| rpittau | good morning ironic! o/ | 06:29 |
|---|---|---|
| queensly[m] | Good morning | 07:33 |
| abongale | gm ironic! | 08:24 |
| opendevreview | Syed Haseeb Ahmed proposed openstack/ironic master: Add interface name to inventory for inspection hooks https://review.opendev.org/c/openstack/ironic/+/953414 | 08:34 |
| frickler | I'm having an issue with ipa from master, it fails to find any ethernet interfaces' mac on my cisco boxes. version from about 2 weeks ago is working fine, stable/2025.1 too. anybody seen something similar? good ideas for debugging? sadly remote access to the site is a bit ... complicated | 10:16 |
| dtantsur | frickler: a side effect of https://opendev.org/openstack/ironic-python-agent-builder/commit/e06a6cf64cc8dee786ae728624afbff758939809? | 10:51 |
| dtantsur | could be https://opendev.org/openstack/ironic-python-agent/commit/9db3cd1e4d72462eb9303fd917f2d19a823cf4f0 too but I hardly see how | 10:51 |
| dtantsur | these are two directions I'd be looking into | 10:52 |
| TheJulia | cardoe: okay, then perhaps an explicit disabled field and we just add that to the logic today to ignore the port entirely | 13:21 |
| TheJulia | Also, good morning | 13:21 |
| dtantsur | morning TheJulia | 13:25 |
| opendevreview | Merged openstack/ironic-prometheus-exporter master: Zuul: do not use USE_PYTHON3 https://review.opendev.org/c/openstack/ironic-prometheus-exporter/+/953498 | 13:32 |
| frickler | dtantsur: thx, I've been looking at the second patch already, but it does look fine for me just from the sight of it. maybe I'll need to do some bisecting. the builder change is unlikely, since building from stable/2025.1 is not affected | 13:38 |
| dtantsur | frickler: you use master builder for stable/2025.1? | 13:48 |
| frickler | dtantsur: well until yesterday I only built ipa from master, because it just worked. when I noticed the regression, I just did set the dib reporef var to switch ipa+reqs to 2025.1, and that worked fine again. I wasn't aware that the builder is branched, too, until now | 13:58 |
| TheJulia | frickler: I guess the big question is do these network interfaces work at all? If they require firmware blobs to initalize, that would be a pretty solid data point | 15:34 |
| frickler | TheJulia: I don't think they need firmware blobs, but I also didn't understand what exactly the builder patch does. the interfaces are all present as expected in lshw output, just somehow the master ipa fails to find them | 15:37 |
| TheJulia | weird | 15:37 |
| TheJulia | so the builder patch just strips excess firmware blobs out to make the ramdisk size drop | 15:37 |
| dtantsur | frickler: you have lshw, does it mean you also have the agent logs? | 15:48 |
| dtantsur | TheJulia: drafted my ideas around scaling sensor data https://bugs.launchpad.net/ironic/+bug/2117178 (can be applied to power actions later) | 15:50 |
| frickler | dtantsur: I can look at them, but not copy&paste from them. or could, currently everything worked with the 2025.1 ipa. sadly with this issue ipa also doesn't get far enough such that it could save logs on the conductor | 15:51 |
| dtantsur | Got it. Do you have any downstream hardware managers in IPA? | 15:52 |
| frickler | I don't think so. from a screenshot the relevant message seems to be: WARNING: ironic_python_agent.hardware - Cannot find detailed information about interface ens1f0np0: StopIteration | 15:54 |
| frickler | which is the first of 8 interfaces. this is repeated a couple of times, none of the other ifs is mentioned in the log | 15:55 |
| dtantsur | This is about inability to find this interface in lshw; not fatal | 15:56 |
| frickler | but the interface is there. and in the end the list of gathered mac addresses is empty and thus the node fails to identify itself to the conductor | 16:04 |
| opendevreview | Syed Haseeb Ahmed proposed openstack/ironic master: Add interface name to inventory for inspection hooks https://review.opendev.org/c/openstack/ironic/+/953414 | 16:34 |
| TheJulia | frickler: easy way to check, you could set IPA_REMOVE_FIRMWARE to an empty string on building a new ramdisk, and it will keep all the firmware blobs | 17:29 |
| rm_work[m] | off-topic, anyone happen to know the state of Zun? TheJulia maybe you have had some interaction with them? | 18:00 |
| TheJulia | I don't think it really got off the ground beyond the original authors/consumers. | 18:02 |
| TheJulia | dtantsur: your dynamic pool patch seems to kind of work, letting my local heavily patched ironic roll for a while :) | 18:30 |
| cardoe | TheJulia: you good with the style in https://review.opendev.org/c/openstack/ironic/+/953414 or I can refactor that to combine the summary and the other mock. | 19:15 |
| cardoe | Just trying to get Haseeb over the finish line with this one. | 19:15 |
| TheJulia | cardoe: LGTM at this point | 19:37 |
| cardoe | Thanks. I gave it my +2 as well. | 20:08 |
| opendevreview | Nahian Pathan proposed openstack/sushy master: Support expanded Chassis and Storage for redfish https://review.opendev.org/c/openstack/sushy/+/955211 | 21:03 |
| opendevreview | Nahian Pathan proposed openstack/sushy master: Support expanded Chassis and Storage for redfish https://review.opendev.org/c/openstack/sushy/+/955211 | 21:05 |
| opendevreview | Nahian Pathan proposed openstack/sushy master: Support expanded Chassis and Storage for redfish https://review.opendev.org/c/openstack/sushy/+/955211 | 21:07 |
| TheJulia | dtantsur: so, overall your futurist chagne actually seems to do exactly what is needed minus what was already noted on change. Ironic just sitting syncing power runs 26-30 threads. Some further worker changes and need to revisit the rejection logic. | 21:24 |
| TheJulia | Overall spread on requests, does semi-match just eventlet... although maybe a little (oddly) slower, but... my local ironic likely needs a fresh state and can re-measure, but I do think the futurist change is the right path at this point as it aligns and keeps ironic itself from having to watch over/modify the thread pool. | 21:26 |
| TheJulia | which also seems like a bad idea | 21:26 |
| opendevreview | Nahian Pathan proposed openstack/sushy master: Support expanded Chassis and Storage for redfish https://review.opendev.org/c/openstack/sushy/+/955211 | 22:04 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!