| opendevreview | Jacob Anders proposed openstack/ironic master: Batch non-BMC Redfish firmware updates under one reboot https://review.opendev.org/c/openstack/ironic/+/999720 | 00:18 |
|---|---|---|
| opendevreview | Jacob Anders proposed openstack/ironic master: Document batched Redfish firmware updates https://review.opendev.org/c/openstack/ironic/+/999721 | 00:18 |
| opendevreview | Jacob Anders proposed openstack/ironic master: Keep Redfish firmware update state in one record with explicit phases https://review.opendev.org/c/openstack/ironic/+/1010459 | 00:18 |
| opendevreview | Jacob Anders proposed openstack/ironic master: Add Dell LC job status helper for multiple JIDs https://review.opendev.org/c/openstack/ironic/+/1005656 | 05:27 |
| opendevreview | Jacob Anders proposed openstack/ironic master: Add Redfish BootProgress gate helper https://review.opendev.org/c/openstack/ironic/+/1005657 | 05:27 |
| opendevreview | Jacob Anders proposed openstack/ironic master: Verify firmware application after reboot before resuming https://review.opendev.org/c/openstack/ironic/+/1005658 | 05:27 |
| opendevreview | Jacob Anders proposed openstack/ironic master: Validate BMC stability only after BMC firmware updates https://review.opendev.org/c/openstack/ironic/+/1005659 | 05:27 |
| opendevreview | Jacob Anders proposed openstack/ironic master: Batch non-BMC Redfish firmware updates under one reboot https://review.opendev.org/c/openstack/ironic/+/999720 | 05:27 |
| opendevreview | Jacob Anders proposed openstack/ironic master: Document batched Redfish firmware updates https://review.opendev.org/c/openstack/ironic/+/999721 | 05:27 |
| opendevreview | Jacob Anders proposed openstack/ironic master: Keep Redfish firmware update state in one record with explicit phases https://review.opendev.org/c/openstack/ironic/+/1010459 | 05:27 |
| opendevreview | Jacob Anders proposed openstack/ironic master: Keep Redfish firmware update state in one record with explicit phases https://review.opendev.org/c/openstack/ironic/+/1010459 | 05:47 |
| opendevreview | Jacob Anders proposed openstack/ironic master: Batch non-BMC Redfish firmware updates under one reboot https://review.opendev.org/c/openstack/ironic/+/999720 | 06:48 |
| opendevreview | Jacob Anders proposed openstack/ironic master: Document batched Redfish firmware updates https://review.opendev.org/c/openstack/ironic/+/999721 | 06:48 |
| opendevreview | Jacob Anders proposed openstack/ironic master: Keep Redfish firmware update state in one record with explicit phases https://review.opendev.org/c/openstack/ironic/+/1010459 | 06:48 |
| opendevreview | Merged openstack/ironic master: Replace boot property literals with constants https://review.opendev.org/c/openstack/ironic/+/980829 | 08:41 |
| opendevreview | Jakub Jelinek proposed openstack/ironic-specs master: Warm available nodes for multi-tenant networking https://review.opendev.org/c/openstack/ironic-specs/+/1008056 | 09:44 |
| kubajj | Updated the spec for the ~fast-track with cleaning and provisioning network being separate~ warm-available ^ | 09:45 |
| opendevreview | Jed Preist proposed openstack/bifrost master: Add support for authentication when downloading deployment images https://review.opendev.org/c/openstack/bifrost/+/1008093 | 09:51 |
| iurygregory | good morning ironic | 12:08 |
| TheJulia | good morning | 12:46 |
| TheJulia | Let the caffination begin | 12:46 |
| JayF | Feel free to start without me. I don't have much input for firmware UI stuff. Might take a few extra minutes to get breakfast before the things start. | 12:58 |
| iurygregory | JayF, ack | 12:58 |
| iurygregory | enjoy your breakfast! | 12:58 |
| iurygregory | cardoe, first topic is the one about firmware update =) | 13:03 |
| cardoe | I’m rushing home! | 13:04 |
| cardoe | Traffic! 5 minutes | 13:04 |
| shermanm | cardoe: :waves: | 13:47 |
| cardoe | shermanm: we should connect sometime and see where we might overlap so we can work together on landing stuff | 13:55 |
| dtantsur | cardoe, shermanm, https://github.com/metal3-io/metal3-docs/pull/692 | 13:57 |
| shermanm | +1 | 13:57 |
| opendevreview | Doug Goldstein proposed openstack/ironic master: Add inspection-rules check to ironic-manage https://review.opendev.org/c/openstack/ironic/+/996194 | 14:38 |
| dtantsur | JayF: so, runbooks in gophercloud. Whose problem do I make this? :) | 15:01 |
| JayF | dtantsur: github issue in gophercloud, assigned to adamcarthur5 / sharpz7 | 15:01 |
| dtantsur | okie | 15:01 |
| JayF | adamcarthur5: please help dtantsur | 15:01 |
| dtantsur | JayF, adamcarthur5, https://github.com/gophercloud/gophercloud/issues/3642 (I cannot assign people there) | 15:02 |
| opendevreview | Doug Goldstein proposed openstack/ironic master: Add inspection-rules check to ironic-manage https://review.opendev.org/c/openstack/ironic/+/996194 | 15:07 |
| adamcarthur5 | dtantsur JayF can do ;)) | 15:31 |
| dtantsur | thx! | 15:31 |
| shermanm | cardoe: definitely | 15:43 |
| JayF | I'll be out during the first part of the next Ironic session to attend the Nova section, Adam is presenting a spec on placement API changes for openstack-exporter | 15:52 |
| cardoe | JayF: oooo get them to make openstack-exporter work better! | 15:56 |
| JayF | https://review.opendev.org/c/openstack/placement/+/990761 that's the idea, at least in context of placement | 15:56 |
| JayF | Adam is proposing a bulk-read API which would help with data gathering | 15:56 |
| shermanm | I don't know if it's come up yet, but has there been discussion about exposing some kind of operator api/cli for networking-generic-switch? | 17:09 |
| shermanm | e.g. "I get an error about stale vlan on switch X, port Y". can my operator run some kind of "ngs port remove --switch x --port y" instead of having to go ssh to the switch | 17:09 |
| dtantsur | I don't think we've ever talked about it, but it sounds useful to me | 17:09 |
| shermanm | as a small subset of "really, I'd like NGS to reconcile against target state" instead of one shot write-only | 17:09 |
| shermanm | we've been calling out a lot of good areas where the happy path works, but during errors or debugging we have to break the abstraction ironic is giving over the BMC, switches, etc | 17:10 |
| shermanm | and staying within ironic gives some benefits like "add rbac" and "don't modify thing Y during state X" | 17:11 |
| JayF | this is kinda what reattach_networking was for? | 17:11 |
| JayF | NGS didn't even have an Ironic-accessible API until Alain's work on ironic-networking :) | 17:12 |
| shermanm | yeah, I just wasn't sure if we'd successfully pulled that out of the adoption case, was a little fuzzy there | 17:12 |
| JayF | that already exists and is implemented for neutron | 17:12 |
| JayF | there was a patch up to do it for ironic-networking, it got -1'd for technical/implementation reasons and I think it's gonna be rewritten | 17:12 |
| * TheJulia glares at the daystar making the screen difficult to read | 17:16 | |
| shermanm | TheJulia: did we still plan on doing that network chat sometime today? | 17:19 |
| cardoe | shermanm: that's why we don't use NGS directly.. We reconcile against target state. | 17:19 |
| TheJulia | shermanm: cardoe: can we chat around 1pm pacific? I'm in a car right now and will pass through limited cell service for ~20m here in .... 20 minutes or so | 17:20 |
| TheJulia | (I was thinking... friday for some reason) | 17:20 |
| TheJulia | (also, got forced into going and dealing with a thing in person today as well | 17:20 |
| shermanm | that time works for me | 17:20 |
| TheJulia | hopefully the repair tech doesn't show up while we're talking :\ | 17:21 |
| TheJulia | And whee heavy winds today :( | 17:22 |
| shermanm | cardoe: mostly I didn't see anything that was open source or non-vendor-specific that worked that way. I'd been digging into the (now dead) networking-odl for ideas. General structure of: precommit validates request and writes into a journal, postcommit watches for the current state to reach journal state, and a periodic job to make sure the journal state still matches what neutron wants. (e.g. cold startups, event | 17:24 |
| shermanm | dropped somehow,....) | 17:24 |
| shermanm | because if we have a more global view of current vs target, we can make better decisions about how to batch commands, warn if something outside NGS changed switch state, etc | 17:24 |
| shermanm | ... somehow I thought openflow was supposed to solve all of this 10 years ago | 17:25 |
| * TheJulia watches a load shift in the wind on a truck and wonders if she is going to have a very unexpected day | 17:30 | |
| cardoe | 1pm pacific is fine | 17:30 |
| TheJulia | cool cool | 17:30 |
| TheJulia | cardoe: its slightly annoying that to have an idea factory, in gerrit, we would have to do something like inaugust :) | 17:43 |
| TheJulia | This is not my day… https://usercontent.irccloud-cdn.com/file/P149NrxH/1792086324.JPG | 17:45 |
| JayF | it is if you go into old fashioned train robbery mode and make off with a bear statue | 17:56 |
| JayF | add a guitar and we got ourselves an official ironic statue | 17:56 |
| TheJulia | I can go ask the metalworker… he is not far down the road from my hiding place | 18:05 |
| TheJulia | Their home might also be right down the road from our home as well…. It’s a small world. | 18:07 |
| opendevreview | Julia Kreger proposed openstack/ironic master: WIP: Add automatic trait management for accelerators https://review.opendev.org/c/openstack/ironic/+/1010732 | 18:41 |
| TheJulia | JayF: ^^^ | 18:41 |
| cardoe | TheJulia: you're driving behind horses... no code pushing | 18:52 |
| TheJulia | ... we *were* | 18:52 |
| TheJulia | and I was the pasenger | 18:53 |
| TheJulia | cardoe: shermanm: https://meetpad.opendev.org/oct2026-ptg-ironic | 19:58 |
| TheJulia | There will be solutioning until morale improves! | 20:01 |
| shermanm | :) | 20:05 |
| opendevreview | Doug Goldstein proposed openstack/ironic master: Add inspection-rules check to ironic-manage https://review.opendev.org/c/openstack/ironic/+/996194 | 20:32 |
| *** hroy_ is now known as hroy | 20:59 | |
| opendevreview | Doug Goldstein proposed openstack/ironic master: Fix TBN universal port and network constructors https://review.opendev.org/c/openstack/ironic/+/1010775 | 21:24 |
| TheJulia | So, regarding the rusting steel ponies! Turns out they were going to the place near my home. | 21:59 |
| TheJulia | It's a small world after all! | 21:59 |
| TheJulia | Now, sing! | 21:59 |
| * TheJulia ducks | 21:59 | |
| TheJulia | shermanm: looks like we can likely just "trick" the kernel into being a bit of a proxy for us. FWIW. Going to keep digging | 22:07 |
| shermanm | nice! | 22:08 |
| TheJulia | or at least kernel client, some portion software to offer the target back off. | 22:17 |
| TheJulia | Wont' be perfect, but even the linux software server benchmarks I've seen were still moderately impressive | 22:17 |
| shermanm | TheJulia: btw, have you seen idrac "serial data capture"? https://github.com/dell/iDRAC-Redfish-Scripting/blob/master/Redfish%20Python/ExportClearSerialDataLogsREDFISH.py | 22:23 |
| shermanm | redfish API to enable continuous capture of serial console to 512kb circular buffer | 22:23 |
| shermanm | I'm curious if they've solved the blocking / perf issues | 22:23 |
| TheJulia | i have not, but apparently there is a "standard websocket" approach, albeit... likely needs standards folks to clarify | 22:25 |
| shermanm | heh, I might be able to start benchmarking serial console performance | 22:27 |
| TheJulia | be careful of your interrupts | 22:33 |
| TheJulia | Anyway, I'm going to hammer at... maybe a series of specs to try and articulate a model and approach over the next few days. The overall idea is huge | 22:34 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
