| opendevreview | Steve Baker proposed openstack/sushy-tools master: Add virtual-media-boot to openstack driver https://review.opendev.org/c/openstack/sushy-tools/+/906768 | 02:17 |
|---|---|---|
| opendevreview | Steve Baker proposed openstack/sushy-tools master: Add virtual-media-boot to openstack driver https://review.opendev.org/c/openstack/sushy-tools/+/906768 | 03:56 |
| rpittau | good morning ironic! o/ | 07:32 |
| rpittau | JayF, TheJulia, sorry about the duplicate work, I mentioned my patch multiple times yesterday :) | 07:32 |
| rpittau | JayF, TheJulia, did you check the metalsmith legacy job in IPA ? | 07:33 |
| rpittau | I'm wondering if it's an issue with parted on the new tinycore image | 07:45 |
| rpittau | I will disable it for the time being while we investigate, going to modify the same unit test fix patch | 07:47 |
| dtantsur | ++ | 07:52 |
| rpittau | mmmmm my git review gets stuck | 08:15 |
| opendevreview | Riccardo Pittau proposed openstack/ironic-python-agent master: Fix unit tests after ironic-lib changes https://review.opendev.org/c/openstack/ironic-python-agent/+/910480 | 08:15 |
| rpittau | oh there you go :) | 08:15 |
| opendevreview | Riccardo Pittau proposed openstack/ironic master: [WIP] Implement generic redfish vmedia attach detach https://review.opendev.org/c/openstack/ironic/+/912516 | 09:29 |
| opendevreview | Riccardo Pittau proposed openstack/ironic master: [WIP] Implement generic redfish vmedia attach detach https://review.opendev.org/c/openstack/ironic/+/912516 | 09:33 |
| opendevreview | Riccardo Pittau proposed openstack/ironic master: [WIP] Implement generic redfish vmedia attach detach https://review.opendev.org/c/openstack/ironic/+/912516 | 09:38 |
| iurygregory | good morning Ironic | 11:16 |
| opendevreview | Riccardo Pittau proposed openstack/ironic master: [WIP] Implement generic redfish vmedia attach detach https://review.opendev.org/c/openstack/ironic/+/912516 | 13:29 |
| TheJulia | rpittau: legacy bios booting? | 14:41 |
| rpittau | TheJulia: yes, I've disabled the job here https://review.opendev.org/c/openstack/ironic-python-agent/+/910480 cause it's blocking the CI for all IPA | 14:42 |
| TheJulia | I really think we could drop the other image building path at this point and most testing around bios booting as well | 14:44 |
| dtantsur | morning TheJulia. With the new service steps, how far are we from things like updating BIOS settings / firmware upgrades on active nodes? | 14:54 |
| TheJulia | did we add the decorator? | 14:54 |
| TheJulia | I don't remember | 14:54 |
| dtantsur | TheJulia: adding a decorator is not a problem, I'm more interested in the general possibility (if you know) | 14:54 |
| TheJulia | I suspect if we did, it should just work but that specific usage has not been explicitly tested | 14:55 |
| dtantsur | Okay, I know someone who may have to test it urgently :) | 14:55 |
| TheJulia | yay and sorry | 14:55 |
| TheJulia | I do remember I went a bit overboard on some of the unit testing, so sort of hoping it all falls into "just works" | 14:56 |
| dtantsur | Do we recover boot devices and stuff like this? | 14:56 |
| * dtantsur is curious about lenovo with its UEFI quirks | 14:56 | |
| TheJulia | it is just one time boot requests if memory serves | 14:57 |
| TheJulia | lenovo's issues were explicitly we made nvram changes in the running OS and then tried to go "hey, do this" via the bmc which causes a reset | 14:57 |
| TheJulia | speaking of, CI really doesn't love the functional tests these days | 14:57 |
| dtantsur | yeah, so I wonder if we get this sort of a reset here | 14:57 |
| dtantsur | when we temporary switch to IPA | 14:58 |
| TheJulia | I would sincerely hope not, but if it gets recorded as something to persistently boot to by the firmware, then that is sort of a bug | 14:58 |
| TheJulia | possibly one we can teach IPA to navigate around | 14:58 |
| dtantsur | Okay, yeah. I only care about Dell/HPE initially, as usual. | 14:59 |
| dtantsur | I'm glad you haven't said "no way" at least :D | 15:01 |
| TheJulia | Nah, it is all just very fuzzy since I'm pulled in so many different directions | 15:12 |
| dtantsur | oh I hear ya | 15:18 |
| dtantsur | this is one more completely unexpected (not to say unpredictable) direction for me | 15:18 |
| TheJulia | Yeah, I seem to remember discussing "eventually, this will come up" | 15:22 |
| * dtantsur nods | 15:34 | |
| opendevreview | Merged openstack/ironic stable/2023.1: [iRMC] Fix IPMI incompatibility handling error https://review.opendev.org/c/openstack/ironic/+/885275 | 16:00 |
| rpittau | good night! o/ | 17:21 |
| opendevreview | Verification of a change to openstack/ironic master failed: Special case lenovo UEFI boot setup https://review.opendev.org/c/openstack/ironic/+/908946 | 17:23 |
| cid | "introduce an API version where any special treatment of .json is disabled in the API", but still maintain compatibility with the guess_content_type_from_ext feature? | 18:47 |
| cid | https://bugs.launchpad.net/ironic/+bug/1748224 | 18:47 |
| JayF | Thanks for picking that up, I'd absolutely make sure to test it still is an issue ebfore you change it | 18:53 |
| cid | sure thing | 18:56 |
| opendevreview | Julia Kreger proposed openstack/ironic master: docs: augment admin troubleshooting docs for system scope context https://review.opendev.org/c/openstack/ironic/+/908203 | 19:20 |
| opendevreview | Verification of a change to openstack/ironic master failed: Special case lenovo UEFI boot setup https://review.opendev.org/c/openstack/ironic/+/908946 | 19:25 |
| dking | Is there a reason that you don't use any faker in your unit tests? | 20:00 |
| JayF | For IPA, we usually use actual output from physical machines when possible. I'm not sure faker would provide us a lot of vvalue in that case | 20:06 |
| JayF | It's not in openstack global requirements, but we don't have to have testing packages coinstallable so that doesn't mean nothing openstack uses it. hmm | 20:06 |
| dking | JayF: I'm on the IPA unit tests for hardware.py. It seems like there's a bunch of tests, but they all just use the same few device names. It feels like the tests could be a lot more robust if you were using something like `faker.unix_device()` rather than just literally '/dev/sdj' for all of the tests. | 20:08 |
| dking | I suppose that it's from habit, but it feels a little dirty to me to just put in literal values where those actual values aren't what's expected from the code. In my personal tests, I even setup a fake BlockDevice so that I can just fill in the bits that I care about but otherwise quickly generate a fake BlockDevice. Here, even the names faked would make things more robust. | 20:11 |
| JayF | I'm not sure I share that perspective, but that kind of approach-change is best proposed and advocated for at the PTG, when there are lots of people there to build consensus | 20:16 |
| dking | Yeah, it's not something that I would personally change, but I was just wondering if there was a reason for it to be like it is. | 20:16 |
| JayF | For IPA, and unit tests, our primary goal is to mimic the environment it runs in, and that has usually taken the shape of "get real data from a machine and redact it" or "craft data that represents a real machine that exists" | 20:17 |
| TheJulia | We've had to add a number of "oh, $thing is slightly different over here" test cases over the years :( | 20:19 |
| dking | Maybe you'd catch a bunch of those "slightly different" if the code faked some values, catching a lot issues with different variations. | 20:21 |
| dking | Also, are the IPA unit tests still known to be broken? I've seen some tickets regarding that, but running tox from the latest master I pulled still seems to fail with an error. | 20:23 |
| JayF | Patch is in the gate now, one sec | 20:25 |
| JayF | https://review.opendev.org/c/openstack/ironic-python-agent/+/910480 | 20:27 |
| dking | Awesome. | 20:28 |
| opendevreview | Steve Baker proposed openstack/sushy-tools master: Add virtual-media-boot to openstack driver https://review.opendev.org/c/openstack/sushy-tools/+/906768 | 20:41 |
| opendevreview | Merged openstack/ironic-python-agent master: Fix unit tests after ironic-lib changes https://review.opendev.org/c/openstack/ironic-python-agent/+/910480 | 20:54 |
| jamesdenton__ | TIL the inspector processing hooks were required to make lldp stuffs actually... visible. I am dense. | 21:32 |
| JayF | you are not the first person to make that mistake | 21:36 |
| jamesdenton__ | they seemed more like... a suggestion | 21:37 |
| jamesdenton__ | "who WOULDN'T want this" | 21:37 |
| jamesdenton__ | :D | 21:37 |
| JayF | Feel free to reword the docs to emphasize that if you think it'll help | 21:37 |
| jamesdenton__ | hey now | 21:38 |
| jamesdenton__ | i'd rather complain | 21:38 |
| jamesdenton__ | "These are plugins that are enabled by default and should not be disabled, unless you understand what you’re doing" ... "Here are some plugins that can be additionally enabled". Like I said, i'm just dense. | 21:39 |
| JayF | So, I took this feedback and shared it with the contractor working w/GR-OSS right now on Ironic docs | 21:40 |
| JayF | so at least you're not screaming into a well :P | 21:40 |
| jamesdenton__ | appreciate that. If there's a way to word it and make it more obvious, i create a PR | 21:41 |
| JayF | if I were tackling it, I'd probably take the quoted portion and put it in a ..note or ..warn or similar | 21:42 |
| TheJulia | some people know it... | 21:42 |
| jamesdenton__ | yes, that might be helpful. or even something like "by default, lldp information may be gathered in raw format but not processed in human-readable format without these plugins enabled" | 21:42 |
| TheJulia | although I had a call with a customer on friday where I was like "and inspection is optional... oh wait, you use an ml2 plugin... inspection is required unless you populate the port connection data" | 21:43 |
| jamesdenton__ | i mean, part of the problem is that i think i know more than i do about how any of this works :D | 21:43 |
| jamesdenton__ | @TheJulia i am trying to avoid populating the port connection data, abslutely | 21:44 |
| jamesdenton__ | absolutely, rather | 21:44 |
| TheJulia | heh | 21:44 |
| TheJulia | :) | 21:44 |
| jamesdenton__ | But, even then i think i'll need to fudge it some, since these are all port channels and the configuration i end up applying with NGS needs to hit PortChannel11 vs Ethernet1/1 | 21:45 |
| JayF | Good thing those inspection collectors/processors are pluggable (I think?) | 21:45 |
| jamesdenton__ | it says they are | 21:46 |
| JayF | For portchannel stuff, the stackhpc folks know how to set that up with NGS if we don't have it well documented | 21:48 |
| TheJulia | ... can you detect the channel name in static config? | 21:48 |
| TheJulia | ++ | 21:48 |
| * TheJulia should go load up the zed ci as a distraction, since the distraction I want... ice skating, is not an option tonight | 21:49 | |
| JayF | I need to fix up a release patch and write the release notes prelude | 21:49 |
| JayF | but I think I left my focus somewhere outside when walking my dog lol | 21:49 |
| jamesdenton__ | I assumed the port channels needed to already be configured on the switch, then the local_link_connection for a given nic port would reference the portchannel name vs the switchport name. At least, that's how it used to be done, with port groups | 21:49 |
| JayF | I know we work with portchannels downstream, and I don't think that's enabled with patches but I'm not sure | 21:50 |
| JayF | and we absolutely flip things in/out of tenant/provisioning/cleaning vlans | 21:50 |
| JayF | that's just a handwavy part of the infra to me right now, honestly, I should go deeper | 21:50 |
| TheJulia | it is also an area where things start to become vendor specific | 21:51 |
| JayF | I think it's at that perfect confluence of: 1) if you want to do this, you know how; 2) everyone has different/strong opinions about the 'right way' to do it (including maybe other depts at your company) 3) it's vendor specific-ish | 21:51 |
| TheJulia | I know derek once had a customer complain the lldp data suddenly was different.... they upgraded the switch firmware | 21:51 |
| TheJulia | and 4) because of 1-3, people tend not to write docs/share their experiences as much | 21:52 |
| jamesdenton__ | agreed | 21:52 |
| JayF | it's like with the CMDB stuff, I think there's a second dimension | 21:52 |
| JayF | it's not /useful/ to share your experiences | 21:53 |
| TheJulia | even when you also hit a bug on the path | 21:53 |
| JayF | because your network design/philosophy and your cmdbs usually are more of a map to how your company thinks about computers than they are technical entities | 21:53 |
| TheJulia | you also might not recognize that | 21:53 |
| TheJulia | yup | 21:53 |
| jamesdenton__ | Well, i will try to share these bumps and bruises on the blog soon. | 21:54 |
| JayF | so everywhere I've worked (including jamesdenton__'s employer) just hit it with a hammer until it was flat and then was like NOBODY TOUCH ANYTHING :D | 21:54 |
| TheJulia | I hope a forge was involved in this process | 21:54 |
| clarkb | sometimes that lack of even communicating it is possible leads to peopel assuming it isn't possible which is its own set of problems | 21:55 |
| TheJulia | jamesdenton__: just keep in mind, nobody wants you to hit bumps and have bruises as a result | 21:55 |
| jamesdenton__ | of course, but they're inevitable, especially when one doesn't read the instructions and/or uses a deploy tool that doesn't really spell things out, either | 21:56 |
| jamesdenton__ | not the fault of anyone here, really, i promise. you all are top notch in my book :) | 21:57 |
| JayF | jamesdenton__: you still need to give me a place to send some swag | 21:57 |
| jamesdenton__ | ahh yeah i'll ping ya | 21:57 |
| JayF | jamesdenton__: and if I find it, I might pass on (at least one of the two) onmetal flags I possess | 21:57 |
| JayF | assuming you and your team are up to take on the mantle :P | 21:57 |
| jamesdenton__ | we are up for the challenge | 21:58 |
| TheJulia | Okay, public skating time at the ice rink is on Thursday, not Wendesday. | 21:59 |
| JayF | https://usercontent.irccloud-cdn.com/file/O0AxH0dT/1000004465.jpg | 21:59 |
| * TheJulia creates a sql update for her brain | 21:59 | |
| jamesdenton__ | JayF rock on | 22:00 |
| JayF | jamesdenton__: ^ I suspect the white/red/black vinyl one was left in the ck offices or was sent to mark | 22:00 |
| JayF | uh, silence | 22:00 |
| * TheJulia wanders off into the world of zuuuuul | 22:03 | |
| opendevreview | Julia Kreger proposed openstack/ironic stable/zed: stable only/ci: pin CI to dnsmasq 2.85/pin proliantutils/scciclient https://review.opendev.org/c/openstack/ironic/+/911158 | 22:22 |
| opendevreview | Merged openstack/ironic master: Special case lenovo UEFI boot setup https://review.opendev.org/c/openstack/ironic/+/908946 | 22:28 |
| opendevreview | Merged openstack/ironic-lib stable/2024.1: Update TOX_CONSTRAINTS_FILE for stable/2024.1 https://review.opendev.org/c/openstack/ironic-lib/+/911871 | 22:29 |
| JayF | oh wow | 22:44 |
| JayF | so 2023.1 -> 2024.1 slurp this time | 22:44 |
| JayF | would take you from | 22:44 |
| JayF | Ironic 21.4.x to Ironic 24.1.x | 22:44 |
| JayF | not a typo | 22:44 |
| TheJulia | heh, nice | 22:45 |
| opendevreview | Jay Faulkner proposed openstack/ironic master: Release notes prelude for 2024.1/24.1 https://review.opendev.org/c/openstack/ironic/+/912679 | 22:59 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/zed: stable only/ci: pin CI to dnsmasq 2.85/pin proliantutils/scciclient https://review.opendev.org/c/openstack/ironic/+/911158 | 23:29 |
| opendevreview | Jacob Anders proposed openstack/sushy-tools master: [WIP] Replace hardcoded BiosVersion with an updatable field https://review.opendev.org/c/openstack/sushy-tools/+/909487 | 23:52 |
| opendevreview | Jacob Anders proposed openstack/sushy-tools master: [WIP] Replace hardcoded BiosVersion with an updatable field https://review.opendev.org/c/openstack/sushy-tools/+/909487 | 23:54 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
