| opendevreview | Merged openstack/python-ironicclient stable/2026.2: Update TOX_CONSTRAINTS_FILE for stable/2026.2 https://review.opendev.org/c/openstack/python-ironicclient/+/1004631 | 00:00 |
|---|---|---|
| cardoe | We do need to finish the schema stuff. | 01:56 |
| opendevreview | Merged openstack/ironic-python-agent master: Maintain plugin requirements as optional dependencies https://review.opendev.org/c/openstack/ironic-python-agent/+/1000668 | 05:04 |
| opendevreview | Jed Preist proposed openstack/ironic-python-agent master: Retry clock sync if it fails https://review.opendev.org/c/openstack/ironic-python-agent/+/1010031 | 08:54 |
| iurygregory_ | good morning ironic | 12:07 |
| TheJulia | good morning | 12:38 |
| * JayF is racing through a Starbucks drive-through, and may be a minute or two late | 12:57 | |
| * TheJulia hears the electric "vroom vvrooom" from a distance | 12:57 | |
| JayF | https://usercontent.irccloud-cdn.com/file/9h8ezxY9/irccloudcapture5936521028140128229.jpg | 12:58 |
| JayF | No vroom. Only wait. 😂 | 12:58 |
| TheJulia | I'm sure Zathras has a quote for this. | 12:59 |
| JayF | cardoe: can you hop in for team photo | 13:46 |
| JayF | This is what I've got for team photo, wdyt https://usercontent.irccloud-cdn.com/file/tRE110ka/team-photo.png | 15:42 |
| shermanm | side-bar, but not something that PTG needs tbh; is there any interest in functional tests for NGS and for the nova ironic driver? I'm happy to contribute examples, but not sure if there are existing preferences on this topic. | 17:14 |
| JayF | Julia has a *lot* of historical work in this area | 17:18 |
| JayF | devstack can spin up test VMs running demo versions of switch OSes in many cases | 17:18 |
| JayF | but they run so slowly any thoughts of running automated tested on them is rough | 17:19 |
| shermanm | I was doing some llm assisted debugging, and had some decent success following how nova's other functional tests worked, just mocking things like "a thing that sends text back and forth" instead of full vm-based tests per driver | 17:52 |
| shermanm | more for flushing out concurrency bugs, and regression tests for known issues | 17:52 |
| shermanm | I can push up some examples for show and tell | 17:52 |
| TheJulia | I sort of built some setup scripting in to help enable it with selected cases for various drivers, but its also... super hit or miss | 17:59 |
| JayF | 'integration tests' that just validate that we're sending the right strings across the ssh sesh (in NGS context) are just unit tests that are more likely to randomly fail | 18:00 |
| JayF | the value the integration gives us there is that it validates stuff actually works, right? | 18:00 |
| JayF | is there a failure case those tests would expose that our current ones don't? | 18:00 |
| shermanm | what jumps to mind first is etcd and coordination, since its actually pretty cheap and fast to run in CI | 18:03 |
| JayF | we, at least at one point, had a job running that did that | 18:24 |
| TheJulia | I guess it feels way more important now that more modern switches have actual session and configuration management as opposed to "oh, other user logged in and locked my config database flat file" | 18:30 |
| JayF | https://docs.openstack.org/ironic-ui/latest/ I've been asked if we could add a screenshot here | 18:47 |
| * JayF will poke it with devstack if someone doesn't get there first | 18:47 | |
| -opendevstatus- NOTICE: The Etherpad service at etherpad.opendev.org (also used by meetpad.opendev.org) is being restarted now to apply a security update; any open pads will display an error message until refreshed, and author colors may get reset. | 19:52 | |
| opendevreview | Julia Kreger proposed openstack/ironic-specs master: ROCE Enabled BMaaS https://review.opendev.org/c/openstack/ironic-specs/+/990413 | 19:53 |
| opendevreview | Julia Kreger proposed openstack/ironic-specs master: ROCE Enabled BMaaS https://review.opendev.org/c/openstack/ironic-specs/+/990413 | 20:20 |
| TheJulia | just over 500 lines now, slimmer, less excess detail | 20:33 |
| opendevreview | Doug Goldstein proposed openstack/ironic master: Add iDRAC BMC config steps (NTP, DNS, OIDC) https://review.opendev.org/c/openstack/ironic/+/1007083 | 21:07 |
| TheJulia | shermanm: When can we plot/scheme storage and networking and chaos oh my? | 21:23 |
| shermanm | I should have time slots after ptg on thurs/fri, otherwise next week is better | 21:24 |
| TheJulia | cardoe: when would work for you? ^^^ | 21:29 |
| TheJulia | Thursday after the ptg session wraps? | 21:45 |
| cardoe | Thursday after works for me | 21:45 |
| shermanm | works for me too | 21:45 |
| TheJulia | cool cool, sounds like a plan | 21:48 |
| opendevreview | Julia Kreger proposed openstack/ironic-specs master: Extend ironic auth models https://review.opendev.org/c/openstack/ironic-specs/+/1010295 | 22:03 |
| TheJulia | dtantsur: ^^^ | 22:04 |
| cardoe | So the storage HTTP/TFTP... I'd run a ceph and let Ironic push to ceph and let ceph serve it up if that's cool for one of the backends? | 22:16 |
| JayF | The big limitation there, cardoe, is around if the ceph http server acts like a "good" one when hit with things like range requests | 22:17 |
| JayF | and if the long URLs it creates are short enough to fit into a BMC | 22:17 |
| cardoe | Yeah. I'm just trying to think of what is an OpenStack-y accepted storage medium. | 22:18 |
| JayF | hint: I think we still have a bug report from someone whose system will only hit http://hostname/install.iso (details may be slightly wrong; but basically it'd only mount vmedia images located at `/specific_filename` | 22:18 |
| JayF | If the interface is not so small that it's trivial to add support for one, upstream or otherwise, I think it'll be a sign we might be doing the wrong thing :) | 22:18 |
| cardoe | Do we have anywhere that we have an HTTPS certificate chain as an argument... I assume the best way to do that is a file name right? | 22:21 |
| TheJulia | uhh, that was HPE Gen 9 gear afaik | 22:21 |
| TheJulia | (the file needs to end in .iso | 22:22 |
| TheJulia | ) | 22:22 |
| TheJulia | cardoe: only cacert afaik | 22:22 |
| JayF | cardoe: make sure any file:/// urls get filtered thru image service | 22:22 |
| JayF | cardoe: if you are implementing a thing that slurps files off a disk, it's basically a security vuln waiting to happen if you don't :| | 22:23 |
| TheJulia | JayF++ | 22:23 |
| cardoe | JayF: https://review.opendev.org/c/openstack/ironic/+/1007083 | 22:23 |
| opendevreview | Michael Sherman proposed openstack/networking-generic-switch master: Reproduce batch queue reads covering other switches https://review.opendev.org/c/openstack/networking-generic-switch/+/1010297 | 22:23 |
| JayF | none of those are slurping files off disks? | 22:24 |
| JayF | (right?) | 22:24 |
| cardoe | Nothing is slurping a file off the disk BUT | 22:24 |
| shermanm | ^ functional test and I think a real bug | 22:24 |
| cardoe | the iDRAC can use OIDC for letting folks login which is a totally normal protocol that things like Active Directory controllers support | 22:25 |
| JayF | I've known places to use that support | 22:25 |
| cardoe | It can use a hardcoded client ID and client secret or it can use OIDC dynamic registration | 22:25 |
| cardoe | The dynamic registration route requires the client to transmit a secret to the OIDC server. Dell iDRAC happens to be an Apache server that runs mod_auth_oidc and to prevent that secret from leaking to a spoofed endpoint, they require that you tell it your OIDC issuer and provide the HTTPS certificate chain of that OIDC issuer before it'll transmit the secret. | 22:26 |
| * TheJulia is feeling a sense of deja vu | 22:26 | |
| JayF | TheJulia: are you thinking what I'm thinknig | 22:27 |
| * JayF DM's cardoe | 22:27 | |
| cardoe | She might have heard me say this earlier. | 22:27 |
| TheJulia | heh | 22:27 |
| TheJulia | Just don't make me call some phone number and say "conundrum" | 22:27 |
| shermanm | "shibboleth" | 22:28 |
| * TheJulia notes 7 days ago a time shibboleth appeared. | 22:28 | |
| cardoe | shermanm: to the corner with you | 22:29 |
| shermanm | it's ok, I can hang out with whichever orange cat buddy owns the corner today | 22:29 |
| JayF | I have a countdown web game I coded, it has a conundrum at the end | 22:29 |
| TheJulia | shermanm: I think its presently one tabby and one orange cuddling in the corner's sun puddle. If that is the corner with the most excellent window. | 22:31 |
| TheJulia | https://usercontent.irccloud-cdn.com/file/cSNuOuPU/1791930716.JPG | 22:32 |
| shermanm | <3 | 22:32 |
| JayF | https://usercontent.irccloud-cdn.com/file/wT9jOqQO/juno-on-rebounder.png | 22:33 |
| TheJulia | cuteness! | 22:33 |
| TheJulia | the tabby has her paws over her eyes. She is super cute when she is not making the floor run red with the blood of humans | 22:33 |
| TheJulia | (she is very... bitey.) | 22:34 |
| shermanm | just .... excitable, y'know | 22:35 |
| cardoe | shermanm: so https://review.opendev.org/c/openstack/ironic/+/1007083 with that you can call a few servicing steps and enroll the Dell iDRAC into an OIDC issuer and then people can access the iDRAC without having any credentials | 22:35 |
| JayF | I'm calling my EOD since I started at 6a | 22:36 |
| JayF | see you cool cats tomorrow | 22:36 |
| opendevreview | Michael Sherman proposed openstack/networking-generic-switch master: Reproduce batch queue reads covering other switches https://review.opendev.org/c/openstack/networking-generic-switch/+/1010297 | 22:39 |
| TheJulia | … yeah, me too. ENOBRAIN | 22:42 |
| opendevreview | Michael Sherman proposed openstack/networking-generic-switch master: Reproduce batch queue reads covering other switches https://review.opendev.org/c/openstack/networking-generic-switch/+/1010297 | 22:45 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
