| iurygregory | TheJulia, yup | 00:11 |
|---|---|---|
| cid | about the ironic-arm-ci | 00:21 |
| cid | I saw a little light today; the whole configuration is all to populate the VM template. | 00:21 |
| cid | So, CI | devstack loads configuration from local.conf or zuul job, | 00:22 |
| cid | then ironic devstack plugin prepares them (including some magic), | 00:22 |
| cid | and passes them over to create-node.sh which receives them, prepares an arguments | 00:22 |
| cid | list, and passes to the configure-vm.py which in turn populates vm.xml. | 00:22 |
| cid | quite a trip, | 00:25 |
| cid | I'm so happy. | 00:25 |
| cid | I'll just get a little sleep now. | 00:25 |
| cid | good night \o | 00:25 |
| iurygregory | good night cid o/ | 00:33 |
| opendevreview | Merged openstack/ironic-python-agent master: Remove eventlet workaround https://review.opendev.org/c/openstack/ironic-python-agent/+/918087 | 00:48 |
| opendevreview | Takashi Kajinami proposed openstack/ironic stable/2024.1: Remove SQLAlchemy tips jobs https://review.opendev.org/c/openstack/ironic/+/919641 | 03:42 |
| opendevreview | Takashi Kajinami proposed openstack/ironic stable/2024.1: Remove SQLAlchemy tips jobs https://review.opendev.org/c/openstack/ironic/+/919641 | 03:43 |
| opendevreview | Takashi Kajinami proposed openstack/ironic stable/2023.2: Remove SQLAlchemy tips jobs https://review.opendev.org/c/openstack/ironic/+/919640 | 03:46 |
| opendevreview | Verification of a change to openstack/ironic-python-agent stable/2023.1 failed: Unmount config drives https://review.opendev.org/c/openstack/ironic-python-agent/+/917852 | 04:06 |
| opendevreview | Merged openstack/ironic master: Add servicing states to states doc, fix state diagram https://review.opendev.org/c/openstack/ironic/+/919353 | 04:22 |
| opendevreview | Merged openstack/ironic stable/2023.2: Remove SQLAlchemy tips jobs https://review.opendev.org/c/openstack/ironic/+/919640 | 06:19 |
| opendevreview | Merged openstack/ironic stable/2024.1: Remove SQLAlchemy tips jobs https://review.opendev.org/c/openstack/ironic/+/919641 | 06:50 |
| dtantsur | JayF: re running out-of-bands steps before automated cleaning: we needed that, and that's a large reason we added verify steps. | 07:57 |
| dtantsur | Our case was to prune stale iDRAC jobs, which, as you say, must be done before you try to boot a ramdisk. | 07:58 |
| dtantsur | The problem with doing it in the automated cleaning is that inspection runs before that. | 07:58 |
| cid | good morning \o | 09:25 |
| iurygregory | good morning Ironic | 11:04 |
| cid | o/ | 11:19 |
| TheJulia | good morning | 13:35 |
| TheJulia | dtantsur: I *think* based upon a side comment JayF made to me yesterday, it is something being very strongly desired as part of every cleaning, including when reclaiming a node from a user | 13:36 |
| TheJulia | dtantsur: might be best for high bandwidth discussion, loosely scheduled in at 1500 UTC :) | 13:37 |
| * TheJulia tries to wake up | 13:38 | |
| dtantsur | I have some errands here, not sure if I make it | 13:45 |
| dtantsur | We definitely want to avoid booting IPA while servicing though, I guess this is relevant | 13:46 |
| opendevreview | Julia Kreger proposed openstack/ironic master: docs: allow the state diagram to be larger https://review.opendev.org/c/openstack/ironic/+/919585 | 13:48 |
| JayF | I would rather not hijack our 1500utc review jam on a tangent of a feature that likely wouldn't be implemented for a cycle plus anyway | 14:11 |
| JayF | Although I appreciate the enthusiasm 😁 | 14:11 |
| TheJulia | well, we can always start slightly early and hijack it until 1500 | 14:11 |
| TheJulia | or just reach the same page in advance of 1500, or after 1600 | 14:11 |
| TheJulia | you never know! | 14:11 |
| TheJulia | the possibilities are unbound! | 14:12 |
| JayF | Honestly, the person from downstream who had an interest in this, I asked them to write up the full use case. I mainly don't want the discussion to happen until that writeup up exist in a Launchpad rfe/bug | 14:14 |
| JayF | I am a low quality proxy 😂 | 14:14 |
| TheJulia | That is totally fair | 14:14 |
| JayF | Lossy compression like an MP3 at 32bps | 14:14 |
| JayF | Sometimes when I ask questions like that early, I'm fishing for if a feature exists that I lost the decoder ring for | 14:15 |
| JayF | And to be fair, in this case, I'm going to get to add verify steps to my talk because I did forget about them lol | 14:15 |
| dtantsur | I can imagine stuff like.. hmm.. "it's dangerous to boot privileged code or even boot on the provisioning network at all until we make sure the secure boot/attestation/whatever is in order" | 14:33 |
| TheJulia | This is why the exact problem the user is trying to solve is needed | 14:35 |
| JayF | I've been thinking about how to do it in our current model | 14:35 |
| JayF | negative priority values indicating "before a ramdisk is booted" could be interesting | 14:36 |
| JayF | https://us06web.zoom.us/j/82701965732?pwd=U21jZJY7H7bwBnPuAA9M0XAT0qfP3Q.1 I'm going to be in here working on my talk until others are ready, I know it's super early but it's easier to do this than risk forgetting | 14:37 |
| TheJulia | I guess it is time to go make coffeeeeeeee | 14:39 |
| JayF | https://github.com/HewlettPackard/sdflex-ironic-driver *blink* | 14:44 |
| iurygregory | *eyes* | 14:45 |
| iurygregory | well, quite old and no updates in the past 2yrs | 14:46 |
| JayF | looks mostly copied out tbh | 14:46 |
| JayF | with minor tweaks | 14:46 |
| TheJulia | I've heard rumor of a superdome flex driver | 14:46 |
| JayF | rume no more lol | 14:50 |
| JayF | I'm going to fork it just so i have a copy | 14:50 |
| iurygregory | https://us06web.zoom.us/j/82701965732?pwd=U21jZJY7H7bwBnPuAA9M0XAT0qfP3Q.1 is the link right? | 15:00 |
| iurygregory | grabbing some water and I will join | 15:01 |
| JayF | yeah iury that's it | 15:02 |
| *** dmellado4 is now known as dmellado | 15:07 | |
| JayF | https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/913926/ | 15:13 |
| JayF | https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/915170 | 15:17 |
| JayF | https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/917277 | 15:19 |
| JayF | https://review.opendev.org/c/openstack/ironic-specs/+/916126/5/specs/approved/mercury.rst#13 | 15:28 |
| opendevreview | Takashi Kajinami proposed openstack/ironic master: Use specific element to detect database backend https://review.opendev.org/c/openstack/ironic/+/919749 | 15:29 |
| opendevreview | Takashi Kajinami proposed openstack/ironic master: Use specific element to detect database backend https://review.opendev.org/c/openstack/ironic/+/919749 | 15:51 |
| opendevreview | Takashi Kajinami proposed openstack/ironic master: Use specific element to detect database backend https://review.opendev.org/c/openstack/ironic/+/919749 | 15:56 |
| dtantsur | So, a question. Right now, we ~ always boot IPA during cleaning/servicing. Mostly because we expect that some steps may be present both in-band and out-of-band. What if we assume that if all steps are present on the hardware type, we don't need to boot IPA? | 15:57 |
| JayF | note that generally we need to look at https://review.opendev.org/c/openstack/ironic-python-agent/+/912949 and other unmaintained/* changes to land | 16:05 |
| JayF | dtantsur: well, in my use case, I need IPA, too -- I just don't wanna boot it until after I've done some prereq work | 16:05 |
| JayF | I want to be able to just indicate "run these steps BEFORE ever booting an agent" | 16:06 |
| JayF | which is why I was pondering the idea of negative priorities | 16:06 |
| opendevreview | Merged openstack/ironic stable/2024.1: Inject a randomized publisher id https://review.opendev.org/c/openstack/ironic/+/917735 | 16:29 |
| opendevreview | Merged openstack/ironic stable/2023.2: Inject a randomized publisher id https://review.opendev.org/c/openstack/ironic/+/917861 | 16:33 |
| opendevreview | Merged openstack/ironic stable/2023.1: Inject a randomized publisher id https://review.opendev.org/c/openstack/ironic/+/917862 | 16:33 |
| opendevreview | Verification of a change to openstack/ironic-tempest-plugin master failed: Patch to enforce json extension works in existing API behaviour https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/913926 | 16:57 |
| opendevreview | Merged openstack/ironic-python-agent-builder master: Match publisher id https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/915170 | 17:14 |
| JayF | Looks like the OSCaaS changes are saving around 4 minutes a run in ironic | 18:17 |
| opendevreview | Merged openstack/ironic-tempest-plugin master: Vmedia advanced ops: Clarify config drive and don't ping after rebuild https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/917277 | 18:17 |
| JayF | when the devstack change lands that'll be great | 18:17 |
| JayF | because it not only helps our jobs directly, I suspect this may also help with noisy neighbor issues (maybe I should say 'hope' rather than suspect) | 18:17 |
| opendevreview | Julia Kreger proposed openstack/ironic-tempest-plugin master: Exclude ramdisk tests with tinycore in uefi mode https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/919762 | 18:40 |
| TheJulia | JayF: https://bugs.launchpad.net/ironic/+bug/2065811 and ^ :) | 18:42 |
| TheJulia | Another tempest fix, btw https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/919544 | 18:51 |
| *** dmellado6 is now known as dmellado | 20:15 | |
| opendevreview | cid proposed openstack/ironic master: wip: Provision aarch64 fake-bare-metal-vms https://review.opendev.org/c/openstack/ironic/+/915441 | 21:26 |
| opendevreview | cid proposed openstack/ironic master: wip: Provision aarch64 fake-bare-metal-vms https://review.opendev.org/c/openstack/ironic/+/915441 | 21:31 |
| JayF | I just realised that, at least AFAICT, there's no easy way to embed a HWM into ipa-builder images without editing our element | 22:07 |
| JayF | wouldn't be a terrible idea to have a "custom hardware manager" element in that group which would install a given repo into the ipa venv | 22:07 |
| ashinclouds[m] | If you install it as a package, much like proliantutils, then that should be sufficient. All boils down to how you model the thing to be used/consumed | 22:09 |
| JayF | what I'm saying is, we have no way to say "install a random extra package in the IPA venv" without editing the element directly | 22:09 |
| JayF | the mechanism is so trivial I'm just surprised we never plumbed it into our DIB stuff | 22:10 |
| * JayF adds it to the list of things to do on a day when he wants a win | 22:10 | |
| TheJulia | oh yeah I see what you mean now | 22:15 |
| TheJulia | yeah | 22:15 |
| JayF | I'm unearthing a lot of ... sharp edges | 22:18 |
| JayF | trying to read through our docs to not miss anything for this prezi | 22:18 |
| opendevreview | Jay Faulkner proposed openstack/ironic master: Deduplicate and remove invalid information for steps https://review.opendev.org/c/openstack/ironic/+/919779 | 22:20 |
| JayF | e.g. | 22:20 |
| JayF | not to mention https://docs.openstack.org/ironic/latest/admin/cleaning.html#what-cleaning-step-is-running | 22:24 |
| JayF | I have to close that page, I can feel myself stressing out lol | 22:25 |
| TheJulia | FWIW, a bug should likely be opened on that | 23:09 |
| TheJulia | get us back above 140 :) | 23:09 |
| TheJulia | ... I started cleaning up bugs and we were above 156... | 23:09 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!