| zouyu[m] | Hello, do we have any thoughts to support specifying a URL for verify_ca, rather than just a local path?... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/vNTYAPHdckzvzSCesnSGOGGG>) | 02:10 |
|---|---|---|
| TheJulia | zouyu[m]: o/ | 05:01 |
| TheJulia | zouyu[m]: ... That seems reasonable to me, would you be willing to propose that as a change to the code base? | 05:01 |
| TheJulia | stevebaker[m]: dtantsur: rpittau: we likely need to go through the settings and consider the same sort of thing for *anything* which could be a file path on the conductor... maybe with the exception of course as I type this I think there is a case where we wouldn't want to, but I'm too tired and depressed right now to actually recall what that setting is/was | 05:03 |
| TheJulia | w/r/t zouyu[m]'s question | 05:04 |
| *** fenggw[m] is now known as GuangwenFeng[m] | 05:37 | |
| zouyu[m] | TheJulia: Thanks for your reply. Yes, I plan to submit the proposal to storyboard. | 06:31 |
| TheJulia | zouyu[m]: it may just be easier to do it in a gerrit change, since it seems a bit obvious. | 06:45 |
| zouyu[m] | TheJulia: Thanks! Got it. | 06:51 |
| dtantsur | rpittau: re https://review.opendev.org/c/openstack/ironic/+/863873, wdyt about merging it with a reference to your github? | 09:16 |
| rpittau | good morning ironic, happy Friday! o/ | 09:21 |
| rpittau | JayF: awesome, thanks for taking care of that | 09:21 |
| rpittau | dtantsur: sounds good to me, I'll reenable the other jobs and clean it a little first | 09:22 |
| rpittau | And I'll have to remember to not close my branch | 09:23 |
| dtantsur | true :) | 09:32 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/21.3: Configure CI for bugfix/21.3 https://review.opendev.org/c/openstack/ironic/+/876264 | 09:45 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/21.3: Configure CI for bugfix/21.3 https://review.opendev.org/c/openstack/ironic/+/876264 | 09:47 |
| dtantsur | rpittau: FYI ^^ | 09:48 |
| rpittau | dtantsur: thanks for that, going to review ASAP | 09:57 |
| rpittau | mmm with all these numbers I'll go crazy(er) | 10:00 |
| dtantsur | yeah | 10:04 |
| opendevreview | Merged openstack/ironic bugfix/21.3: Do not move nodes to CLEAN FAILED with empty last_error https://review.opendev.org/c/openstack/ironic/+/876173 | 10:10 |
| kubajj | Good morning Ironic o/ | 10:12 |
| opendevreview | Merged openstack/ironic bugfix/21.0: Do not move nodes to CLEAN FAILED with empty last_error https://review.opendev.org/c/openstack/ironic/+/876176 | 10:15 |
| dtantsur | kubajj: morning! JFYI I've seen your email, planning to work on a response over the course of the day. Should I simply reply to you or copy anyone? | 10:24 |
| kubajj | dtantsur: thanks, just reply to the email, I will then extract quotes into the dissertation to do something like a qualitative evaluation 🙂 | 10:25 |
| rpittau | quick review if anyone has a minute, thanks! https://review.opendev.org/c/openstack/bifrost/+/872634 | 10:28 |
| dtantsur | +2 | 10:30 |
| rpittau | tnx! | 10:36 |
| opendevreview | Riccardo Pittau proposed openstack/bifrost master: Fix enabling epel repo for rpm distributions https://review.opendev.org/c/openstack/bifrost/+/875929 | 11:31 |
| rpittau | JayF: as soon as virtualpdu has been moved we should request a new release, or can we even do it now? | 11:35 |
| iurygregory | good morning Ironic | 11:40 |
| dtantsur | rpittau: projects under x/ are not managed by the release team | 11:42 |
| dtantsur | so if you have ACL for pypi, you can release. if not, well. another round of conversations :) | 11:43 |
| dtantsur | good news: openstackci is a maintainer of the pypi project | 11:43 |
| dtantsur | so you should be able to release | 11:43 |
| dtantsur | Hmmmm, why are we trying to PXE-boot an ISO in the CI? | 14:13 |
| iurygregory | O.o | 14:14 |
| iurygregory | wrong config in the job? | 14:14 |
| dtantsur | mmm, no, it's intended apparently. but with redfish we're supposed to do virtual-media... | 14:15 |
| TheJulia | is this a tempest job on a non-virtual-media job? | 14:17 |
| dtantsur | TheJulia: standalone | 14:17 |
| opendevreview | Oleksandr Kozachenko proposed openstack/networking-generic-switch master: aruba: Override check string for config mode https://review.opendev.org/c/openstack/networking-generic-switch/+/876360 | 14:18 |
| TheJulia | so yeah, that is intentional then since it is not the explicit virtual media job I think | 14:18 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic-tempest-plugin master: Split the ramdisk ISO test https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/876361 | 14:18 |
| dtantsur | TheJulia: my guess ^^^ | 14:18 |
| TheJulia | yeah, that makes sense to do | 14:19 |
| dtantsur | let's see what the CI says | 14:20 |
| opendevreview | OpenStack Release Bot proposed openstack/networking-generic-switch stable/2023.1: Update .gitreview for stable/2023.1 https://review.opendev.org/c/openstack/networking-generic-switch/+/876362 | 14:20 |
| opendevreview | OpenStack Release Bot proposed openstack/networking-generic-switch stable/2023.1: Update TOX_CONSTRAINTS_FILE for stable/2023.1 https://review.opendev.org/c/openstack/networking-generic-switch/+/876363 | 14:20 |
| opendevreview | OpenStack Release Bot proposed openstack/networking-generic-switch master: Update master for stable/2023.1 https://review.opendev.org/c/openstack/networking-generic-switch/+/876364 | 14:20 |
| * TheJulia tries to wake up | 14:20 | |
| TheJulia | Good morning everyone | 14:20 |
| rpittau | dtantsur: right! And thanks, going to ask for a release and see what happens :) | 14:36 |
| JayF | rpittau: ++ you got to the ending I was going to suggest | 14:53 |
| * TheJulia binks at the size of a newer centos stream 9 iso | 15:10 | |
| samuelkunkel[m] | Did it get even bigger? | 15:10 |
| samuelkunkel[m] | Are they aiming for the 1GB? ;) | 15:11 |
| TheJulia | 8.8GB | 15:12 |
| samuelkunkel[m] | hu? :D | 15:12 |
| samuelkunkel[m] | did they smuggle in all ubuntu releases of the last 2 years in there? | 15:13 |
| TheJulia | no idea... | 15:13 |
| JayF | I honestly gotta say again, it's tough for me to understand why we wanna be on the stream treadmill vs something like rocky :/ | 15:13 |
| JayF | especially given stuff like this can break us | 15:13 |
| TheJulia | the breaky things actually some of us find them far earlier as opposed to upon the next major upgrade, luckily we've only seen ?2? breaks in stream one intentional changes, the other was a "rutro, revert" if memory servers | 15:14 |
| dtantsur | morning TheJulia, JayF | 15:14 |
| JayF | o/ | 15:14 |
| dtantsur | rpittau: s/ask for release/make a release/ ;) | 15:15 |
| dtantsur | make sure you have ACL to push tags | 15:15 |
| JayF | there's a nonzero chance the ACLs are not setup for custom releases | 15:15 |
| JayF | wait, they should still be | 15:15 |
| JayF | because my project-config change hasn't landed | 15:15 |
| dtantsur | right | 15:15 |
| JayF | virtualpdu-release can cut releases | 15:15 |
| JayF | https://review.opendev.org/admin/groups/cc93c0bed1f8a435b44a06c8539b3dfb5379d0ee,members yeah we're good | 15:16 |
| JayF | ironic-core is in virtualpdu-release | 15:16 |
| JayF | so you might have to do it manually but it can be done | 15:16 |
| dtantsur | s/might// the release team does not handle /x/ stuff | 15:16 |
| JayF | if I were doing it, I'd sync with releases team before just yolo-doing it, but expect the answer to be "do your own release" | 15:16 |
| TheJulia | yeah, they don't handle x namespace stuff | 15:17 |
| JayF | the actual move from x->openstack is likely not happening until post-release | 15:17 |
| dtantsur | which reminds me, I haven't released or touched ironic-staging-drivers for... ahemmmm.... | 15:17 |
| JayF | even though we are, from a governance standpoint, running virtualpdu | 15:17 |
| TheJulia | c'est la vie | 15:17 |
| JayF | (that merged yesterday afternoon) | 15:17 |
| TheJulia | dtantsur: I'd wonder if it makes sense to... | 15:17 |
| dtantsur | yay, ironic-staging-drivers does not even have stable/zed \o/ \o/ | 15:18 |
| TheJulia | w/r/t staging-drivers | 15:18 |
| dtantsur | if we do it, I'd prefer that we switch to a group ownership | 15:18 |
| dtantsur | i.e. bring it under ironic | 15:18 |
| JayF | so we have already done it | 15:19 |
| JayF | https://opendev.org/openstack/governance/commit/db8597ad9245ce2c115b4d5140a6d81d63b2a9af | 15:19 |
| dtantsur | yeah, I meant ironic-staging-drivers | 15:19 |
| JayF | oh. | 15:19 |
| JayF | I would not be in favor of us adopting that... | 15:20 |
| dtantsur | sorry, side-tracking. just remembered that it's been dead since zed | 15:20 |
| dtantsur | rpittau and myself cannot keep it alive | 15:20 |
| dtantsur | if we're not taking it in ironic, I suggest adopting the wakeonlan driver and dropping everything else | 15:20 |
| rpittau | forgot about that :/ | 15:20 |
| JayF | Let me ask it this way: are you all trying to keep it alive out of courtesy? RH/business requirements? Something else? | 15:20 |
| dtantsur | rpittau: no blame here, we're terribly busy, both of us | 15:20 |
| JayF | the idea is that like, it's in staging-drivers because there was nobody to maintain it | 15:20 |
| dtantsur | JayF: out of courtesy. OpenShift could not care less. | 15:21 |
| JayF | so it seems like if it doesn't get maintained, that's probably OK? Someone using it should step up and help if they care | 15:21 |
| dtantsur | OSP used to use oVirt, probably not any more | 15:21 |
| JayF | at a certain point we only have so many brains and hands and we have to point them at the primary concerns of the project | 15:21 |
| TheJulia | I'm starting to wonder if anyone actually uses any of those drivers at this point | 15:21 |
| dtantsur | I could argue that having a WakeOnLan driver in-tree may be pretty cool | 15:21 |
| dtantsur | for people who don't have fancy hardware at all | 15:21 |
| TheJulia | dunno about iboot, maybe WOL... | 15:21 |
| dtantsur | everything else is a pure burden for me | 15:21 |
| JayF | I think that's neat, I struggle with it being hard to test. I don't have strong opinions about any individual driver in there; I do have opinions about you all throwing your time into an unused black hole lol | 15:22 |
| dtantsur | to be clear, that's all the code: https://opendev.org/x/ironic-staging-drivers/src/branch/master/ironic_staging_drivers/wol/power.py | 15:22 |
| dtantsur | I can live with these 200 lines not having a CI, but I understand the others may feel differently | 15:23 |
| JayF | My bigger issue is that if it exists as a driver, people will expect a certain level of quality. Not being able to implement, e.g. `reboot`, seems like a serious impediment | 15:23 |
| dtantsur | I sincerely hope that people who know what WOL is also understand that it cannot power off | 15:24 |
| dtantsur | but dunno. I don't care much in the end. | 15:24 |
| JayF | My bigger concern is that like, ironic-staging-drivers exists so that we have a place for that potentially-useful code to live even when folks have abandoned it | 15:25 |
| JayF | if you aren't using those drivers, and don't know anyone who is, I implore you to leave it alone | 15:25 |
| dtantsur | leave it alone = forget or deprecate | 15:25 |
| dtantsur | ? | 15:25 |
| JayF | it's x/ namespace, right? | 15:25 |
| dtantsur | yes | 15:26 |
| JayF | like I'd just not prioritize maintainence of it | 15:26 |
| JayF | wither on the vine is more or less my call | 15:26 |
| dtantsur | I'm not operating this way, sorry | 15:26 |
| dtantsur | If I'm responsible for it, I either need to keep being responsible or to officially stop | 15:26 |
| dtantsur | I feel bad for not doing anything since Yoga | 15:26 |
| rpittau | so summarizing for virtualpdu, if I create the release patch now, and it gets acceppted, we'll have to create tags and whatever ourselves, like we do for staging-drivers ? | 15:26 |
| JayF | rpittau: that's our understanding, yes | 15:27 |
| rpittau | ok, thanks | 15:27 |
| JayF | rpittau: they might not even accept the patch tbh | 15:27 |
| dtantsur | I can put up a call on the ML. If nobody volunteers, I'll deprecate the project and leave it with a warning in the docs. | 15:27 |
| rpittau | I'll give that a try anyway, we need a new release | 15:27 |
| dtantsur | rpittau: it won't be accepted until it's under our governance | 15:27 |
| JayF | dtantsur: ++ I'm OK with that. And if you feel strongly about WoL driver I won't -1 it or anything | 15:27 |
| rpittau | ah :/ | 15:27 |
| JayF | dtantsur: it is under our governance, just the location hasn't moved | 15:27 |
| JayF | I think releases cares about repo location, not governance | 15:27 |
| * rpittau grabs a beer | 15:28 | |
| dtantsur | JayF: they care about both, they expect the two not to conflict :) | 15:28 |
| JayF | well we're going to be in the nebulous grey area for a few weeks :D | 15:28 |
| dtantsur | then you'll need to bootstrap the openstack/releases repo with the new deliverable | 15:28 |
| dtantsur | yeah | 15:28 |
| rpittau | yeah virtuapdu is not there | 15:28 |
| dtantsur | I'd release first, then finalize the processes (including removing the ACL from us) | 15:28 |
| JayF | might be valuable to email the list, in that same thread, when we cut a release | 15:29 |
| JayF | it's good to be loud about how efficient we are, and this virtualpdu thing has been annoying, but we've been sprinting ever since we were handed the ball | 15:29 |
| rpittau | it's fun, although it might be the load of the week to talk | 15:31 |
| dtantsur | email sent | 15:31 |
| dtantsur | (about staging-drivers)\ | 15:31 |
| TheJulia | so intel node-manager became datacenter manager | 15:32 |
| rpittau | should we treat virtualpdu as virtualbmc? as independent I mean | 15:37 |
| JayF | I trust you alls' take on that. I don't have a strong opinion. | 15:39 |
| * JayF afk for a few hours | 15:41 | |
| dtantsur | I'd keep it the same way, yes | 15:43 |
| rpittau | btw just found a bug in the new_release script, and filed a fix | 15:50 |
| opendevreview | Merged openstack/networking-generic-switch stable/2023.1: Update .gitreview for stable/2023.1 https://review.opendev.org/c/openstack/networking-generic-switch/+/876362 | 16:12 |
| opendevreview | Merged openstack/networking-generic-switch stable/2023.1: Update TOX_CONSTRAINTS_FILE for stable/2023.1 https://review.opendev.org/c/openstack/networking-generic-switch/+/876363 | 16:15 |
| opendevreview | Merged openstack/networking-generic-switch master: Update master for stable/2023.1 https://review.opendev.org/c/openstack/networking-generic-switch/+/876364 | 16:15 |
| dtantsur | looking for a 2nd +2 on https://review.opendev.org/c/openstack/ironic/+/876264/ please, needed for the CI | 16:36 |
| dtantsur | also https://review.opendev.org/c/openstack/ironic/+/855042 has been around since august :) | 16:36 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/21.2: Fixes for tox 4.0 https://review.opendev.org/c/openstack/ironic/+/876409 | 16:39 |
| dtantsur | this is needed for the CI as well ^^ | 16:39 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/21.2: Update .gitreview for bugfix/21.2 https://review.opendev.org/c/openstack/ironic/+/867826 | 16:39 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/21.2: Do not move nodes to CLEAN FAILED with empty last_error https://review.opendev.org/c/openstack/ironic/+/876174 | 16:40 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/21.2: Configure CI for bugfix/21.2 https://review.opendev.org/c/openstack/ironic/+/876410 | 16:41 |
| dtantsur | someone just came to the metal3 slack with https://orange-opensource.gitlab.io/kanod/reference/architecture/added-value.html#bmaas | 16:57 |
| dtantsur | this is curious (cc JayF TheJulia for a Friday reading) | 16:58 |
| dtantsur | also https://orange-opensource.gitlab.io/kanod/reference/kiab/broker.html#kiab-broker | 17:00 |
| * dtantsur is sad nothing of the ownership work is available to metal3 currently | 17:01 | |
| TheJulia | legacy bios support only, wow | 17:02 |
| TheJulia | well, inherently if you want to use a proejct-scoped user to an existing ironic... | 17:02 |
| TheJulia | but it all depends on the use case and access model | 17:03 |
| * TheJulia is flabbergasted that centos9 stream is still installing | 17:03 | |
| dtantsur | it's streaming \o/ | 17:04 |
| TheJulia | surprisingly enough, it has 2.2GB of free ram on a 4gb VM | 17:06 |
| dtantsur | $ free -h | 17:07 |
| dtantsur | total used free shared buff/cache available | 17:08 |
| dtantsur | Mem: 23Gi 1.6Gi 16Gi 297Mi 5.8Gi 21Gi | 17:08 |
| dtantsur | This is with a whole bifrost running! | 17:08 |
| rpittau | about bifrost, in the patch where I'm trying to fix the EPEL repo, removing and reinstalling the epel-release package does the trick | 17:16 |
| rpittau | I have the suspect he pkg is installed but the repos entries are purged from the original node, haven't verified it yet | 17:16 |
| dtantsur | ask infra? | 17:17 |
| rpittau | yeah, it's really puzzling | 17:17 |
| dtantsur | well.. after the issue with wiped tzdata or what was it? | 17:18 |
| rpittau | oh yes, right! | 17:18 |
| rpittau | in the container image | 17:19 |
| dtantsur | yep | 17:19 |
| dtantsur | https://zuul.opendev.org/t/openstack/build/1b381b57ef304017b772da1b91d7839a WUT | 17:22 |
| dtantsur | https://1183a60a708f5672e1c1-38c49964a79149719549049b602122d6.ssl.cf1.rackcdn.com/876409/1/check/openstack-tox-py310/1b381b5/job-output.txt | 17:23 |
| samuelkunkel[m] | Am I assuming correctly that I need a AMQP based ironic setup to use the ironic prometheus exporter? From scrolling through the docs it reads like that | 17:24 |
| samuelkunkel[m] | s/a/an/ | 17:24 |
| samuelkunkel[m] | (We use json_rpc currently) | 17:24 |
| rpittau | dtantsur: remove skisdist | 17:25 |
| rpittau | skipsdist | 17:25 |
| dtantsur | samuelkunkel[m]: I don't think so, but I'll let iurygregory answer | 17:25 |
| dtantsur | rpittau: thanks!! | 17:25 |
| rpittau | sorry, put +2 and missed that | 17:25 |
| TheJulia | samuelkunkel[m]: no, it loads as a notifier plugin and afaik can be used without | 17:25 |
| samuelkunkel[m] | That sounds good, thanks! | 17:26 |
| TheJulia | dtantsur: wut indeed | 17:26 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/21.2: Fixes for tox 4.0 https://review.opendev.org/c/openstack/ironic/+/876409 | 17:26 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/21.2: Configure CI for bugfix/21.2 https://review.opendev.org/c/openstack/ironic/+/876410 | 17:26 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/21.2: Do not move nodes to CLEAN FAILED with empty last_error https://review.opendev.org/c/openstack/ironic/+/876174 | 17:27 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/21.2: Update .gitreview for bugfix/21.2 https://review.opendev.org/c/openstack/ironic/+/867826 | 17:27 |
| iurygregory | samuelkunkel[m], AMQP is not required | 17:29 |
| dtantsur | I'd still love if somebody could take a quick look at https://review.opendev.org/c/openstack/ironic/+/875944 and tell me in which parts I'm too crazy even for me :) | 17:34 |
| * TheJulia takes a stick and pokes diskiamge-builder | 17:35 | |
| dtantsur | is it too early in your day to suggest a flame-thrower? | 17:36 |
| rpittau | bye everyone, have a great weekend! o/ | 17:37 |
| dtantsur | o/ | 17:37 |
| dtantsur | I'm going as well, enjoy the weekend | 17:38 |
| samuelkunkel[m] | Thanks iurygregory . Is that also the case for the upcoming addition of ironic internal metrics? | 17:39 |
| TheJulia | guestfish: cannot mix --ro and --rw options <-- i just feel like drinking | 17:45 |
| TheJulia | dtantsur: i dunno, it could be good idea | 17:45 |
| TheJulia | samuelkunkel[m]: yes | 17:46 |
| iurygregory | samuelkunkel[m], correct (but you will need to use the antelope version for ironic and for IPE) | 17:55 |
| opendevreview | Merged openstack/ironic bugfix/19.0: Do not move nodes to CLEAN FAILED with empty last_error https://review.opendev.org/c/openstack/ironic/+/876179 | 17:56 |
| opendevreview | Merged openstack/ironic bugfix/20.2: Do not move nodes to CLEAN FAILED with empty last_error https://review.opendev.org/c/openstack/ironic/+/876177 | 18:04 |
| opendevreview | Merged openstack/ironic stable/zed: Do not move nodes to CLEAN FAILED with empty last_error https://review.opendev.org/c/openstack/ironic/+/876175 | 18:04 |
| opendevreview | Merged openstack/ironic bugfix/21.3: Configure CI for bugfix/21.3 https://review.opendev.org/c/openstack/ironic/+/876264 | 18:04 |
| opendevreview | Merged openstack/ironic master: Add Yoga versions to release notes https://review.opendev.org/c/openstack/ironic/+/855042 | 18:04 |
| TheJulia | dtantsur: done | 18:08 |
| JayF | > Servers can only have one owner. It is not easy to trade servers between different tenants. | 18:17 |
| JayF | false | 18:17 |
| JayF | lol | 18:17 |
| * JayF is catching up | 18:17 | |
| TheJulia | JayF: I thought the same thing, and I'm fairly sure dissertations have been done on that subject | 18:18 |
| JayF | This honestly reads like a ... shallow understanding of what already exists and where the pitfalls lie | 18:18 |
| TheJulia | I had a similar impression, in that it was clearly written from their context, and in their view this is it | 18:34 |
| JayF | we should try to reach out and talk to them | 18:34 |
| JayF | open source to me, part of it is about trying to reduce duplication of effort and if we can help them, or convince them to use ironic tools as a part of their project, all the better | 18:35 |
| JayF | hell, maybe another user of virtualbmc/sushy-tools/virtualpdu | 18:35 |
| samuelkunkel[m] | iurygregory sounds good. Once Antelope is released I will give it a try. Is the versioning of the IPE aligned with the openstack versioning? Or is it more like sushy / ipa where certain versions are aligned with certain openstack release versions? | 18:50 |
| JayF | IPA and IPE are both 'cycle-with-intermediary' which means they get a release cut at least once when the openstack integrated release is cut | 18:54 |
| JayF | so there should be a version of it released and maintained directly corrolating to the OpenStack 2023.1 (aka Antelope) release. | 18:54 |
| JayF | (I think sushy is an independent library? but I'm not sure) | 18:54 |
| TheJulia | sushy I think is independent | 18:59 |
| TheJulia | but we still need to cut releases ourselves | 19:00 |
| TheJulia | for ipe/ipa | 19:00 |
| jlvillal | Does Ironic have a IPA kernel/ramdisk based on aarch config option? Kind of like there is a ipxe_bootfile_name_by_arch option. | 19:05 |
| JayF | We talked about that, IDR if we actually did it | 19:06 |
| jlvillal | But this is probably just showing my lack of knowledge on how things are supposed to be done. | 19:06 |
| JayF | lemme check release notes | 19:06 |
| jlvillal | Thanks! | 19:07 |
| TheJulia | it is not there | 19:07 |
| JayF | https://files.openstack.org/docs/ironic/victoria/install/configure-pxe.html#pxe-multi-architecture-setup | 19:07 |
| JayF | oh, that's pxe bootfile, not ipa | 19:08 |
| JayF | dang | 19:08 |
| JayF | jlvillal: https://etherpad.opendev.org/p/ironic-bobcat-ptg line 39 | 19:08 |
| jlvillal | Thanks. No worries. Sounds like I should configure the json/yaml file in bifrost with ipa_kernel_url | 19:08 |
| JayF | jlvillal: it's on the schedule to discuss at PTG | 19:08 |
| TheJulia | realistically, given nodes can vary so much, I would generally just expect they be set on the node level, the general options in ironic.conf are really only for more general user cases | 19:08 |
| JayF | yeah, you can absolutely do it per node, but it's not-ideal to have any "run an API call N times for N nodes" which is what that creates | 19:09 |
| JayF | alias-based assignment, as laid out in that ehterpad, is likely the route to go | 19:09 |
| JayF | probably not that complex either? we should JFDI for bobcat | 19:09 |
| jlvillal | Okay. Thanks all :) I think I know what I need to do for my use case. I'm hoping I can set it in the JSON I pass to `./bifrost-cli deploy` | 19:12 |
| TheJulia | yeah, I'd rather just review a patch than discuss it ;) | 19:12 |
| TheJulia | there is likely a little more complexity in the option handling that might be refactoring | 19:12 |
| TheJulia | but I don't know if that is really bikeshedable | 19:12 |
| jlvillal | everything is bikesheddable ;) | 19:13 |
| jlvillal | hehe | 19:13 |
| * TheJulia gets out the bikeshed paint can labeled "purple" | 19:13 | |
| * TheJulia suggests we debate "purple" | 19:13 | |
| TheJulia | ;) | 19:13 |
| jlvillal | :) | 19:15 |
| JayF | I don't believe in purple. Can we find some empirical evidence this 'purple' actually exists? | 19:19 |
| TheJulia | it is purple, it doesn't need to define itself | 19:21 |
| JayF | That's exactly the attitude I'd expect from a supposed royal color | 19:22 |
| JayF | lol | 19:22 |
| TheJulia | The snails gave their lives for the purple! | 19:22 |
| * TheJulia suggests next debate is "green", but not to the extent the drazi went to with green and yellow. | 19:26 | |
| jlvillal | Hmmm. So I set the values `ipa_kernel_url` and `ipa_ramdisk_url` in the JSON file that I passed to `./bifrost-cli deploy`. But it seemed to ignore those values and in the `/var/lib/ironic/httpboot/` directory it created a UUID directory and the config file is pointing at the default `ipa.kernel` file which was copied into the UUID directory. :( | 19:27 |
| JayF | TheJulia: wasn't it purple and red? | 19:28 |
| JayF | if I've forgotten it's clearly time for a rewatch lol | 19:29 |
| jlvillal | Maybe this will make more sense than my comment: https://paste.opendev.org/show/b2xesiwUW7TOzDQEZvAT/ | 19:29 |
| JayF | green and purple lol | 19:29 |
| TheJulia | oh, it was purple! | 19:30 |
| JayF | jlvillal: I'm looking but bifrost is a weak spot so maybe someone else will know off the top of their head | 19:31 |
| TheJulia | see! purple exists! | 19:31 |
| JayF | dang it, you got me | 19:31 |
| JayF | but it only exists in sci-fi | 19:31 |
| JayF | clearly :D | 19:31 |
| JayF | it's a bike shed, not a spaceship, we can't paint it a space color :P | 19:31 |
| JayF | jlvillal: can you share the JSON, redacted if needed? | 19:32 |
| * TheJulia urges purple, in accordance with the KSP2 agency color editor | 19:32 | |
| * TheJulia gets does her best corgi side-eye impression with diskimage-builder | 19:33 | |
| jlvillal | JayF, Sure. https://paste.opendev.org/show/byi0dSAa0qJx5cR72lUi/ | 19:33 |
| TheJulia | jlvillal: under driver_info, deploy_ramdisk and deploy_kernel keys can be set | 19:34 |
| TheJulia | and those can just be URLs | 19:34 |
| JayF | yeah AFAICT in bifrost, ipa_kernel_url and the like can only be used to set it globally, not per node | 19:35 |
| JayF | so I think you need to move those two keys into driver_info and you'll be set | 19:35 |
| JayF | so bifrost doesn't do anything w/them except plug them into ironic | 19:35 |
| jlvillal | Thanks. I was trying to follow this example: https://github.com/openstack/bifrost/blob/0e6be25ee17ea75d60eb4f32fea37db0f79af52d/playbooks/inventory/baremetal.json.example#L28-L29 | 19:35 |
| jlvillal | But I can update that to deploy_kernel/ramdisk | 19:36 |
| jlvillal | I will try your suggestion! | 19:36 |
| TheJulia | huh... | 19:36 |
| jlvillal | huh about what part? | 19:37 |
| TheJulia | I just don't remember that change | 19:37 |
| JayF | that example appears broken | 19:37 |
| JayF | or I'm bad at reading ansible (very possible) | 19:37 |
| jlvillal | As a note it is also in the example YAML. https://github.com/openstack/bifrost/blob/0e6be25ee17ea75d60eb4f32fea37db0f79af52d/playbooks/inventory/baremetal.yml.example#L22-L23 | 19:39 |
| TheJulia | so in theory, I guess it could be used, but in practice I don't see it offhand | 19:39 |
| TheJulia | yeah, looking at the original change it wouldn't have bene a thing | 19:40 |
| TheJulia | been | 19:40 |
| jlvillal | Woohoo! Success with `deploy_kernel` and `deploy_ramdisk` :) | 19:45 |
| JayF | \o/ | 19:46 |
| TheJulia | if your using yaml you could also template the key and just rinse/repeat on a variable, fwiw | 19:50 |
| jlvillal | I'm using JSON at the moment. But will keep that in mind. Thanks. | 19:51 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!