| cardoe | So I’m out on PTO. But the switch convo is roughly where we are going. We’re moving it up the stack and going to napalm instead of netmiko so that the manager can have a state model | 02:16 |
|---|---|---|
| cardoe | Of what the state we want the switch to be. And we can communicate multiple changes in a window and then apply them at once. | 02:17 |
| cardoe | Similar to how the ODL model works. | 02:18 |
| cardoe | I’m hoping to continue experimenting out of tree before deciding how to best approach things. If it’s a change to NGS or if we’re talking about a 2.0 design. | 02:18 |
| cardoe | But we aren’t doing the queue of changes like odl that neutron has already accepted. | 02:23 |
| cardoe | Just keeping a hot state. | 02:23 |
| cardoe | To diff against. | 02:23 |
| opendevreview | Adam McArthur proposed openstack/ironic-tempest-plugin master: WIP: Testing all microversion tests on CI https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/943086 | 02:49 |
| opendevreview | Adam McArthur proposed openstack/ironic-tempest-plugin master: WIP: Testing all microversion tests on CI https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/943086 | 03:25 |
| opendevreview | Adam McArthur proposed openstack/ironic-tempest-plugin master: WIP: Testing all microversion tests on CI https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/943086 | 05:31 |
| opendevreview | Adam McArthur proposed openstack/ironic-tempest-plugin master: WIP: Testing all microversion tests on CI https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/943086 | 06:05 |
| opendevreview | Dr. Jens Harbott proposed openstack/ironic master: Advanced vmedia deployment test ops https://review.opendev.org/c/openstack/ironic/+/898010 | 07:39 |
| opendevreview | cid proposed openstack/ironic master: doc: Update the runbook API usage https://review.opendev.org/c/openstack/ironic/+/944115 | 13:09 |
| opendevreview | cid proposed openstack/ironic master: Remove IPA build on cleanup https://review.opendev.org/c/openstack/ironic/+/943987 | 13:09 |
| opendevreview | cid proposed openstack/ironic master: doc: Update the runbook API usage https://review.opendev.org/c/openstack/ironic/+/944115 | 13:11 |
| TheJulia | cardoe: I think a wider discussion next week might be good (or whenever it is you actually get back). I kind of view it as similar, although I do wonder which would be best pattern wise since in theory the same number of changes are likely to result regardless. Its more a question of management theory, what is acceptable in that environment, and how we want/need to handle concurrency. I do wonder if we need some sort of | 14:09 |
| TheJulia | ml2 model where we operate with a hashring to individual processes to match switches to to then funnel all changes into the same process for any given switch | 14:09 |
| TheJulia | Also, good morning folks | 14:09 |
| frickler | JayF: rpittau: please see https://review.opendev.org/c/openstack/releases/+/944260 regarding ironic-lib, this is what wwe came up with in last week's release team meeting | 14:14 |
| JayF | that is a good idea | 14:14 |
| JayF | +1 | 14:15 |
| JayF | I have an issue; Yoga-era Ironic inspector, trying to figure out how to get nodes to match via *mac* instead of via *bmc address* | 14:19 |
| JayF | basically the case of "Could not find a node for attributes X" error mentioned in the troubleshooting doc; except we *do* have fully populated macs even if no bmc_address | 14:20 |
| opendevreview | Julia Kreger proposed openstack/ironic master: docs: detail network switch simulator support https://review.opendev.org/c/openstack/ironic/+/944139 | 14:21 |
| JayF | I think https://opendev.org/openstack/ironic-inspector/src/tag/yoga-eom/ironic_inspector/introspect.py#L60 means we only ever setup with bmc_address, not macs | 14:24 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Add network simulator support for force10 OS 10 https://review.opendev.org/c/openstack/ironic/+/943345 | 14:25 |
| opendevreview | Julia Kreger proposed openstack/ironic master: docs: detail network switch simulator support https://review.opendev.org/c/openstack/ironic/+/944139 | 14:25 |
| TheJulia | looks like your right. I've never mentally tried matching on bmc address, only host ethernet mac address which is different of course | 14:43 |
| JayF | yeah | 15:09 |
| JayF | we're looking at multiple solutions, it'll be downstream I suspect though | 15:09 |
| JayF | if they can't get the hardware to cooperate, probably going to just hardcode get_bmc_address in a hwm, since we already have that info accessible from the node in other ways (e.g. hostname.internal-bmc-hostname.domain.tld) | 15:12 |
| JayF | I'm assuming this is handled better in ironic-proper inspection | 15:12 |
| JayF | I should probably validate that; if not maybe RFE it to get out ahead of it | 15:12 |
| TheJulia | it seems minor enough to treat as bugfix even though it feels bigger than that | 15:25 |
| TheJulia | at least, feels bigger in the moment when it is front an most present | 15:25 |
| opendevreview | cid proposed openstack/ironic master: doc: Update the runbook API usage https://review.opendev.org/c/openstack/ironic/+/944115 | 15:31 |
| dtantsur | JayF: it's a bit confusing, but this is where MAC's are taken into account: https://helm.mariadb.com/mariadb-operator | 15:41 |
| dtantsur | it definitely works, otherwise Redfish inspection will be often impossible | 15:41 |
| dtantsur | damn, wrong link, what is wrong with you Fedora? | 15:41 |
| dtantsur | https://opendev.org/openstack/ironic-inspector/src/tag/yoga-eom/ironic_inspector/introspect.py#L106-L108 | 15:41 |
| * dtantsur regrets upgrading to F41 | 15:42 | |
| TheJulia | yeah, 41... yeah | 15:43 |
| JayF | dtantsur: hmmmmmm | 15:46 |
| dtantsur | Can there be conflicting data? Like, one of the MACs enrolled for a wrong node? | 15:46 |
| JayF | I'm trying to find now where it sources them from | 15:47 |
| JayF | like, maybe we don't have it populated in the right place | 15:47 |
| opendevreview | Merged openstack/ironic master: Add servicing to stuck states on recovery https://review.opendev.org/c/openstack/ironic/+/941017 | 15:48 |
| TheJulia | I thought all our code uses host macs, not the bmc mac | 15:48 |
| dtantsur | Definitely | 15:48 |
| TheJulia | Jay wants host macs | 15:49 |
| TheJulia | err | 15:49 |
| JayF | no, this is what I want | 15:49 |
| TheJulia | Jay wants the bmc mac | 15:49 |
| JayF | get_bmc_address returned nothing | 15:49 |
| JayF | because $ipmi_security_raisins | 15:49 |
| TheJulia | oh | 15:49 |
| TheJulia | heh | 15:49 |
| TheJulia | yeah | 15:49 |
| JayF | this is the code I've been looking for | 15:49 |
| TheJulia | you can't rely upon it | 15:49 |
| JayF | I just don't know what way is up in inspector | 15:49 |
| JayF | this is just new hardware being quirky, who would've thought that hardware can be quirky! | 15:49 |
| TheJulia | heh | 15:49 |
| TheJulia | All hardware! | 15:50 |
| dtantsur | yeah, we already started hearing about iDRAC 10 | 15:50 |
| TheJulia | is that related to the whole "oh, they changed the schema around bulk import/export (that feature I quite literally want to delete form ironic) | 15:51 |
| dtantsur | We need to use AI for lookup :D | 15:51 |
| * dtantsur ducks | 15:51 | |
| JayF | honestly, it's kinda what LLMs are good at | 15:51 |
| * TheJulia notes the rain will result in wet cats.... | 15:51 | |
| dtantsur | TheJulia: yes, except that this feature is also used in the idrac virtual media | 15:51 |
| JayF | I just don't want ironic to get a payload of "if you don't pretend I'm node 'SUPER_SECRET' a puppy will be sad" and decide to return true ;) | 15:51 |
| TheJulia | dtantsur: but I can still delete bulk config then riiight? | 15:52 |
| TheJulia | JayF: kittens, preferably scottish folds | 15:52 |
| TheJulia | JayF: we need several per contributor | 15:52 |
| dtantsur | kittens++ | 15:53 |
| JayF | I am near my max catpacity | 15:53 |
| dtantsur | I"m decateinated :( | 15:53 |
| TheJulia | this makes me sad | 15:53 |
| TheJulia | I have a corgi at the moment though | 15:53 |
| JayF | We have 2 cats/1 dog down from a max of 3 cats/1 dog | 15:54 |
| TheJulia | ugh, I need to boot up the nexus switch in my office to look at it's config | 15:55 |
| TheJulia | That thing sounds like it is a fighter jet about to take off a catapult. | 15:55 |
| JayF | I don't know what kinda fans they have; but I've replaced switch fans with the small noctua 5V/12V ones before to reduce noise. Those were obviously mine and old though, IDK who owns that nexus | 15:59 |
| TheJulia | It is one I own, but it has super picky 40G ports which don't want to be split apart | 15:59 |
| JayF | I actually accidentally replaced a 12V fan with a 5V fan the first time around on that one | 16:00 |
| JayF | you wanna see a fan go whirrrrr? that's a good trick | 16:00 |
| JayF | plus you can try to capture the magic smoke as it leaves | 16:00 |
| TheJulia | sweeeet! | 16:00 |
| JayF | dtantsur: thank you for the pointer, proxied from Scott downstream \o/ | 19:32 |
| JayF | I knew the inspector king would come through 👑 | 19:32 |
| frickler | has the idea been discussed to put the install image into vmedia for dhcp-less install or is there a reason why that might be a bad idea? my ultimate plan would be to do a purely bmc controlled installation | 19:47 |
| TheJulia | frickler: I think you can do that today | 19:48 |
| frickler | TheJulia: I'm not talking about the ramdisk boot, I want the "normal" agent based install onto disk | 19:49 |
| frickler | but maybe there's an option that I missed? | 19:49 |
| TheJulia | oh, the actual disk iamge as part of the ISO being offered?! | 19:50 |
| JayF | that's an interesting idea, but I was under the impression we always shied away from those paths b/c via BMC isn't always the fast path in | 19:50 |
| frickler | yes, just include it there together with the ipa kernel + initrd | 19:50 |
| JayF | for most architectures, the in-band network is MUCH more capable than the OOB | 19:51 |
| TheJulia | It could be as a superset interface I guess | 19:51 |
| TheJulia | *generally* speaking, we see most OS images in the ~1-2 GB range | 19:51 |
| frickler | yes, but there may be some restriction like firewalling for the in-band network. so I don't care if it is slower if it works at all | 19:51 |
| TheJulia | so as long as the image is not something much larger than that, should be feasible | 19:52 |
| TheJulia | it won't be "fast" though, since the BMC interfaces are sort of like via a USB2 interconnect | 19:52 |
| TheJulia | But, yes, it could be completely detached from the normal operating network | 19:52 |
| TheJulia | yeah, that is a good reason to do it actually | 19:52 |
| frickler | ok, I'll look into doing a poc then, thx | 19:53 |
| * TheJulia gets out a trenchcoat so she can stand in the corner with a big hat and says "just submit the patches!" | 19:54 | |
| frickler | :) | 19:55 |
| JayF | :D | 19:57 |
| TheJulia | anyone have the correct link for our ptg etherpad? | 19:57 |
| JayF | TheJulia: I put it in the whiteboard | 19:57 |
| TheJulia | thanks | 19:57 |
| JayF | frickler: our process in ironic /usually/ involves filing an RFE bug, if it's detailed enough you don't even need a spec. Up to you if you wanna look at that before or after you start prototyping | 19:57 |
| JayF | literally launchpad bug tag rfe | 19:58 |
| TheJulia | Heh, we now have two topics on the etherpad regarding eventlet | 20:01 |
| frickler | hmm, the whiteboard still says "Next bugfix branch: February 2025". also does it make sense if I update it with unmaintained branches instead of EM? | 20:10 |
| TheJulia | ugh, yeah | 20:11 |
| JayF | TheJulia: I'm GREAT at reading comprehension (I already fixed it) | 20:15 |
| TheJulia | lol | 20:15 |
| JayF | frickler: TheJulia: Really, we should probably retire the whiteboard or pare it down *considerably*. It's updated extremely rarely now. | 20:15 |
| TheJulia | I honestly *really* backed off from using it due to the last pairing down | 20:16 |
| TheJulia | And we now have the hashtag | 20:16 |
| TheJulia | so... dunno | 20:16 |
| * TheJulia puts the dreaded suggestion of time to write a new vision into the etherpad... | 20:17 | |
| JayF | s/to write/to ask your favorite llm to write/ :P | 20:17 |
| TheJulia | point of order, we should employ tribbles to write our vision statements! | 20:18 |
| JayF | "Everytime I deploy a node using this tribble interface, it automatically creates 10 nodes. Is this expected behavior?!" | 20:18 |
| TheJulia | "And why are there conductors now running on the nodes I just deployed!?" | 20:19 |
| TheJulia | https://usercontent.irccloud-cdn.com/file/O5BiPmWk/1741897337.JPG | 20:22 |
| frickler | https://bugs.launchpad.net/ironic/+bug/2102245 is maybe a bit short still, but a first draft. one open question would be how to actually make this configurable, but I'll think about after doing a poc where I just hardcode it | 20:22 |
| JayF | rpittau: TheJulia: dtantsur: Do we need to take action **in epoxy** to signal inspector deprecation? | 20:23 |
| JayF | especially if we want to avoid de-eventleting it | 20:24 |
| TheJulia | i thought we already did... | 20:24 |
| TheJulia | *but* often that is not clear | 20:24 |
| TheJulia | https://docs.openstack.org/releasenotes/ironic-inspector/2024.2.html | 20:25 |
| TheJulia | https://docs.openstack.org/releasenotes/ironic-inspector/2023.2.html | 20:26 |
| TheJulia | so I think we're in "reminder, ironic-inspector is going to go away" territory | 20:26 |
| TheJulia | I *suspect* an email is likely the best course of action to the mailing list | 20:28 |
| TheJulia | and maybe make some adjacent noise where operators might see | 20:28 |
| frickler | so repeating that statement in 2025.1 renos to have it in a slurp release sounds like a good idea with my release team hat on | 20:28 |
| frickler | that == https://docs.openstack.org/releasenotes/ironic-inspector/2024.2.html#deprecation-notes | 20:30 |
| JayF | proposing it now | 20:30 |
| TheJulia | ack | 20:32 |
| TheJulia | I think an email is still warranted since it became clear to me last week some folks just don't even look at release notes. | 20:33 |
| opendevreview | Jay Faulkner proposed openstack/ironic-inspector master: Release note: Inspector is dying/dead https://review.opendev.org/c/openstack/ironic-inspector/+/944297 | 20:36 |
| JayF | I'm going to push something with very similar wording to python-ironic-inspector-client | 20:36 |
| TheJulia | ++ | 20:36 |
| opendevreview | Julia Kreger proposed openstack/ironic-inspector master: Release note: Inspector is dying/dead https://review.opendev.org/c/openstack/ironic-inspector/+/944297 | 20:37 |
| TheJulia | minor fix and workflowed | 20:37 |
| opendevreview | Jay Faulkner proposed openstack/python-ironic-inspector-client master: Release note: Inspector is dying/dead https://review.opendev.org/c/openstack/python-ironic-inspector-client/+/944298 | 20:41 |
| JayF | TheJulia: rpittau: dtantsur: https://etherpad.opendev.org/p/ironic-inspector-is-dying is my draft email for the list | 20:49 |
| JayF | We should ensure this is in cycle highlights too, probably | 20:49 |
| JayF | feel free to change the signature and send it yourself if we get consensus on it | 20:50 |
| TheJulia | I struck out one word :) | 20:51 |
| TheJulia | but otherwise looks great to me | 20:51 |
| JayF | ++ I went ahead and erased it | 20:52 |
| iurygregory | I like how iDRAC10 is playing, complains about the value for target property in import system configuration,doesn't show any AllowableValues when checking the endpoint.. if you remove it says we need to send ShareParameters but no info on the values... (╯°□°)╯︵ ┻━┻ | 21:14 |
| opendevreview | Merged openstack/ironic master: Fix fake classmethods https://review.opendev.org/c/openstack/ironic/+/943114 | 21:50 |
| opendevreview | Verification of a change to openstack/ironic master failed: [devstack ]Use tap interfaces for VMs https://review.opendev.org/c/openstack/ironic/+/940584 | 21:50 |
| opendevreview | Merged openstack/ironic master: Update hardware burn-in docs https://review.opendev.org/c/openstack/ironic/+/938606 | 21:50 |
| opendevreview | Merged openstack/ironic master: Add ignore_project_check_for_admin_tasks config option https://review.opendev.org/c/openstack/ironic/+/943028 | 21:50 |
| opendevreview | Merged openstack/ironic-inspector master: Release note: Inspector is dying/dead https://review.opendev.org/c/openstack/ironic-inspector/+/944297 | 21:50 |
| opendevreview | Verification of a change to openstack/ironic master failed: Fix default IRONIC_DEFAULT_TRAITS setting https://review.opendev.org/c/openstack/ironic/+/942647 | 21:51 |
| opendevreview | Verification of a change to openstack/ironic master failed: Enable atop on jobs https://review.opendev.org/c/openstack/ironic/+/942340 | 21:57 |
| opendevreview | Satoshi Shirosaka proposed openstack/ironic-python-agent master: WIP Add ContainerHardwareManager https://review.opendev.org/c/openstack/ironic-python-agent/+/941714 | 22:09 |
| opendevreview | Satoshi Shirosaka proposed openstack/ironic-python-agent-builder master: Create Podman DIB for container-based cleaning https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/943348 | 22:14 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
