| opendevreview | Merged openstack/networking-generic-switch master: Add default vlan to driver interfaces https://review.opendev.org/c/openstack/networking-generic-switch/+/962037 | 00:08 |
|---|---|---|
| opendevreview | Clif Houck proposed openstack/ironic master: WIP: Generate network plan based on trait based networking config https://review.opendev.org/c/openstack/ironic/+/964895 | 00:59 |
| rpittau | good morning ironic! o/ | 06:27 |
| Ayo[m] | Good morning rpittau :) | 07:05 |
| Ayo[m] | Good morning ironic! | 07:05 |
| rpittau | hey Ayo[m] :) | 07:39 |
| Ayo[m] | We’re still eagerly waiting on that update😅 rpittau | 07:50 |
| rpittau | Ayo[m]: sorry, which update? too many things this week! | 07:51 |
| Ayo[m] | rpittau: outreachy project | 07:53 |
| rpittau | Ayo[m]: we're going to send out an email today with some updates | 07:54 |
| Ayo[m] | Great!! | 07:54 |
| Ayo[m] | rpittau: i just had to ask because I could see that a whole lot has been going on | 07:56 |
| rpittau | Ayo[m]: yeah, this week we have the PTG and it's taking a lot of our time | 07:57 |
| Ayo[m] | PTG? I’ve been seeing this a lot in the group? Don’t really know what it is tho | 07:58 |
| rpittau | Project Team Gathering, it's a week long of (virtual) meetings focused on planning the next release cycle | 07:59 |
| Ayo[m] | Ok | 08:01 |
| Ayo[m] | Is there a deadline or time limit for each release tho? or it gets released based on issue updates? | 08:04 |
| Ayo[m] | I’m new to this, please don’t mind my many questions rpittau | 08:05 |
| rpittau | Ayo[m]: no worries! We have 2 release cycles per year, roughly 6 months each, that end with a release, usually at the end of March and September | 08:06 |
| rpittau | currently we're in the Gazpacho release cycle https://releases.openstack.org/gazpacho/schedule.html | 08:06 |
| Ayo[m] | Expectations are really detailed rpittau | 08:15 |
| opendevreview | Riccardo Pittau proposed openstack/ironic-python-agent master: Add retry logic for heartbeat connection errors https://review.opendev.org/c/openstack/ironic-python-agent/+/965201 | 09:20 |
| rpittau | hopefully this makes sense ^ | 09:20 |
| dtantsur | rpittau: hmm, I'm not sure. The expectation is that any heartbeat is retried anyway, why do we need more retry logic? | 10:48 |
| dtantsur | rpittau: https://opendev.org/openstack/ironic-python-agent/src/branch/master/ironic_python_agent/agent.py#L151-L160 | 10:49 |
| rpittau | dtantsur: yep, I think I lost that logic! | 10:50 |
| iurygregory | good morning ironic | 11:07 |
| iurygregory | TheJulia, JayF cardoe https://paste.opendev.org/show/bubmOHDoL8kjsZLAOqgg/ does it look ok ? | 11:08 |
| dtantsur | cid, iurygregory and everyone interested: I've started the async WG etherpad https://etherpad.opendev.org/p/ironic-asyncio | 11:56 |
| iurygregory | tks dtantsur | 11:56 |
| cid | ack | 11:58 |
| TheJulia | iurygregory: I'd simplify and soften what you've written. "As the library was never intended for others to consume, yet that has happened, and it being a library has been a noted hurdle to quickly resolving user issues." | 12:00 |
| TheJulia | rpittau: I guess I'm curious what spawned that change, granted the existing retry logic is a bit heavy on the leaning towards backing-off if there is an issue (at least, that was the impression burned into my brain from ages ago) | 12:08 |
| iurygregory | TheJulia, ohhh I though I shouldn't directly mention that! | 12:11 |
| TheJulia | I think its fine to mention a full context | 12:12 |
| TheJulia | more than anything, we want to make just enough noise to cause people to come to the table to discuss | 12:12 |
| iurygregory | gotcha! | 12:13 |
| iurygregory | TheJulia, only [ironic] in the subject is ok? or should we add something more? | 12:26 |
| TheJulia | I think that is fine | 12:30 |
| JayF | iurygregory: ++ I was going to make the same suggestion as Julia, to be explicit that our support policy has always been that we don't support it | 12:39 |
| iurygregory | JayF, I've changed towards what Julia suggested, the email is out \o/ | 12:39 |
| iurygregory | now grab coffee and wait | 12:39 |
| iurygregory | :D | 12:39 |
| TheJulia | coffee coffee coffee coffee coffee | 12:55 |
| iurygregory | coffee \o/ | 12:56 |
| iurygregory | PTG starting | 13:00 |
| rpittau | TheJulia: a bug that I saw downstream :) | 13:17 |
| TheJulia | rpittau: a slightly better thing is maybe capture the error and *not* raising a backtrace | 13:20 |
| TheJulia | RemoteDisconnected is ... interesting | 13:20 |
| TheJulia | transparent proxy?! | 13:20 |
| * dtantsur back in ~ 30 mins | 13:29 | |
| Mahnoor | I'm wondering if PTGs are recorded? Just in case someone wants to refer to a topic's discussion later... | 13:35 |
| iurygregory | Mahnoor, it's not recorded | 13:55 |
| TheJulia | rpittau: didn't entirely read, but I do wonder if its really the root cause or just an episode of "I seen an error, I blame this error" | 13:56 |
| iurygregory | normally people would use the etherpad to refer to something that was discussed | 13:56 |
| rpittau | TheJulia: no, I was not thinking that that was the actual root cause, it just made me questioning if we had a retry logic for the heartbeat, but it was a side thought | 13:58 |
| TheJulia | rpittau: ahh, yeah, I don't think logging is handled the same way with the different exception | 14:01 |
| TheJulia | looks like a MitM proxy?! | 14:01 |
| rpittau | mmm could be | 14:03 |
| cardoe | Sorry. I fail at wrapping my head around Zuul stuff. | 14:24 |
| cardoe | So I just didn't know how to fix it. | 14:25 |
| cardoe | clif: https://review.opendev.org/c/openstack/ironic/+/964570 so I wanted to do the portgroup correctly layered after your change. | 14:26 |
| clif | will look | 14:27 |
| opendevreview | Pierre Riteau proposed openstack/ironic-python-agent-builder stable/2025.2: Publish CentOS Stream 10 IPA dib images https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/965242 | 14:38 |
| opendevreview | Steve Baker proposed openstack/ironic master: Switch graphical console image to ubuntu by default https://review.opendev.org/c/openstack/ironic/+/964201 | 14:42 |
| opendevreview | Steve Baker proposed openstack/ironic master: Add a kubernetes provider for console container https://review.opendev.org/c/openstack/ironic/+/962242 | 14:42 |
| TheJulia | iurygregory: what room are we joining for the next topic? | 15:30 |
| iurygregory | TheJulia, everything in Mitaka | 15:30 |
| iurygregory | we don't need to leave | 15:30 |
| TheJulia | woohoo | 15:30 |
| TheJulia | quick! grab a screenshot | 15:31 |
| iurygregory | done | 15:31 |
| opendevreview | Riccardo Pittau proposed openstack/ironic master: Handle HTTP 409 race condition in Redfish power operations https://review.opendev.org/c/openstack/ironic/+/965248 | 15:54 |
| opendevreview | Jay Faulkner proposed openstack/ironic-specs master: Add work items template, including maint checklist https://review.opendev.org/c/openstack/ironic-specs/+/965249 | 16:11 |
| opendevreview | Riccardo Pittau proposed openstack/ironic master: Handle HTTP 409 race condition in Redfish power operations https://review.opendev.org/c/openstack/ironic/+/965248 | 16:14 |
| cardoe | The volume attach complete already works. If https://github.com/openstack/nova/blob/30bf8c1025ea24823ebde3eacede4ea34ec6be8c/nova/virt/driver.py#L756 doesn't throw an exception then the volume reports attached. | 16:23 |
| cardoe | https://github.com/rackerlabs/understack/blob/main/containers/nova/patches/ironic-attach-debug.patch that patch will make nova report the volume attachment being successful | 16:25 |
| cardoe | The real patch I'm using sets up the properties in the BIOS in that call. | 16:26 |
| TheJulia | I'd just wire on the ironic side to say done. | 16:34 |
| TheJulia | at least, from my point of view, its not different from the host being rebooted | 16:34 |
| TheJulia | or the vm not really started | 16:34 |
| JayF | None of this is as bad as us saying "ACTIVE" when we power on instead of when OS boots | 16:35 |
| JayF | in terms of lying to other apis lol | 16:35 |
| dtantsur | heh | 16:58 |
| dtantsur | With Redfish, we can at least try fixing that | 16:58 |
| * dtantsur has also been pondering an idea of optional in-instance agents | 16:59 | |
| TheJulia | I love how ideas return | 16:59 |
| TheJulia | :) | 16:59 |
| JayF | -1.5 | 17:01 |
| JayF | loll | 17:01 |
| JayF | you get a half point of downvote reduction over 10 years | 17:01 |
| JayF | maybe in 2035 /s | 17:02 |
| cardoe | TheJulia: soooo VXLAN to VMs... | 17:05 |
| cardoe | If ya got like 60 seconds | 17:05 |
| TheJulia | sure | 17:07 |
| TheJulia | cardoe: https://meetpad.opendev.org/oct2025-ptg-ironic | 17:08 |
| cardoe | https://github.com/ovn-org/ovn/blob/557020bac583d60ebcd6def3f1473180d5427a5c/NEWS#L248 | 17:53 |
| opendevreview | Merged openstack/ironic-python-agent-builder stable/2025.2: Publish CentOS Stream 10 IPA dib images https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/965242 | 17:54 |
| cardoe | https://bugs.launchpad.net/neutron/+bug/2114451 is the bug for https://review.opendev.org/c/openstack/neutron/+/840418 which won't select the right segment. | 18:14 |
| cardoe | TheJulia/clif: oh. so today Ironic selects the port based on the physical_network field but let's say my new EVPN VXLAN driver lands. I've kept the physical_network field as the VLAN pool/range/cabinet... I would need to then potentially have traits or something to match the EVPN VXLAN physical_network to the right ports. That's what I meant to bring up. | 22:03 |
| TheJulia | so, Idealy I'd split the desired physical_network and then subdivide it further based upon delimiters. I guess the weirdness, at least that I've felt is the range/cabinet being part of that mix | 22:07 |
| TheJulia | I get why, but I need it on a diagram | 22:07 |
| cardoe | I'm coloring as we speak. | 22:07 |
| cardoe | Fed my first draft of docs and diagrams to Claude and told it to give me constructive feedback one statement or question at a time so that I can revise. | 22:08 |
| * TheJulia expects crayon | 22:08 | |
| TheJulia | :) | 22:08 |
| TheJulia | cool cool | 22:08 |
| TheJulia | I'm going to go for an early (happy hour at the local bar) dinner and crash because of jetlag | 22:09 |
| cardoe | That was one of the things it pointed out that I wasn't very clear between the two physical_networks | 22:09 |
| TheJulia | I mean, at some point the same basic idea applies, but in theory not as much because you can't tie it on the neutron/vm side really | 22:12 |
| cardoe | We could call it "leaf_id" if ya want. | 22:20 |
| cardoe | We need the physical_network the port is on as well as the leaf its on. | 22:20 |
| JayF | for the most part, in TBH, physical_network becomes less relevant, yeah? | 22:41 |
| JayF | because that's the simple match case | 22:41 |
| cardoe | Well in my case I've got 2 ports hooked to a network where tenants can make networks. Then I've got 2 ports that are utilized for the storage traffic. They're completely separate fabrics (the name of a leaf-spine network of switches that have interconnection). I can have VNI 10000 on both of them. But they're different instances of VNI 10000. | 22:44 |
| cardoe | I'd like to use neutron to IP the Redfish side of the storage traffic | 22:46 |
| cardoe | I don't want Ironic to try and attach regular traffic to the ports that are for storage either. | 22:46 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!