| opendevreview | nidhi proposed openstack/ironic master: Add PCIe function fields to redfish inspection https://review.opendev.org/c/openstack/ironic/+/963179 | 06:34 |
|---|---|---|
| opendevreview | Abhijith P C proposed openstack/ironic master: Fix grammar errors in bare metal state machine documentation https://review.opendev.org/c/openstack/ironic/+/965877 | 06:53 |
| rpittau | good morning ironic! happy friday! o/ | 08:18 |
| rpittau | more like broken CI friday :/ | 10:05 |
| dtantsur | alegacy: hey, so regarding switch configuration. I know that the inotify idea has been shot down so far. But we have a feature where mutable configuration can be updated on receiving SIGHUP. I wonder if we can do the same thing for switch configs. | 11:07 |
| dtantsur | Bonus: it will simplify things for Metal3 (no requirement for IrSO to manage switch CRD) | 11:08 |
| opendevreview | Merged openstack/ironic bugfix/31.0: Fail flat driver rebind when no VIFs are bound https://review.opendev.org/c/openstack/ironic/+/968331 | 11:16 |
| opendevreview | Merged openstack/ironic bugfix/30.0: Fail flat driver rebind when no VIFs are bound https://review.opendev.org/c/openstack/ironic/+/968332 | 11:17 |
| opendevreview | Verification of a change to openstack/ironic stable/2025.1 failed: Fail flat driver rebind when no VIFs are bound https://review.opendev.org/c/openstack/ironic/+/968330 | 11:19 |
| *** srelf__ is now known as continuity | 11:31 | |
| ykarel | Hi is the failure known one failing since two days https://zuul.openstack.org/builds?job_name=ironic-tempest-ovn-uefi-ipmi-pxe&branch=master&skip=0 | 11:45 |
| ykarel | bm logs waiting for something https://61409e8eaf5941f00fa2-27a42f979dc9642e645167ee17c588d5.ssl.cf2.rackcdn.com/openstack/df7dec9efb864043afcdf7537ac06b6b/controller/logs/ironic-bm-logs/node-0_no_ansi_2025-11-27-03%3A34%3A32_log.txt | 11:46 |
| rpittau | ykarel: there are multiple issues in ci at the moment, I'm investigating one related to uefi, might be the same | 12:04 |
| ykarel | rpittau, thx | 12:04 |
| alegacy | dtantsur: I'm willing to work that in as a separate patch, but i'd like to focus on getting the current ones merged first if you agree. | 12:27 |
| alegacy | dtantsur: i'll reach out about the metal3 thing... i don't quite see how that would avoid the need for IrSO to manage the switches, or why it would want to avoid that. | 12:37 |
| rpittau | there's clearly an issue with the ubuntu edk II, all the centos 10 jobs are ok | 13:22 |
| rpittau | from what I can see it may be due to a recent ovmf update | 13:22 |
| rpittau | looks like a regression in the most recent ovmf ubuntu pkg | 13:28 |
| opendevreview | Riccardo Pittau proposed openstack/bifrost master: Pin ubuntu ovmf package https://review.opendev.org/c/openstack/bifrost/+/968790 | 13:38 |
| opendevreview | Riccardo Pittau proposed openstack/bifrost master: [DNM] TEST CI https://review.opendev.org/c/openstack/bifrost/+/968791 | 13:40 |
| opendevreview | Riccardo Pittau proposed openstack/ironic master: Pin ubuntu ovmf package https://review.opendev.org/c/openstack/ironic/+/968795 | 13:48 |
| opendevreview | Takashi Kajinami proposed openstack/ironic master: drac: Deprecate unused options https://review.opendev.org/c/openstack/ironic/+/968807 | 14:51 |
| opendevreview | Riccardo Pittau proposed openstack/bifrost master: Pin ubuntu ovmf package https://review.opendev.org/c/openstack/bifrost/+/968790 | 15:24 |
| dtantsur | alegacy: if ironic-networking can reload switch config on SIGHUP, we can do something very similar to what we do for tracking TLS secrets in Metal3: https://github.com/metal3-io/ironic-image/blob/main/scripts/tls-common.sh#L101-L122 | 16:54 |
| dtantsur | alegacy: that brings up back to your initial idea: IrSO just provides ConfigMaps, but BMO writes to them | 16:54 |
| alegacy | so BMO would own/manage the switch crd? | 16:55 |
| dtantsur | alegacy: yeah, that's the thought | 16:56 |
| dtantsur | nicely consistent with other resources (BMH, HNA, etc) | 16:56 |
| alegacy | i'll have to think about that. i think that would be tricky for switch's configured to use ssh keys | 16:58 |
| opendevreview | Riccardo Pittau proposed openstack/bifrost master: Pin ubuntu ovmf package https://review.opendev.org/c/openstack/bifrost/+/968790 | 16:59 |
| dtantsur | alegacy: you mean, we'll need to inject the SSH keys into the container? | 16:59 |
| alegacy | ya, likely | 17:00 |
| dtantsur | alegacy: right. so you were thinking to link them as secrets from switch config? | 17:00 |
| alegacy | dtantsur: yes, I was considering (not prototyped yet) putting all of the keys in a secret and then mounting that to a directory and then referencing each file for each switch. | 17:02 |
| dtantsur | alegacy: you know workflow better, are there going to be different keys per switch? or is it more of a traditional workflow: an admin has a SSH key, and it works for all/most switches? | 17:03 |
| opendevreview | Riccardo Pittau proposed openstack/ironic master: Pin ubuntu ovmf package https://review.opendev.org/c/openstack/ironic/+/968795 | 17:04 |
| alegacy | dtantsur: I assume we won't have control over how they assign their user accounts for the switches so could be worse case of 1 key per switch. | 17:06 |
| alegacy | 1 key total would be easy, but I've never been that lucky. | 17:06 |
| dtantsur | alegacy: well, if you anyway plan to merge all keys into one secret, we can give this secret the same treatment as I suggested for the configmap, I guess? | 17:07 |
| dtantsur | Somewhere in the happy future, we'll probably create API for switches in Ironic. But we're very much not there yet. | 17:09 |
| dtantsur | Which is another reason I don't want to link switch configuration into the installation process. | 17:09 |
| dtantsur | (The 3rd reason is that some potential users don't use IrSO (yet?)) | 17:09 |
| alegacy | dtantsur: ok, i'll keep this in mind. | 17:12 |
| dtantsur | it looks like this hook point https://opendev.org/openstack/oslo.config/src/commit/ea0f0c3271f1e889659aad4e5ba1cf81b5986cd6/oslo_config/cfg.py#L3232 can be used to reload switches when oslo.config reloads itself | 17:25 |
| dtantsur | hmm, does SIGHUP still work though? | 17:36 |
| dtantsur | I mean, it works, but restarts the entire conductor, which is a bit nuclear | 17:41 |
| dtantsur | TheJulia: when you back from holidays ^^ | 17:43 |
| cardoe | rpittau: will you cut a sushy release? | 17:49 |
| dtantsur | I think it has been proposed already | 17:49 |
| dtantsur | cardoe: 5.9.0 is out | 17:49 |
| cardoe | awesome thanks | 18:00 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!