seems it has to be fixed with | 13:30 |
vsaienko_ | some manual steps and legacy neutron-dhcp agent I was able to deploy BM node with openvswitch vtep emulator. On ironic side this patch is needed https://review.opendev.org/c/openstack/ironic/+/945964 | 13:30 |
frickler | https://review.opendev.org/c/opendev/irc-meetings/+/945965 if someone wants to verify | 13:33 |
TheJulia | good morning folks | 13:38 |
opendevreview | Julia Kreger proposed openstack/networking-baremetal master: DNM: Check CI status https://review.opendev.org/c/openstack/networking-baremetal/+/945966 | 13:45 |
TheJulia | vsaienko_: so, we've got a long list of issues with ovn, if you wouldn't mind adding a doc update to it. The DHCP stuff just is not surprising in the slightest given the history there. | 14:08 |
JayF | https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/943086/28 should be a nice easy review, we even have proven the tests work via a test change | 14:50 |
rpittau | "Generated-By: Gemini 2.0 Flash" lol | 14:53 |
JayF | adamcarthur5 uses those tools to great efficiency | 14:53 |
JayF | A lot of his gophercloud implementations (getting full coverage of ironic api) are AI-aided too | 14:53 |
JayF | I'm getting more and more curious how well one could make an AI coding assistant for openstack if we specifically trained it on our code/libraries/mailing lists/etc | 14:54 |
JayF | a good 15 years of public context | 14:54 |
opendevreview | Verification of a change to openstack/ironic master failed: Do not silence the actual error in prepare_instance_boot https://review.opendev.org/c/openstack/ironic/+/945844 | 14:57 |
rpittau | #startmeeting ironic | 15:00 |
opendevmeet | Meeting started Mon Mar 31 15:00:06 2025 UTC and is due to finish in 60 minutes. The chair is rpittau. Information about MeetBot at http://wiki.debian.org/MeetBot. | 15:00 |
opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 15:00 |
opendevmeet | The meeting name has been set to 'ironic' | 15:00 |
dtantsur | o/ | 15:00 |
rpittau | Hello everyone! | 15:00 |
rpittau | Welcome to our weekly meeting! | 15:00 |
rpittau | The meeting agenda can be found here: | 15:00 |
rpittau | https://wiki.openstack.org/wiki/Meetings/Ironic#Agenda_for_March_31.2C_2025 | 15:00 |
WiniciusAllan[m] | o/ | 15:00 |
kubajj | o/ | 15:00 |
JayF | \o | 15:00 |
TheJulia | o/ | 15:01 |
rpittau | let's wait one more minute for people to join | 15:01 |
cid | o/ | 15:01 |
rpittau | alright, let's go | 15:02 |
rpittau | #topic Announcements/Reminders | 15:02 |
rpittau | #info Standing reminder to review patches tagged ironic-week-prio and to hashtag any patches ready for review with ironic-week-prio: | 15:03 |
rpittau | #link https://tinyurl.com/ironic-weekly-prio-dash | 15:03 |
rpittau | we're back to pre-release numbers here, need to get to that! | 15:03 |
TheJulia | huh? | 15:04 |
rpittau | we have a lot of patches to review :D | 15:04 |
TheJulia | ahh, yes | 15:04 |
TheJulia | this is the way :) | 15:04 |
rpittau | :) | 15:04 |
rpittau | I'm actually approving the network patches, I had that in list and just needed to press the red button | 15:05 |
rpittau | I'll try to get to the rest during the week, going to be a busy week! | 15:06 |
rpittau | #info 2025.1 Epoxy Release Schedule | 15:07 |
rpittau | #link https://releases.openstack.org/epoxy/schedule.html | 15:07 |
rpittau | we're at R-0! We're releasing in a couple of days! | 15:07 |
rpittau | and don't forget Open Infra Live is this Thursday April 3 | 15:07 |
rpittau | #link https://www.youtube.com/watch?v=qX2-8oyFVPk | 15:07 |
rpittau | #info Flamingo PTG will take place place April 7-11, 2025! | 15:08 |
rpittau | next week! | 15:08 |
rpittau | meeting will be cancelled on Monday April 7 | 15:08 |
rpittau | please review the topics one more time https://etherpad.opendev.org/p/ironic-ptg-april-2025 before we order them for the schedule | 15:09 |
rpittau | also do we need nova cross-team session? | 15:09 |
rpittau | do we need neutron cross-team session? | 15:09 |
JayF | Not afaik from our end | 15:10 |
JayF | I would advise us to schedule around the eventlet removal and for as many people to attend that as possible | 15:10 |
rpittau | JayF: when's that? | 15:10 |
JayF | it's on the ptg board; or at least; will be | 15:10 |
JayF | just another set of ptg sessions | 15:10 |
rpittau | ok | 15:10 |
JayF | ah it's not up yet | 15:11 |
TheJulia | cardoe: do you think cross-neutron may be helpful? | 15:11 |
JayF | probably going to specifically wait and hit a quiet time so more folks can attend | 15:11 |
rpittau | ack | 15:11 |
cardoe | Yes for sure | 15:11 |
rpittau | cardoe: in that case we will need topics to propose and a slot that fits the two projects | 15:12 |
TheJulia | so consensus seems to be that we think we're good on the nova front, but not the neutron front and need a cross-project session if we can pull one off? | 15:12 |
rpittau | I would like to finalize the schedule before Thursday so it's ready for the Open Infra Live | 15:14 |
TheJulia | would there be any befit to ad-hoc discussions in advance of that so we can finalize topics? | 15:15 |
TheJulia | Just thinking it might help move our own clarity forward? | 15:15 |
rpittau | TheJulia: would probably help | 15:17 |
TheJulia | so maybe we try and get through meeting and would a walk through the etherpad be helful on a higher bandwidth call? | 15:18 |
rpittau | sure | 15:19 |
opendevreview | Verification of a change to openstack/ironic master failed: Do not silence the actual error in prepare_instance_boot https://review.opendev.org/c/openstack/ironic/+/945844 | 15:20 |
rpittau | there are no Discussion topics actually for today, unless someone has something :) | 15:20 |
rpittau | we can probably close here then | 15:21 |
rpittau | #endmeeting | 15:22 |
opendevmeet | Meeting ended Mon Mar 31 15:22:01 2025 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 15:22 |
opendevmeet | Minutes: https://meetings.opendev.org/meetings/ironic/2025/ironic.2025-03-31-15.00.html | 15:22 |
opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/ironic/2025/ironic.2025-03-31-15.00.txt | 15:22 |
opendevmeet | Log: https://meetings.opendev.org/meetings/ironic/2025/ironic.2025-03-31-15.00.log.html | 15:22 |
TheJulia | so, high bandwidth discussion regarding etherpad and networking topics? | 15:23 |
rpittau | yep, I'm updating the etherpad with the days | 15:24 |
rpittau | adding an epoxy retrospective topic for monday | 15:24 |
frickler | does anyone have experience with cisco server hardware? I'll likely have the pleasure soon and would be happy to learn about prior art | 15:24 |
TheJulia | frickler: I'm so very sorry | 15:25 |
frickler | TheJulia: that's very comforting, but kind of what I expected, thank you ;) | 15:26 |
rpittau | (don't say it, don't say it) | 15:26 |
rpittau | frickler I do have experience with cisco servers, unfortunately (damn!) | 15:26 |
TheJulia | frickler: Some of us have had our run-ins with Cisco server gear. Everything can vary based upon firmware/season/alignment | 15:27 |
rpittau | to be fair the latest generations/firmwares are not that bad | 15:27 |
TheJulia | rpittau: is that the ones where they decided all redfish is just read-only? | 15:27 |
TheJulia | Or is this the version which likes redfish and acts funky with ipmi?! | 15:28 |
rpittau | lol I think they went beyond that | 15:28 |
rpittau | ipm... what?!? :D | 15:28 |
TheJulia | frickler: ^^^ should sort of demonstrate some of the challenges you may or may not encounter | 15:28 |
TheJulia | how... comforting | 15:28 |
* TheJulia wonders if everyone needs more coffeeeee this morning | 15:34 | |
JayF | I need more weekend :) | 15:36 |
rpittau | I Was thinking the same | 15:36 |
TheJulia | ... This would be nice | 15:36 |
frickler | for me it is evening and I think I need something else now ... but thanks anyway. at least I'm hoping that new hardware comes with new firmware | 15:36 |
rpittau | frickler: if you use redfish )highly recommended) you should be ok-ish using 5.2 version or later for UCS-X and 4.3 or later for UCS-C | 15:39 |
TheJulia | I know we've seen some new versions with odd behavior as well, but generally folks complain loudly and then they get fixed/updated firmware and seem "okay" | 15:40 |
rpittau | TheJulia: one thing I remember we discussed last PTG was about the starting time, I put 1300 UTC this time but I'm now wondering if it's too early for America based people | 15:44 |
rpittau | JayF, iurygregory ^ | 15:44 |
JayF | lets do some math, it's 1545 UTC / 8:45am PDT right now | 15:45 |
JayF | - 2:45 == 6:00am | 15:45 |
TheJulia | rpittau: that is presently 6 AM US Pacific. Could be better, could be worse. | 15:45 |
JayF | I can swing it but don't expect me to have energy :) | 15:45 |
TheJulia | Likewise | 15:45 |
rpittau | we can move 1 hour later, no problem | 15:45 |
cardoe | is https://review.opendev.org/c/openstack/ironic/+/944769 ready for +W? | 15:45 |
rpittau | I will move it 1 hour later :) | 15:46 |
TheJulia | cardoe: you suggested anaconda stuffs were not required on your end any longer? Do we still need/want the ptg topic? | 15:47 |
cardoe | We can scrap it. | 15:47 |
cardoe | I need to go in and do the updates I've promised I'd make. | 15:47 |
cardoe | Sorry skating in to a Q1 deliverable by the skin of my teeth. | 15:48 |
TheJulia | I marked anaconda stuffs as struckthrough | 15:49 |
frickler | rpittau: yes, redfish is the plan, I'll note down those versions, thx | 15:52 |
iurygregory | rpittau, for me it's ok | 15:52 |
iurygregory | it's 10am so no problem on my side | 15:53 |
rpittau | ack | 15:53 |
rpittau | I moved the starting time 1 hour later anyway, I know we have more people from North America that may join | 15:54 |
opendevreview | Merged openstack/ironic master: Fix confusing logs on tenant network configuration https://review.opendev.org/c/openstack/ironic/+/945462 | 15:55 |
TheJulia | rpittau: I guess we might not be able to gain clarity on specific overlap items with the neutron folks until tomorrow sometime? That might also make scheduling different in general | 16:05 |
rpittau | TheJulia: that's ok, I will review the topics and make a temp schedule in a different file, then update based on the neutron outcome | 16:06 |
TheJulia | JayF: fyi, I added two comments to https://review.opendev.org/c/openstack/ironic-specs/+/945642 | 16:09 |
JayF | 👀 | 16:10 |
opendevreview | Merged openstack/ironic master: network testing: hooking in an external network simulator https://review.opendev.org/c/openstack/ironic/+/942298 | 16:20 |
opendevreview | Merged openstack/ironic master: Add network simulator support for force10 OS 10 https://review.opendev.org/c/openstack/ironic/+/943345 | 16:20 |
opendevreview | Merged openstack/ironic master: docs: detail network switch simulator support https://review.opendev.org/c/openstack/ironic/+/944139 | 16:20 |
opendevreview | Merged openstack/ironic master: network simulator support for Cisco Nexus 9k https://review.opendev.org/c/openstack/ironic/+/944873 | 16:20 |
* TheJulia does a happy dnace | 16:21 | |
TheJulia | dance | 16:21 |
opendevreview | Merged openstack/networking-generic-switch master: Adding notes for SONiC switches https://review.opendev.org/c/openstack/networking-generic-switch/+/945887 | 16:27 |
opendevreview | Satoshi Shirosaka proposed openstack/ironic-python-agent master: WIP Implement manual cleaning for ContainerHardwareManager https://review.opendev.org/c/openstack/ironic-python-agent/+/945862 | 16:59 |
opendevreview | Merged openstack/ironic-python-agent master: Remove eventlet greendns patching https://review.opendev.org/c/openstack/ironic-python-agent/+/945043 | 17:38 |
* JayF is wrangling failing unit tests and failing miserably | 18:34 | |
JayF | I think I've got a mock missing or screwed up somewhere based on the errors; all the tests are failing with things like ironic.common.exception.DriverNotFoundInEntrypoint: Could not find the following items in the 'ironic.hardware.interfaces.firmware' entrypoint: fake, no-firmware. | 18:34 |
TheJulia | hey, I think I just hit that as well! | 18:48 |
opendevreview | Jay Faulkner proposed openstack/ironic master: WIP: Automated cleaning by runbook https://review.opendev.org/c/openstack/ironic/+/945259 | 18:48 |
JayF | TheJulia: the thing that screws with me: it works on master | 18:49 |
TheJulia | JayF: regarding training an AI in openstack context, the board likely smile for such an effort. :) | 18:49 |
JayF | adamcarthur5: ^ | 18:49 |
JayF | TheJulia: so master unit tests work, 945259 unit tests go boom with those entrypoint errors | 18:50 |
JayF | I'm going to step away for a minute before I put a laptop sized hole in the wall, but if anyone has any ideas please pass them along | 18:51 |
TheJulia | I'm just starting to look at it myself | 18:51 |
TheJulia | I've got a stupidly minor change which shouldn't break this way but it is | 18:52 |
opendevreview | Merged openstack/ironic master: Do not silence the actual error in prepare_instance_boot https://review.opendev.org/c/openstack/ironic/+/945844 | 18:53 |
TheJulia | JayF: rm ironic.egg-info | 18:55 |
frickler | oh, that's what stephenfin and me were seeing like a week ago, I wouldn't have matched it if TheJulia hadn't remembered the solution | 19:00 |
frickler | (or reinvented?) | 19:00 |
TheJulia | semi-remembered | 19:00 |
frickler | would still be interesting to know what is triggering this, but in tox or pip or setuptools or where? | 19:01 |
JayF | oh hell | 19:01 |
frickler | s/but/bug/ | 19:02 |
TheJulia | no idea, super curious as well | 19:02 |
JayF | > - Failed: 7 | 19:05 |
JayF | Hey look, we fixed about 4000 unit tests compared to previous runs | 19:05 |
JayF | lol | 19:05 |
JayF | I have a question for folks | 20:19 |
* JayF pauses to verify before he says it | 20:21 | |
JayF | So the fact we have an ironic.cmd module appears to be breaking the ability for vscode to discover unit tests in ironic | 20:23 |
JayF | due to there being a built-in cmd module that is used by pdb | 20:23 |
JayF | how would folks feel about just renaming that module? | 20:23 |
opendevreview | Jay Faulkner proposed openstack/ironic master: Rename ironic cmd module https://review.opendev.org/c/openstack/ironic/+/945999 | 20:28 |
TheJulia | I wouldn't be opposed | 20:33 |
JayF | lets see how that change looks in CI | 20:34 |
JayF | I've been down a rabbithole of testing AI coding tools, and cursor is based on vscode | 20:35 |
JayF | combined with fighting unit tests and you can see why I want this to work lol | 20:35 |
cardoe | So TheJulia, JayF's https://review.opendev.org/c/openstack/ironic/+/945259 is kinda where I was going with the PTG item " require deploy templates for ALL deploy steps " | 20:55 |
JayF | that has nothing to do with deployment? | 20:56 |
JayF | and there are technical reasons why deploy templates are distinct from runbooks | 20:56 |
cardoe | It doesn't but along the similar vein | 20:56 |
cardoe | That we use deploy templates | 20:56 |
JayF | are you suggesting getting rid of the ability to manually provide deployment steps in standalone? | 20:56 |
JayF | I've read through your comments along those lines several times and I just don't get it | 20:56 |
cardoe | So the steps that are followed today require you to look at what's enabled in the ironic.conf and then what's set in the IPA | 20:57 |
cardoe | What if you could have a complete deploy template with all the steps encoded in that. | 20:57 |
cardoe | There wouldn't be a combination of sources if that makes sense. | 20:58 |
cardoe | I'm failing at describing this. | 20:58 |
cardoe | Honestly we can strike it until I can dedicate more brain cycles to it. | 20:58 |
JayF | why don't we approach it from the base thing | 20:58 |
JayF | like, what are you trying to achieve? what's the end goal? | 20:58 |
JayF | in this case, my downstream wants to be able to change what happens in automated cleaning *without* having to push a new IPA ramdisk | 20:59 |
JayF | which runbook-automated-cleaning + ContainerHWM delivers | 20:59 |
JayF | what's the larger story for your ask? | 20:59 |
cardoe | Given an Ironic system, what will the deployment steps be that will run. Similarly what will the cleaning steps that will run. | 21:00 |
cardoe | How can I probe two different systems and see if they'll execute the same steps or not. | 21:00 |
TheJulia | It is a chicken/egg problem | 21:00 |
TheJulia | ... I guess I'll need some whiskey for next week | 21:01 |
JayF | cardoe: there's a ... lot that needs to be tackled if we want to do this | 21:03 |
JayF | cardoe: nova has a philosophy that every user-facing behavior difference should be visible via api | 21:04 |
JayF | that's almost impossible to do in ironic AND maintain some semblance of hardware-specific behavior | 21:04 |
TheJulia | Well, that is side effect, the underlying challenge is state is ultimately determined by where things are at and where they need to go. You can't identify that outside of process which is the challenge for repeatability, but not everything is at a clean state and you furthermore have other variables to address like what happens when an IPA image is updated. | 21:05 |
cardoe | So I think there's some valid things to caveat with this. Because we ultimately do create the list of deployment steps because we set them on the node. | 21:06 |
JayF | I might suggest a more achievable goal might be to make an API-readable/modifyable path for everything, even if it's not the only path | 21:06 |
JayF | which is sorta where we are with the combination of runbooks + cleaning when my patch is complete | 21:07 |
TheJulia | Yes, and where this whole thing broke down last time was that the demand was to have an API surface where we go and update that as soon as you go ask for that data. | 21:07 |
TheJulia | which means... powering down workloads to discover updated data | 21:07 |
JayF | I mean, service makes all this possible, if we want to serve the operator at the cost of the user :D | 21:07 |
TheJulia | technically, yeah | 21:08 |
JayF | "your instance will brb, running service to see what steps are available, brb" | 21:08 |
JayF | that is being pointed out explicitly as a negative side effect example :); I do not think we should do that to be explicit | 21:08 |
cardoe | So I was thinking about that availability thing with the priorities. | 21:08 |
TheJulia | "node powers off" "node gets iso attached" "node powers into PARTY RAMDISK!" | 21:08 |
cardoe | e.g. if I edited ironic.conf to add step X at priority 60 | 21:08 |
cardoe | Have that visible. Or if I turned the step off | 21:09 |
cardoe | I understand today that IPA drives what steps are available. | 21:09 |
JayF | in the past, the closest we've gotten to this is caching steps from the last time we fetched them | 21:10 |
cardoe | I didn't have a final proposed "this is how I think it should work" plan. | 21:10 |
JayF | the conversation I wanted to happen is happening now though; I understand your goal so at least can grok where you're going | 21:10 |
TheJulia | This is useful context | 21:11 |
TheJulia | so less of a need to just begin with the whiskey | 21:11 |
TheJulia | :) | 21:11 |
cardoe | I'm sitting here thinking I've got Ironic right now deployed to 4 DCs. I want the same deploy steps and the same clean steps in all 4. | 21:11 |
cardoe | And today auditing that isn't programmatic at all. | 21:11 |
JayF | I wonder if something like ... | 21:12 |
TheJulia | I *hate* the idea of revealing tons of ironic's internal state config via the api, but perhaps there is way to raise up enough of it in a limited scope | 21:12 |
JayF | store the set of steps run in [thing] in node history, optionally | 21:12 |
JayF | I think the way to approach this is to record things that have happened, for post-action auditing | 21:12 |
JayF | rather than make a proactive api | 21:12 |
TheJulia | +1 to also recording what has happened | 21:12 |
TheJulia | we should definitely be doing that in a higher level of detail | 21:13 |
cardoe | I was then taking it a step further and saying "should we then use deploy templates" always so that I can set the same deploy template and it errors because something isn't available in 1. | 21:13 |
JayF | I think deploy templates + custom_agent behaves like you want | 21:13 |
cardoe | Rather than discovering things didn't behave correctly AFTER one node got built or cleaned wrong to catch the problem before hand. | 21:13 |
JayF | with no implied automatic behavior | 21:13 |
JayF | for deployment | 21:14 |
JayF | custom_agent w/deploy templates + my feature == fully API contained steps | 21:14 |
cardoe | Yeah I'm pretty confident I can handle this today with a little bit of custom code. I just threw this out there as is this a valid case that others might be interested in... and if so how would we have that. | 21:15 |
cardoe | I'm just trying to ask if my downstream cases are valid to other users and if they are, how can I upstream it. | 21:16 |
JayF | I guess I'm asking what is 'that' if not use of custom_agent driver w/deploy_templates + my new automated-cleaning-via-runbook | 21:16 |
cardoe | So it's also fair that a lot of my questions are "no, go away and do your weird downstream thing" | 21:16 |
JayF | I think I still am missing a subtlety about what you're asking | 21:16 |
cardoe | I've got multiple hardware types. Different clean steps for each. Different deploy steps for them potentially as well. | 21:17 |
JayF | how are those clean/deploy steps changing? HWM magic? Using a different deploy_template? Combination? | 21:18 |
cardoe | But my hardware types are consistent between my DCs (they better be or I'm bringing a pipe wrench and discovering people's passwords). So I want to know my different Ironics are doing the same thing. | 21:18 |
JayF | If that's the single high level question; I almost feel like the answer is "this is the job of your config management system / deploy mechanism for ironic" because so much of this is about getting ironic.confs synced up | 21:19 |
cardoe | Right now HWM magic in IPA. But I was asking the team if we could do this all with different deploy templates. | 21:19 |
JayF | unless you just want more stuff shaped like "use X template/runbook for Y resource class" | 21:19 |
cardoe | That's exactly what I want. | 21:19 |
cardoe | I see your container stuff giving that power | 21:20 |
JayF | You want to map default deploy template by resource class if none is specified by trait hinting from nova? | 21:20 |
cardoe | That's my thought. | 21:20 |
JayF | only other thing which might make sense along those lines is if you need IPA ramdisk by resource_class | 21:21 |
cardoe | Didn't realize nova could hint the deploy template. | 21:21 |
JayF | it already exists by_arch | 21:21 |
JayF | yeah, that's the primary use case | 21:21 |
cardoe | oh. :-D didn't glean that from the docs. | 21:21 |
JayF | nova flavor bm-with-smt vs bm-without-smt and each triggering a deploy_template with bios.apply_configuration :) | 21:21 |
cardoe | oh yeah yes | 21:21 |
cardoe | nvm yes I knew that | 21:22 |
JayF | yeah so more stuff indexed by resource class I think makes sense for config that could be different based on hardware | 21:23 |
JayF | I would potentially be onboard for default deploy template by resource class if not specified otherwise | 21:23 |
JayF | the "potentially" is only because I don't know where any of the dragons would lie :P | 21:23 |
cardoe | I'm also wanting something that would be simpler. If its not simpler then never mind. | 21:26 |
cardoe | when I chased down the anaconda deploy stuff finding all the places we've got "if this then that else other" where the conditionals were the same. | 21:27 |
JayF | I think the only arbiter of "simpler" is a user's opinion | 21:27 |
JayF | in terms of UX like this | 21:27 |
JayF | (you are also the user in this case I think) | 21:30 |
opendevreview | Satoshi Shirosaka proposed openstack/ironic-python-agent master: WIP Implement manual cleaning for ContainerHardwareManager https://review.opendev.org/c/openstack/ironic-python-agent/+/945862 | 21:37 |
ayo_ | . | 21:49 |
ayo_ | . | 21:59 |
opendevreview | Julia Kreger proposed openstack/ironic master: redfish: permit ipmitool-socat console for redfish usage https://review.opendev.org/c/openstack/ironic/+/946004 | 22:01 |
TheJulia | so, that seemed easiest. Going the fully right way was a ton of work in the end. | 22:08 |
JayF | ayo_: - | 22:11 |
JayF | TheJulia: +2 with a concrete suggestion if you have 5 min | 22:12 |
TheJulia | sure sure | 22:13 |
opendevreview | Jay Faulkner proposed openstack/ironic master: Rename ironic cmd module https://review.opendev.org/c/openstack/ironic/+/945999 | 22:13 |
opendevreview | Julia Kreger proposed openstack/ironic master: redfish: permit ipmitool-socat console for redfish usage https://review.opendev.org/c/openstack/ironic/+/946004 | 22:16 |
JayF | same deal again, +2 but a concrete suggestion (using gerrit suggestions this time, no less!) | 22:18 |
opendevreview | Merged openstack/ironic-python-agent-builder master: Convert README.md to README.rst https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/945773 | 22:25 |
opendevreview | Julia Kreger proposed openstack/ironic master: redfish: permit ipmitool-socat console for redfish usage https://review.opendev.org/c/openstack/ironic/+/946004 | 22:44 |
TheJulia | JayF: afaik, the constraint applies to irmc as well, so I made it generic. | 22:54 |
JayF | ack, sgtm, I already restored my +2 | 22:55 |
opendevreview | Satoshi Shirosaka proposed openstack/ironic-python-agent master: Implement manual cleaning for ContainerHardwareManager https://review.opendev.org/c/openstack/ironic-python-agent/+/945862 | 23:40 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!