| cardoe | These test issues suck. ☹️ | 00:04 |
|---|---|---|
| cardoe | https://review.opendev.org/c/openstack/ironic/+/933194 that'll at least let us iterate on pre-commit with less failures. | 02:49 |
| cardoe | JayF: that _ missing bug happened with 39cd751a9063ae04e1649ccbca5496c3754d706e from 2021. | 02:52 |
| opendevreview | Verification of a change to openstack/ironic master failed: The i18n function is used but not imported https://review.opendev.org/c/openstack/ironic/+/937255 | 03:07 |
| *** diablo_rojo_phone is now known as Guest2639 | 08:38 | |
| opendevreview | Thomas Goirand proposed openstack/ironic master: Stop using crypt stdlib API https://review.opendev.org/c/openstack/ironic/+/936357 | 10:19 |
| opendevreview | Thomas Goirand proposed openstack/ironic master: Stop using crypt stdlib API https://review.opendev.org/c/openstack/ironic/+/936357 | 10:20 |
| adam-metal3 | Hello Ironic | 10:23 |
| adam-metal3 | JayF, TheJulia, rpittau: rpittau mentioned that I should bring it up right here. We have this issue in Metal3 https://github.com/metal3-io/baremetal-operator/issues/2110 , we are booting live ISO and detecting both iso and gpt image formats and failing with the deployment. We have looked into and AFAIK it comes from thre "recent" changes done in the oslo utils. I can fix this in IPA and Ironic (It | 10:26 |
| adam-metal3 | fails in Ironic for us but the inspection code is basically the same in IPA too) I have already started to write a fix . | 10:26 |
| adam-metal3 | BUT | 10:26 |
| adam-metal3 | I really really feel that "gpt" shouldn't be a separate "image format" in oslo_utils.imageutils rather we should just add the gpt "security check" to all the other image format's security checks | 10:28 |
| adam-metal3 | The floks from "nova" said themselvs that oficially only GPT is supported whatever image we boot, so I really don't see any reason to handle the partition table type as an image format | 10:29 |
| adam-metal3 | This is the olo change I am reffering to https://github.com/openstack/oslo.utils/commit/91af49beed6eea273863094bd62dcfb6688bfb11 and if you would like to go with the fix for only IPA and Ironic I would bring in the new newly introduced "formats" property https://github.com/openstack/oslo.utils/commit/3d4ae167cc1cc3509669a2c30f44e8d31ba39e3f to IPA and Ironic so that we can handle situations where | 10:33 |
| adam-metal3 | e.g. both gpt and ISO is detected | 10:33 |
| adam-metal3 | So what I am asking/saying is that, Is the current approach in oslo_utils (handling gpt as an image format) okay or not? If you say that you think the current oslo_utils approach is reasonable because of some wider openstack requirement I will just work on fixes for Ironic cumminty but if you feel like me that the oslo approach is just a bit weird I could work on a fix there | 10:37 |
| dtantsur | adam-metal3: ugh. Let's make sure we have a launchpad entry for this. | 11:22 |
| dtantsur | I feel like we shouldn't validate ISO images as well | 11:27 |
| adam-metal3 | dtantsur: I will create the launchpat issue, no problem | 11:28 |
| opendevreview | Thomas Goirand proposed openstack/ironic master: Stop using crypt stdlib API https://review.opendev.org/c/openstack/ironic/+/936357 | 13:30 |
| opendevreview | Thomas Goirand proposed openstack/ironic-python-agent master: Python 3.13: crypt.crypt is removed https://review.opendev.org/c/openstack/ironic-python-agent/+/937454 | 13:42 |
| opendevreview | Thomas Goirand proposed openstack/ironic-python-agent master: Python 3.13: crypt.crypt is removed https://review.opendev.org/c/openstack/ironic-python-agent/+/937454 | 13:43 |
| cardoe | I feel like someone else has opened some of those too? | 15:14 |
| zigo | tkajinam: Do you feel it's ok now? https://review.opendev.org/c/openstack/ironic/+/936357 | 15:33 |
| opendevreview | Thomas Goirand proposed openstack/ironic-python-agent master: Python 3.13: crypt.crypt is removed https://review.opendev.org/c/openstack/ironic-python-agent/+/937454 | 16:07 |
| JayF | zigo: I think you might not understand what the option you changed default on does there | 16:09 |
| JayF | zigo: I -1'd, I think we'll need a deprecation period for that change -- essentially that is fed directly into /etc/shadow on the rescue ramdisk; so if we change this without operators knowing it could break rescue | 16:09 |
| zigo | JayF: So I should switch back to sha-256 ?!? | 16:10 |
| JayF | If my understanding is right, we should billboard the deprecation (put something in the release notes, and a comment by the default), then we'll change it next cycle | 16:10 |
| JayF | It's always possible i'm wrong :) | 16:10 |
| zigo | At this point, I may give-up, I'm scratching my own itch of the package not working in Debian Unstable (ie: Debian 13, as Dalmatian will be there). Maybe someone form the Ironic team should take-over? | 16:11 |
| JayF | it's up to you, but I'll note we don't have python 3.13 support billboarded for this release and so it's likely a higher priority to you-personally than it is to openstack as a whole this cycle. I'm happy to review and help land, but I personally won't pick up and finish that patch anytime soon | 16:16 |
| JayF | I also know there's a py3.13 fix in eventlet that is pending release as well | 16:16 |
| zigo | I already have the eventlet patches backported to 0.36.1 ! :) | 16:17 |
| zigo | Even the one from Stefano that he wrote yesterday. | 16:17 |
| zigo | As for 3.13 in the gate, well, Maybe this will be added after Epoxy, though that's too late for me (since Debian 13 will be released with Python 3.13). | 16:18 |
| JayF | nice, so you probably are up to date | 16:18 |
| zigo | I believe so. :) | 16:18 |
| JayF | I was talking to itamar about those fixes yesterday | 16:18 |
| zigo | This fixed a bunch of failure in oslo.utils. | 16:18 |
| zigo | Sorry, in oslo.messaging. | 16:18 |
| zigo | I suppose you're talking about this one: https://patch-diff.githubusercontent.com/raw/eventlet/eventlet/pull/1000.patch | 16:19 |
| * zigo got to go back home | 16:20 | |
| JayF | I honestly don't know what I'm talking about, just I know itamar told me he was blocked on new eventlet release, which also was waiting another py3.13 patch to go in | 16:41 |
| JayF | etc etc | 16:42 |
| opendevreview | Verification of a change to openstack/ironic master failed: The i18n function is used but not imported https://review.opendev.org/c/openstack/ironic/+/937255 | 18:35 |
| *** atmark_ is now known as atmark | 19:37 | |
| opendevreview | cid proposed openstack/ironic master: [WIP] Agentless Kea DHCP backend. https://review.opendev.org/c/openstack/ironic/+/935645 | 22:54 |
| opendevreview | cid proposed openstack/ironic master: [WIP] Agentless Kea DHCP backend. https://review.opendev.org/c/openstack/ironic/+/935645 | 23:18 |
| JayF | I put this in #cluster-api-baremetal on slack as well, but metal3-integration is failing with 404 errors | 23:19 |
| JayF | https://github.com/metal3-io/metal3-dev-env hasn't had a commit in a week, so I assume it's broken | 23:20 |
| cardoe | which Slack? K8s? | 23:28 |
| opendevreview | Jay Faulkner proposed openstack/ironic master: CI: Detail job removal permission https://review.opendev.org/c/openstack/ironic/+/933151 | 23:34 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!