| kaloyank | morning Ironic | 07:06 |
|---|---|---|
| dtantsur | JayF: 17:00 UTC is problematic for me, but I can make it, I guess | 08:59 |
| dtantsur | JayF: otherwise, the schedule looks good to me | 09:03 |
| rpittau | good morning ironic! o/ | 09:05 |
| kaloyank | generic question, is this https://docs.openstack.org/ironic/latest/contributor/contributing.html#feature-submission-process still valid? I want to submit an RFE for saving local disk deployments to images | 09:14 |
| dtantsur | kaloyank: this is more or less up-to-date, yes | 09:18 |
| dtantsur | have you seen the previous attempt on snapshots in ironic? | 09:19 |
| kaloyank | nope, it'd be interesting if you can share some links :) | 09:19 |
| kaloyank | I found this: https://specs.openstack.org/openstack/ironic-specs/specs/not-implemented/snapshot-support.html | 09:33 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic-specs master: [WIP] Merge Inspector into Ironic https://review.opendev.org/c/openstack/ironic-specs/+/878001 | 09:36 |
| dtantsur | kaloyank: yeah, I meant that | 09:36 |
| kaloyank | if I understand correctly, the spec just needs to be implemented? | 09:40 |
| opendevreview | Riccardo Pittau proposed openstack/sushy stable/zed: Exclude all files starting with . from flake8 tests https://review.opendev.org/c/openstack/sushy/+/878018 | 10:17 |
| opendevreview | Riccardo Pittau proposed openstack/sushy stable/yoga: Exclude all files starting with . from flake8 tests https://review.opendev.org/c/openstack/sushy/+/878019 | 10:17 |
| opendevreview | Riccardo Pittau proposed openstack/sushy stable/xena: Exclude all files starting with . from flake8 tests https://review.opendev.org/c/openstack/sushy/+/878020 | 10:18 |
| opendevreview | Riccardo Pittau proposed openstack/ironic master: Use main branch of metal3-dev-env to run metal3 integration job https://review.opendev.org/c/openstack/ironic/+/877600 | 11:05 |
| opendevreview | Riccardo Pittau proposed openstack/ironic master: Use main branch of metal3-dev-env to run metal3 integration job https://review.opendev.org/c/openstack/ironic/+/877600 | 11:07 |
| iurygregory | good morning Ironic | 11:19 |
| opendevreview | Merged openstack/ironic-python-agent-builder master: Add DIB_IPA_HARDWARE_RDO to define repo behaviour https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/877567 | 12:41 |
| TheJulia | good morning | 13:01 |
| TheJulia | kaloyank: yes, but going down that road, it doesn't have to be an exact match to the spec. you may learn things along the way! | 13:04 |
| TheJulia | We can also change the spec if new ideas surface | 13:04 |
| TheJulia | Some aspects have changed since that was written, we might be able to find a way to get authentication data out of the agent | 13:06 |
| dtantsur | morning TheJulia | 13:27 |
| dtantsur | TheJulia, JayF, arne_wiebalck, 2023Q2 is around the corner, who is going to schedule the next BM SIG meetup? | 13:32 |
| TheJulia | I guess the question which comes to mind is what would we like to present/engage in regards to? Would this be an opportunity to present some of the under-discussion ideas (as long as we preface them as such)? | 14:00 |
| JayF | dtantsur: if you have specific suggestions how I can shift things around to make your life easier; please let me know (re: vPTG) | 14:08 |
| dtantsur | JayF: nothing specific. If we manage to finish by 16:30 UTC, it will be much better for me, but it's fine. | 14:09 |
| dtantsur | TheJulia: "Things I learned by first creating ironic-inspector and then destroying it again" :D | 14:10 |
| sschmitt | I've got a question about this spec: https://specs.openstack.org/openstack/ironic-specs/specs/9.0/physical-network-awareness.html | 14:26 |
| sschmitt | it seems to be fully implemented, but there are some usecases it seems like it doesn't support. For one this seems to work only for vlan networks. We'd like to do it with vxlan networks (assuming we have some ml2 mechanism to setup the vxlan vteps on the top of rack switches) | 14:28 |
| dtantsur | I'm not sure anything on the ironic side is actually preventing vxlan from working? Or even: I've heard people confidently saying "it works if the ML2 driver and the switch do the right thing". | 14:29 |
| sschmitt | Secondly, as far as im aware you must specify these physical networks in your ml2 conf file. Which means if you want to add one (like if you got a new bm node that has 8 interfaces but you only configured 6 physnets), you have to restart neutron | 14:29 |
| sschmitt | i think this also means that they are implicitly provider networks, which isnt ideal from an RBAC perspective | 14:30 |
| sschmitt | @dtantsur you might be right, i know its hard coded in NGS for vlan. It also might on the neutron side as vxlan provider networks claim to be supported, but i can find almost no info on them | 14:32 |
| dtantsur | I'll tag hjensas and mgoddard who know much more about these things | 14:32 |
| TheJulia | sschmitt: I think the model was largely for where you had distinct networking fabrics and to scope the fabric based and help map port->fabric->neutron's connectivity for neutron | 15:46 |
| TheJulia | unfortunately the term fabric has other means | 15:48 |
| TheJulia | meanings | 15:48 |
| * TheJulia can't type today | 15:48 | |
| TheJulia | dtantsur: That sounds like a whorthwhile presentation | 15:48 |
| TheJulia | dtantsur: I think talking about RBAC would actually be a good idea. I've been on the scientific-sig slack a few rbac related questions have come up (besides, I have a talk on this subject too..) | 15:49 |
| sschmitt | yeah i think that works for some, but i think "put this network on this baremetal interface, otherwise fail" is also important | 15:49 |
| sschmitt | especially when you want mixed vm and baremetal topologies | 15:50 |
| TheJulia | sschmitt: yeah, If you want to implicitly direct it yeah. The physnet awareness was always intended so the humans didn't have to be concious of it | 15:50 |
| sschmitt | we get around this today by scheduling a bunch of dummy networks until we get the right network on the right interface, which is less than ideal | 15:52 |
| TheJulia | ouch :( | 15:53 |
| hjensas | sschmitt: In neutron "physical_network" is only a thing for provider networks. (provider:physical_network) | 16:01 |
| hjensas | sschmitt: I think if don't intend to use provider networks there is a gap in functionality. | 16:03 |
| hjensas | sschmitt: also, reading backlog - In theory using vxlan should work if the is an ML2 that can configure vtep on the switch. (But I'm not aware of any plug-in that does) | 16:05 |
| sschmitt | so it sounds like a new spec for functionality to explicitly hook up a neutron tenant network to a baremetal port would be needed | 16:07 |
| hjensas | sschmitt: yes, I think this would be an RFE, ideas: 1. additional field in the binding-profile 2. a tag on the neutron port (I think binding-profile would be preffered) | 16:12 |
| sschmitt | yeah i think binding profile would be the best as well | 16:14 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic-specs master: [WIP] Merge Inspector into Ironic https://review.opendev.org/c/openstack/ironic-specs/+/878001 | 16:54 |
| dtantsur | This is becoming an interesting document ^^^. Essentially, the whole Inspector (except for the introspection rules) explained as a text :D | 16:54 |
| rpittau | I'm too tired to read that now, but wow! | 16:58 |
| TheJulia | I'm almost afraid?! | 17:00 |
| dtantsur | C'mon, I tried to keep it easy and entertaining! | 17:06 |
| dtantsur | don't make me regret even starting a spec :D | 17:06 |
| JayF | My eyes are glazed over and all I've read is the url ;) | 17:06 |
| JayF | seriously, thanks for writing that up and I'll open it up and read it today at some point :D | 17:07 |
| dtantsur | Thanks! It's WIP but it's actually almost-almost complete. | 17:07 |
| rpittau | good night! o/ | 17:10 |
| * TheJulia hears of virtualization inside of firmware levels of code interaction utilizing linux running on cards in firmware on cards before the main host OS boots and wonders how many levels of nesting is taking place. | 17:23 | |
| dtantsur | "We need to go deeper" (c) | 17:26 |
| TheJulia | I suspect that is always the pattern. The substrate is often hidden as you can only observe the floor (for lack of a better word) | 17:36 |
| kubajj | dtantsur: am I safe to quote this in my dissertation? This seems kind of relevant | 18:32 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!