Wednesday, 2025-04-30

opendevreviewIury Gregory Melo Ferreira proposed openstack/ironic-specs master: NIC Firmware Updates  https://review.opendev.org/c/openstack/ironic-specs/+/94850301:42
arne_wiebalckGood morning, Ironic!06:25
AmarachiOrdor[m]Good morning Ironic!06:26
abongaleGood Morning Ironic!08:25
opendevreviewIvan Anfimov proposed openstack/ironic master: Remove tags from README  https://review.opendev.org/c/openstack/ironic/+/94847908:48
queensly[m]Good morning08:59
queensly[m]Hi JayF I saw your bug https://bugs.launchpad.net/ironic/+bug/2072364 report about explicitly documenting the documentation contributor processes and I'm planning to work on addressing this issue. Per the bug description, OWNERS or MAINTAINERS file need to be created in the docs directory. I would like to know who the OWNERS are and how do I write it in the file. Is there a format (eg name, email, name on IRC channel etc ) Any09:02
queensly[m]suggestions you have would be greatly appreciated!09:02
opendevreviewIury Gregory Melo Ferreira proposed openstack/ironic-specs master: NIC Firmware Updates  https://review.opendev.org/c/openstack/ironic-specs/+/94850310:10
iurygregorygood morning ironic10:10
opendevreviewIury Gregory Melo Ferreira proposed openstack/ironic-specs master: NIC Firmware Updates  https://review.opendev.org/c/openstack/ironic-specs/+/94850310:17
opendevreviewQueensly Kyerewaa Acheampongmaa proposed openstack/ironic master: doc: Add contributor guide for making documentation changes  https://review.opendev.org/c/openstack/ironic/+/94854110:25
opendevreviewQueensly Kyerewaa Acheampongmaa proposed openstack/ironic master: doc: Add contributor guide for making documentation changes  https://review.opendev.org/c/openstack/ironic/+/94854110:27
Sandzwerg[m]Morning ironic o/13:46
Sandzwerg[m]I have a question regarding secureboot. If the capability="secure_boot:true" is set and I use the redfish and iPXE (or even virtual media) then ironic should disable/enable secure boot before each deployment, correct?13:48
TheJuliaSandzwerg[m]: did you sign yourself an ipxe binary?13:50
TheJuliaIronic *should* if I'm recalling correctly, that being said you'll need to do virtual media based deployments most likely or install a Machine Owner Key (MOK), and sign your ipxe binary.13:51
TheJuliaWe sort of expect operators to lean more towards leaving it on all the time though13:52
TheJuliaAlso, given whole industry verticals have a mandate to have it enabled at all times13:53
Sandzwerg[m]No I didn't sign any ipxe yet because my understanding was that with this I wouldn't have to (as the deployment that has issues with secure boot would work without it). At least short term that would be the fastest way for us to implement it. Midterm would be virtual-media or even HTTPS, althoug it seems our self-build IPA image is not secure boot ready yet.13:58
TheJuliaso I would be curious why we think the ipa ramdisk shouldn't work out of the box, that is likely a distinct bug which should be worked. Doing as your leaning towards should work though if it is just on the flavor. 14:00
Sandzwerg[m]This time it switched secure boot on during the deployment and then disabled it again for the deployed image, not sure what happens there yet. We haven't tagged the flavor yet. Does that needs to be tagges secure boot as well?14:08
Sandzwerg[m]Is there a centos-ipa build around that should work with secure boot? We used to use the provided tinycore images some years ago but then switched to building our own when we had issues with hardware, but we never used the disk-image builder14:09
TheJuliaSo, the secure_boot field is the node overall, on deployment should still work from the capability as supplied from the flavor14:10
TheJuliatinycore is purely for upstream testing, please use anything other than those images14:10
Sandzwerg[m]yeah we're now build something based on fedora but we build it with mkosi so we deviate there from ironic as well ^^''14:13
Sandzwerg[m]But for now a known-good secure-boot enabled image would work for me. Just to test if it works at all. Afterwards I can try to get our image in shape14:13
cardoeSo looking at https://docs.openstack.org/ironic/latest/install/configure-pxe.html instead of telling people to copy stuff should we use loader_file_paths?14:14
TheJuliacardoe: likely14:14
cardoeCause I find that ironic sometimes tries to be helpful and clean up snponly.efi for me but loader_file_paths seems to make it not do that. So I'd update the docs if you all agree.14:15
TheJuliaSandzwerg[m]: ahh, I'd expect centos to work, the kernel will be signed and all the kernel artifacts should also be appropriately signed afaik14:15
Sandzwerg[m]yeah but is there a ready centos&IPA image somewhere?14:18
TheJuliahttps://tarballs.opendev.org/openstack/ironic-python-agent/dib/14:18
Sandzwerg[m]That's what I was looking for, thank you very much14:20
*** dking is now known as Guest1484014:48
*** Guest14840 is now known as dking15:02
opendevreviewKaifeng Wang proposed openstack/ironic master: Allow to unprovision instance from service wait states  https://review.opendev.org/c/openstack/ironic/+/94830115:14
JayFhttps://zuul.opendev.org/t/openstack/build/9d3bb67618f84180abca5f65d88e1b5b metal3 jobs are broken in a way that appears systemmatic15:51
JayFeither bugfix branch no longer compatible with updated constraints or constraints not being used to install it seems to me15:52
JayFrpittau: ^ FYI15:52
JayF+ python3 -m pip install --no-cache-dir --ignore-installed --prefix /usr -r /tmp/ironic-packages-list-final -c /tmp/upper-constraints.txt15:53
JayFit's using constraints :(15:53
JayFmaybe it's just temporary and a recheck will fix 15:53
fricklerthat's with py39? we dropped py39 from u-c16:12
JayF> + sudo update-alternatives --install /usr/bin/python python /usr/bin/python3.12 116:15
JayFis in the log16:15
JayFso if it's not using 3.12 it'd be a bug I believe16:15
fricklerJayF: if I read things correctly, the install is happening in a container and e.g. doing this before, so that looks very much like python3.9 to me: Downloading greenlet-3.1.1-cp39-cp39-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl.metadata16:19
JayFoooh yeah the cp39 16:20
JayFgood catch16:20
JayFrpittau: dtantsur: ^^ it looks like metal3-dev-env is trying to install Ironic with python 3.9, which is unsupported and breaking CI16:20
opendevreviewJulia Kreger proposed openstack/ironic master: WIP: Patch configdrive metadata  https://review.opendev.org/c/openstack/ironic/+/94667716:52
dtantsurJayF: I'm not sure what to do about it: we're not ready to switch fully to CS10 yet17:20
dtantsur(I'm not sure Ironic even supports CS10 officially)17:20
dtantsurFYI I"ll be out on Thu and Fri17:21
JayFThat sounds like we should mark it nonvoting in the meantime, then?17:21
dtantsuryeah, and wait for Riccardo to return from PTO17:21
JayFAh, didn't know he was out17:21
JayFbut I'm a bit discombobulated from missing earlier this week sick17:21
dtantsurI hope you're better now17:21
JayF(if back issues preventing computing counts as 'sick')17:21
JayFeh, I can sit and use a laptop all day17:22
JayFwhich is better than I was :)17:22
JayFI would tell you not to get old, but it is better than the alternative17:22
dtantsurright17:24
TheJuliaSo perhaps the thing to do is to nv the job in the mean time? It seems like the job needs undergo at least some updates outside of the default minimums17:35
TheJuliawell, defaults/minimums17:35
JayFI'm doing that now17:41
opendevreviewJay Faulkner proposed openstack/ironic master: Make metal3 job non-voting temporarily  https://review.opendev.org/c/openstack/ironic/+/94856817:44
dxterslabhas anyone worked with SMC redfish iKVM and been successful using noVNC with the websocket provided from it?19:35
TheJuliadxterslab: kind of... and not really19:45
TheJuliadxterslab: so the tl;dr is the patch for encoding to novnc never merged19:45
TheJuliadxterslab: so the solution stevebaker[m] worked on was to leverage an explicit vnc service in the middle to help navigate it from a browser instance19:46
dxterslabWhere can I find the changes that were proposed by stevebaker?20:28
dxterslabI think I found it20:30
stevebaker[m]dxterslab: Supermicro has a non standard implementation of VNC, so stock noVNC doesn't work. There is a fork of noVNC which does20:37
stevebaker[m]dxterslab: https://github.com/kelleyk/noVNC20:37
stevebaker[m]the ironic graphical console implementation for Supermicro is not nearly as robust as iLO or iDRAC. It simulates a user logging in and clicking on the console thumbnail image20:40
JayFSupermicro? a non-standard implementation!? [shocked pikachu]20:40
stevebaker[m]lol right?20:40
TheJuliaAnyone ever setup a spine/leaf setup with devstack ?20:51
opendevreviewMerged openstack/ironic master: Make metal3 job non-voting temporarily  https://review.opendev.org/c/openstack/ironic/+/94856820:53
opendevreviewMerged openstack/ironic master: docs: update openstackdocstheme version to 3.5.0  https://review.opendev.org/c/openstack/ironic/+/94824620:53
opendevreviewKulsoom Saboor proposed openstack/bifrost master: Add note about registering nodes  https://review.opendev.org/c/openstack/bifrost/+/94859120:56
opendevreviewKulsoom Saboor proposed openstack/bifrost master: Add note about registering nodes  https://review.opendev.org/c/openstack/bifrost/+/94859122:39
opendevreviewMerged openstack/ironic master: CI: Coverage for neutron with automated cleaning  https://review.opendev.org/c/openstack/ironic/+/94753523:56

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!