| opendevreview | Merged openstack/ironic master: Correct Image properties lookup for paths https://review.opendev.org/c/openstack/ironic/+/852694 | 01:07 |
|---|---|---|
| opendevreview | Merged openstack/ironic master: CI: Changes to support Anaconda CI jobs https://review.opendev.org/c/openstack/ironic/+/849587 | 01:07 |
| opendevreview | Jay Faulkner proposed openstack/ironic bugfix/21.0: Correct Image properties lookup for paths https://review.opendev.org/c/openstack/ironic/+/858459 | 02:26 |
| opendevreview | Merged openstack/ironic-inspector master: Ignore connection errors when getting ironic ports https://review.opendev.org/c/openstack/ironic-inspector/+/836847 | 02:32 |
| JayF | should https://review.opendev.org/c/openstack/ironic-inspector/+/791251 and https://review.opendev.org/c/openstack/ironic-inspector/+/836847 be backported? Does anyone even use inspector on stable (has to be a yes, right?) | 02:32 |
| lecris[m] | TheJulia: dtantsur: Unfortunately the newer bios and ipmi does not fix the issue | 03:55 |
| TheJulia | What about network cards? | 03:56 |
| lecris[m] | TheJulia: Don't have external cards to test and I don't see how to update the firmware for that | 04:12 |
| lecris[m] | One thing worth checking is, is it possible to combine dhcpopts and regular conf file? | 04:17 |
| lecris[m] | Oh, but I see the tags for pxe are added to each individual port | 04:19 |
| opendevreview | SONG SHUKUN proposed openstack/ironic master: Add support auth protocols for iRMC https://review.opendev.org/c/openstack/ironic/+/857035 | 04:45 |
| lecris[m] | Is there an issue open for migrating from grub2-install to grub2-mkconfig? | 04:55 |
| lecris[m] | This issue is popping up: https://bugs.launchpad.net/diskimage-builder/+bug/1930878 | 04:55 |
| TheJulia | We place the assets and set the Nvram entry now. I think we even have a patch out there or did to flat out ignore grub2-install | 05:02 |
| TheJulia | lecris[m]: w/r/t combining, not really. The issue is ultimately for some reason the card explicitly doesn’t read/care about field length, and leans on the null value as a terminator… something I’ve personally never seen before | 05:03 |
| TheJulia | JayF: many… I would backport them to wallaby | 05:05 |
| * TheJulia goes to bed… because of hours of calls tomorrow | 05:07 | |
| lecris[m] | <TheJulia> "We place the assets and set..." <- Is this for nova deployed images built from disk-image-builder. So why am I getting grub2 issues? | 05:07 |
| lecris[m] | Kay, thnx for the help TheJulia: | 05:08 |
| TheJulia | Wholedisk or partition images? | 05:08 |
| lecris[m] | Not sure, I've built with baremetal and added kernel and ramdisk | 05:09 |
| TheJulia | And are their contents already valid? Dib images should be fine, but we have made numerous updates as grub maintainers have been pushing things such as grub2-install behavior changes as they break signing for assets thus breaking secure boot | 05:10 |
| lecris[m] | I'll take another look at the docs to see if I'm doing all the steps correctly | 05:11 |
| TheJulia | lecris[m]: heading down the wholedisk image path is highly recommended. If your setting a kernel/RAM disk then it is a partition image | 05:11 |
| lecris[m] | Thnx for the hint | 05:12 |
| TheJulia | And EFI booting looks for a partition. Ironic’s agent code does try to preserve/place efi booting assets if they are in the image, but some cloud image bases don’t include those binary assets, fwiw | 05:13 |
| TheJulia | Anyway, sleep() | 05:14 |
| adam-rozman | Good day Ironic ! | 06:03 |
| arne_wiebalck | Good morning adam-rozman and Ironic! | 06:07 |
| opendevreview | Ebbex proposed openstack/bifrost master: Use a more traditional ansible approach to include_vars https://review.opendev.org/c/openstack/bifrost/+/855806 | 06:22 |
| opendevreview | Ebbex proposed openstack/bifrost master: Refactor use of include_vars https://review.opendev.org/c/openstack/bifrost/+/855807 | 06:22 |
| rpittau | good morning ironic! o/ | 07:51 |
| lecris[m] | Trying to boot an ironic instance from volume but my cinder backend is ceph and the deploy complains that it is not iscsi | 08:20 |
| lecris[m] | I've seen that there is rdb_iscsi but not sure what that enables | 08:20 |
| lecris[m] | So how is openstack supposed to be configured here? | 08:21 |
| opendevreview | KOBAYASHI DAISUKE proposed openstack/ironic stable/wallaby: Check Fujitsu backport CI https://review.opendev.org/c/openstack/ironic/+/858352 | 09:41 |
| dtantsur | folks, https://review.opendev.org/c/openstack/bifrost/+/858391 is a bit of a mess, but it's needed to fix stable/yoga | 09:42 |
| ajya | TheJulia: it should be updated. What is is used for? Should distribution list that we manage internally would be better than one individual? | 10:16 |
| opendevreview | Jakub Jelinek proposed openstack/ironic-inspector master: Introduce skip list to inspector https://review.opendev.org/c/openstack/ironic-inspector/+/853304 | 10:23 |
| iurygregory | good morning Ironic | 11:05 |
| iurygregory | lecris[m], I never used the boot from volume, but there is some docs around it https://docs.openstack.org/ironic/latest/admin/boot-from-volume.html | 11:10 |
| dtantsur | lecris[m]: yeah, you'll need to use the iSCSI feature (whatever it's called - I'm also not strong in this area). Which, I'm afraid, Ceph is slowly deprecating :( TheJulia is again the best person to talk to. | 11:16 |
| lecris[m] | I guess it's a long way till we could mount ceph volumes in ironic? | 11:18 |
| dtantsur | lecris[m]: well... I don't think iPXE understands rdb, which leaves us without many options. | 11:18 |
| dtantsur | The idea was for Ceph to offer iSCSI support, which iPXE understands and can boot from. | 11:19 |
| dtantsur | It has been developed, but then they apparently backed away from this idea. I don't know why. | 11:19 |
| lecris[m] | Wouldn't it be possible to boot a lightweight agent to mount and configure from there? | 11:20 |
| dtantsur | lecris[m]: and then kexec into the actual OS? possible, but really error-prone and difficult to debug. | 11:21 |
| dtantsur | also no idea of an rbd mount survives kexec | 11:21 |
| lecris[m] | True, I've had my fair share when using warewulf | 11:21 |
| dtantsur | lecris[m]: the good news is: if you know how to do it in your environment, you can probably emulate it with https://docs.openstack.org/ironic/latest/admin/ramdisk-boot.html | 11:22 |
| dtantsur | so, the idea is not unheard of, based on quick googling: https://www.mail-archive.com/ceph-users@ceph.io/msg10371.html | 11:24 |
| lecris[m] | Hmm looks promising | 11:25 |
| lecris[m] | Isn't it already possible to have controller node translate ceph volumes to iscsi? | 11:28 |
| dtantsur | I cannot answer this unfortunately | 11:32 |
| lecris[m] | Well at least I've managed to get the whole stack running minus some features I can work around for now | 11:36 |
| lecris[m] | Still puzzled by the dnsmasq dhcp issue though | 11:37 |
| *** tosky is now known as Guest1012 | 12:14 | |
| *** tosky__ is now known as tosky | 12:14 | |
| zigo | Hi there ! Building Ironic 21.0.0 is no problem under Debian Unstable, but building the backport fails with these unit tests: | 12:28 |
| zigo | https://paste.opendev.org/show/blI3zZWCpsHL5VqnhKr7/ | 12:28 |
| zigo | Does anyone have a clue which python lib needs updating ? | 12:28 |
| zigo | dtantsur: ^ | 12:28 |
| zigo | TheJulia: ^ | 12:29 |
| iurygregory | zigo, I remember we have to fix some problems for jsonschema | 12:29 |
| zigo | Oh, jsonschema 4 ? | 12:29 |
| zigo | No, got the same version (backported the one from Unstable too...). | 12:30 |
| *** tosky is now known as Guest1013 | 12:30 | |
| *** tosky__ is now known as tosky | 12:30 | |
| iurygregory | yeah 4 something had some problems and we fixed | 12:30 |
| iurygregory | there is also the problem with oslo.db oslo.db==12.1.0 | 12:31 |
| iurygregory | =) | 12:31 |
| iurygregory | we merged a workaround for this already | 12:31 |
| zigo | The pip freeze is this: https://paste.opendev.org/show/bi8jHmEHq6JXeHNQR2fu/ | 12:32 |
| zigo | iurygregory: Oh, it was broken with latest oslo.db ? | 12:33 |
| zigo | I need the patch then! | 12:33 |
| zigo | That could be it ... | 12:33 |
| iurygregory | zigo, a lot of things were broken with this oslo.db version :D | 12:33 |
| zigo | This ? https://github.com/openstack/ironic/commit/74795abf2ff972b2b848eee98426bcdfee2a9c20 | 12:34 |
| dtantsur | zigo: oslo.db is a likely candidate | 12:34 |
| dtantsur | that's the workaround for now | 12:34 |
| zigo | Let me try to force oslo.db 12.1.0 in the Debian Experimental build and see if it fails. | 12:34 |
| zigo | Thanks for the clues. | 12:35 |
| iurygregory | zigo, correct =) | 12:35 |
| iurygregory | good luck! | 12:35 |
| zigo | BINGO ! | 12:42 |
| zigo | Forcing oslo.db 12.1.0 fails the build, now trying the patch ... :) | 12:42 |
| zigo | Thanks again. | 12:42 |
| iurygregory | zigo, nice \o/ | 12:48 |
| kubajj | Good afternoon ironic! | 12:52 |
| kubajj | I just looked at Zuul and it says Post Failure, has anybody seen that before? https://zuul.opendev.org/t/openstack/build/5e35cd8742b24f38adde8fadde1dbb07 | 12:53 |
| iurygregory | kubajj, sometimes it can happen, you can comment "recheck post failure jobs" | 12:54 |
| iurygregory | without the "" =) | 12:54 |
| kubajj | iurygregory: thanks, will do | 12:54 |
| TheJulia | yay for putting meeting on the wrong day | 12:59 |
| TheJulia | Also... good morning everyone! | 12:59 |
| kubajj | good morning TheJulia | 13:00 |
| iurygregory | TheJulia, the PTG one? | 13:04 |
| iurygregory | we can move for tomorrow lol | 13:04 |
| TheJulia | huh?! | 13:04 |
| TheJulia | I don't have that on my calendar | 13:04 |
| TheJulia | eeek | 13:04 |
| iurygregory | you accepted yesterday O.o | 13:04 |
| iurygregory | company email .-. | 13:04 |
| TheJulia | Wait, is that the meeting in 2+ hours? | 13:04 |
| iurygregory | 2h 55min | 13:05 |
| iurygregory | 16 UTC | 13:05 |
| TheJulia | okay, I have that one | 13:07 |
| TheJulia | I thought I had a 4 hour block of meetings starting at 6am local time this morning | 13:07 |
| TheJulia | turns out it is only 3 hours starting at 7 am | 13:07 |
| iurygregory | O.O | 13:07 |
| iurygregory | "only" | 13:07 |
| zigo | IMO, it'd be nice to release new versions for ironic + ironic-inspector to fix this oslo.db compat 12.1.0 thing... | 13:11 |
| iurygregory | zigo, it will be released this week | 13:14 |
| zigo | gr8 | 13:14 |
| iurygregory | + zed stable/branch | 13:14 |
| dtantsur | iurygregory: I might have a free evening in the end, could you add me to the invite? | 13:19 |
| iurygregory | dtantsur, sure | 13:20 |
| iurygregory | done | 13:20 |
| opendevreview | Uemit Seren proposed openstack/sushy master: Implement allowed values for BootSourceOverrideEnabled https://review.opendev.org/c/openstack/sushy/+/850461 | 13:22 |
| opendevreview | Uemit Seren proposed openstack/sushy master: Implement allowed values for BootSourceOverrideEnabled https://review.opendev.org/c/openstack/sushy/+/850461 | 13:23 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Concurrent Distructive/Intensive ops limits https://review.opendev.org/c/openstack/ironic/+/840006 | 13:47 |
| TheJulia | iurygregory: might just be easier to edit https://review.opendev.org/c/openstack/ironic/+/857615 | 13:53 |
| TheJulia | as opposed ot waiting | 13:54 |
| rpittau | iurygregory: I guess there's no way to move the PTG meeting? | 13:58 |
| iurygregory | rpittau, well it would depend ... only 3 ppl voted so it makes a bit complicated, but I'm all for more ppl participating =) | 14:31 |
| iurygregory | TheJulia, ok, moved to +2 | 14:32 |
| iurygregory | JayF, TheJulia thoughts on what rpittau asked about moving the PTG meeting? | 14:32 |
| JayF | I'm mostly open during working hours today | 14:33 |
| iurygregory | tomorrow 15 UTC would work for both of you checking the doodle (not sure if still applies) | 14:33 |
| JayF | I have a meeting at 8am (in 27 minutes) | 14:33 |
| JayF | that's 8am my time, I am clear | 14:33 |
| * TheJulia is soooo confused | 14:33 | |
| JayF | TheJulia: AIUI: rpittau wants us to move the ptg meeting so he can attend. Does tomorrow @1500UTC aka 8am PT work? | 14:34 |
| TheJulia | but my calendar is open tomorrow morning :) | 14:34 |
| TheJulia | ++ | 14:34 |
| TheJulia | \o/ | 14:34 |
| iurygregory | ok, updating the invitation | 14:34 |
| rpittau | iurygregory: it's ok, I see also dtantsurgave his availability for today, don't want to reschedule for all of you just for me, and it would mean also thursday, not even tomorrow | 14:34 |
| * TheJulia does the one less meeting today dance | 14:34 | |
| iurygregory | rpittau, oh .-. tomorrow 15 UTC you can't? | 14:35 |
| rpittau | iurygregory: unfortunately not :/ | 14:35 |
| dtantsur | tomorrow is probably out of question for me as well (unless I'm still too sick to go to the driving school) | 14:36 |
| dtantsur | although.. 15 UTC is doable, if not for long | 14:36 |
| dtantsur | (same for Thu) | 14:36 |
| JayF | I can go backwards from that too, 14UTC works | 14:37 |
| iurygregory | for Thursday TheJulia doesn't have any free time it seems | 14:37 |
| iurygregory | at least for 15/16 UTC | 14:37 |
| TheJulia | I may have marked myself out on Thursday | 14:39 |
| TheJulia | Can look in a moment | 14:39 |
| TheJulia | I have a friend coming into town this weeeknd and going to hang out with them on Friday now that we’ve figured things out | 14:39 |
| iurygregory | nice =) | 14:40 |
| iurygregory | if 15 UTC on Thursday works for you I can update the invitation for everyone, otherwise we keep today's time | 14:41 |
| JayF | I'm clear all thursday | 14:41 |
| JayF | but will not be here Friday (it's my birthday and I'm taking it off) | 14:41 |
| iurygregory | no meetings on Friday lol | 14:42 |
| iurygregory | time off on birthday ++ | 14:43 |
| TheJulia | iurygregory: what your seeing is all of hte meetings I'm invited to but have not accepted | 14:43 |
| TheJulia | iurygregory: I'm free during my morning on Thursday | 14:43 |
| iurygregory | TheJulia, I'm only looking at the doodle XD | 14:43 |
| dtantsur | 15:00 Thu is fine | 14:44 |
| TheJulia | oh, I likely didn't answer that then | 14:44 |
| iurygregory | yeah =) | 14:44 |
| iurygregory | ok, updating to Thu 15 UTC | 14:44 |
| * iurygregory also sends email to openstack-discuss now | 14:45 | |
| rpittau | thanks all :) | 14:45 |
| TheJulia | speaking of PTG... it seems the operators have expressed interest in their being operator centric discussions hosted by project teams, so if we can carve an hour out of the schedule for operators that hopefully doesn't conflict with other teams, I think it would be good | 14:45 |
| iurygregory | yes | 14:46 |
| iurygregory | I was planning to discuss this during this session | 14:47 |
| opendevreview | Uemit Seren proposed openstack/sushy master: Implement allowed values for BootSourceOverrideEnabled https://review.opendev.org/c/openstack/sushy/+/850461 | 14:53 |
| iurygregory | Congratulations to JayF \o/ he will be the PTL for the Antelope cycle =D | 15:04 |
| JayF | I feel like the dog that caught the car, what do I do now ;) | 15:04 |
| JayF | seriously, I'll be leaning heavily on iurygregory, Julia, and anyone else who is willing to listen :D I've got a lot to learn over the next few months | 15:05 |
| iurygregory | JayF, we will help you =) | 15:05 |
| iurygregory | the release part is very well documented (I never had any problems :D) | 15:06 |
| rpittau | congrats JayF :) | 15:09 |
| JayF | I have a handful of stable reviews that should be quick and easy if folks want to help with some quick wins so I can get them another level down: https://review.opendev.org/c/openstack/ironic/+/858459 https://review.opendev.org/c/openstack/ironic/+/857771 https://review.opendev.org/c/openstack/ironic/+/856205 | 15:21 |
| JayF | https://review.opendev.org/c/openstack/ironic-python-agent/+/857772 | 15:21 |
| dtantsur | JayF: congrats/condolences, we're here to help! :) | 15:23 |
| arne_wiebalck | Thanks iurygregory for all your PTL work during the last cycle! | 15:24 |
| dtantsur | yep, iurygregory++ | 15:24 |
| arne_wiebalck | And congratulations JayF for both your new roles! | 15:24 |
| iurygregory | Thanks arne_wiebalck dtantsur =) | 15:35 |
| rpittau | bye everyone, see you tomorrow o/ | 15:39 |
| TheJulia | JayF: noticed I missed one thing on https://review.opendev.org/c/openstack/ironic/+/840006, fixed it this morning | 15:57 |
| JayF | :eyes: | 15:57 |
| TheJulia | wow that was some giant eyes | 15:58 |
| JayF | oh yeah, I reviewed this on mobile but didn't login and vote | 15:58 |
| TheJulia | ahh, yeah | 15:58 |
| JayF | TheJulia: it worked on your client? lol | 15:58 |
| TheJulia | yeah, irccloud | 15:58 |
| JayF | TheJulia: and also, giant bitmap emoji not being downsized is a buf in your font configuration | 15:58 |
| JayF | assuming you're on linux, yeah | 15:58 |
| TheJulia | 👓 | 15:58 |
| JayF | I do not want to admit how many hours I spent troubleshooting emoji on my gentoo install lol | 15:59 |
| JayF | it still doesn't do compound emoji (e.g. color modifiers for emoji, where the unicode is like brown-smile) | 15:59 |
| TheJulia | heh | 16:10 |
| * TheJulia makes fresh coffee before trying to actually be productive now | 16:10 | |
| opendevreview | Merged openstack/ironic master: tests: Add a WarningsFixture https://review.opendev.org/c/openstack/ironic/+/856345 | 16:11 |
| kubajj | iurygregory: fixed https://review.opendev.org/c/openstack/ironic-inspector/+/853304 | 16:13 |
| arne_wiebalck | We just instantiated the first ARM node with Ironic at CERN \o/ | 16:19 |
| TheJulia | \o/ | 16:20 |
| kubajj | arne_wiebalck: \o/ | 16:20 |
| iurygregory | kubajj, ack I will take a look | 16:21 |
| iurygregory | arne_wiebalck, congrats! | 16:21 |
| TheJulia | Any objection if I rev https://review.opendev.org/c/openstack/ironic/+/857615 to include a release note? | 16:21 |
| iurygregory | TheJulia, none on my side, happy to +2 again | 16:23 |
| iurygregory | and is such a small thing I'm also happy to +W =) | 16:24 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Fix nodes stuck at cleaning on Network Service issues https://review.opendev.org/c/openstack/ironic/+/857615 | 16:32 |
| TheJulia | posted and +2'ed | 16:32 |
| opendevreview | Julia Kreger proposed openstack/ironic-python-agent stable/wallaby: Fix software raid output poisoning https://review.opendev.org/c/openstack/ironic-python-agent/+/858464 | 16:39 |
| TheJulia | JayF: w/r/t https://review.opendev.org/c/openstack/ironic/+/857984 I think we have a terminology issue, happy to take a shot at fixing it real quick | 16:47 |
| opendevreview | Jay Faulkner proposed openstack/ironic-python-agent bugfix/8.1: Fix software raid output poisoning https://review.opendev.org/c/openstack/ironic-python-agent/+/858465 | 16:48 |
| JayF | TheJulia: ^^ that one needs landing before wallaby | 16:48 |
| JayF | TheJulia: either fix it or leave a comment and I can circle back | 16:48 |
| TheJulia | workflowed | 16:50 |
| JayF | TheJulia: I had a feeling something around that was wrong, ty | 16:51 |
| TheJulia | JayF: happy to revise in a few | 16:53 |
| JayF | I already have it open | 16:53 |
| TheJulia | ok | 16:53 |
| JayF | then looks like that ipa fix has conflicts going further back than wallaby :( | 16:53 |
| TheJulia | JayF: raid wasn't completely sorted until wallaby | 16:54 |
| JayF | aha, so probably just needs to stop at wallaby then, good stuff | 16:54 |
| TheJulia | yeah | 16:54 |
| TheJulia | so... w/r/t bugfix branches, did we ever actually abandon the old ones, because i thought we had a list to abandon | 16:55 |
| JayF | Honestly, that was something I was generally going to inquire about | 16:55 |
| JayF | including if we know if there are folks consuming them | 16:55 |
| JayF | because it does add quite a bit of overhead | 16:55 |
| TheJulia | RH consumes some of them, but generally it should only be the last few. We've got a couple years of them on IPA | 16:56 |
| JayF | Maybe this should be a PTG topic? | 16:57 |
| TheJulia | ++ | 16:57 |
| JayF | Only maintaining bugfix branches until the named-version of them is released would be one route to potentially go | 16:57 |
| JayF | so e.g. once zed is released, any bugfix branches would go away until we decided to cut the first pre-antelope bugfix branch? | 16:57 |
| JayF | idk | 16:57 |
| JayF | I have no insight into how folks consume these, and frankly am surprised given how painful upgrading openstack can be operationally that folks would do it more than twice a year lol | 16:58 |
| TheJulia | so, we actually wrote this down | 16:58 |
| TheJulia | and it summed up to the last 2 bugfix branches, but in practice... that has not panned out since use has lagged more than anticipated | 16:58 |
| JayF | Can you be more explicit about what you mean by "use has lagged" | 16:59 |
| opendevreview | Merged openstack/sushy master: Fix misuse of assertTrue https://review.opendev.org/c/openstack/sushy/+/856907 | 16:59 |
| TheJulia | like being bugfix-3 and not moving forward | 16:59 |
| JayF | okay, so we haven't done a good job of following up and actually retiring bugfix branches beyond the 2 we support? | 17:00 |
| TheJulia | That is my understanding as well | 17:02 |
| TheJulia | iurygregory: ^^ | 17:02 |
| JayF | ack; I can add that to my list as long as we confirm that's what we want. | 17:02 |
| JayF | actually, I'll add an item for Monday's meeting | 17:02 |
| TheJulia | sounds good to me | 17:02 |
| opendevreview | Jay Faulkner proposed openstack/ironic master: Document existence of non-production "fake" driver https://review.opendev.org/c/openstack/ironic/+/857984 | 17:06 |
| JayF | TheJulia: ^ is that better? | 17:06 |
| TheJulia | JayF: much | 17:08 |
| TheJulia | JayF: a minor nit/suggestion | 17:09 |
| JayF | ack, will fix that pluralization after finishing this rebase | 17:10 |
| JayF | probably 15m | 17:10 |
| TheJulia | k | 17:10 |
| TheJulia | no worries, still trying to fix my router | 17:10 |
| * TheJulia glaers at the big white 5G cube of doom not playing nicely with the load balancer | 17:10 | |
| * TheJulia attempts the age old reboot | 17:14 | |
| JayF | TheJulia: when you fix your router, if you have 5 minutes I could use a second pair of eyes | 17:33 |
| JayF | trying to get https://review.opendev.org/c/openstack/ironic/+/858459 back to bugfix/20.2, it needed a rebase, and now unit tests are failing | 17:33 |
| * TheJulia raises an eyebrow | 17:37 | |
| TheJulia | I8791623be95e7e47739ee051753de97eb0e5e2a3 needs to merge first | 17:41 |
| opendevreview | Merged openstack/ironic stable/xena: Do not reboot into nowhere after BIOS settings with fast-track https://review.opendev.org/c/openstack/ironic/+/856205 | 18:01 |
| JayF | OH | 18:06 |
| JayF | I was thinking it might be something like that, but didn't see anything in bugfix21.0 | 18:06 |
| JayF | I guess thi slanded before 21.0 was cut | 18:06 |
| JayF | good stuff ty | 18:06 |
| JayF | TheJulia: backported that, literally the exact same two tests are filing :| | 18:11 |
| opendevreview | Merged openstack/ironic-inspector master: Introduce skip list to inspector https://review.opendev.org/c/openstack/ironic-inspector/+/853304 | 18:12 |
| opendevreview | Merged openstack/ironic bugfix/21.0: Correct Image properties lookup for paths https://review.opendev.org/c/openstack/ironic/+/858459 | 18:12 |
| JayF | ooh bad rebase, I love it when it's my fault | 18:12 |
| TheJulia | woot :( | 18:30 |
| TheJulia | okay, router fixed | 18:30 |
| TheJulia | hopefully this will work better now | 18:30 |
| JayF | TheJulia: is it possible there's yet-another-prereq for that backport? | 18:31 |
| * JayF is about to go digging down git logs | 18:31 | |
| TheJulia | quite | 18:31 |
| TheJulia | I'd git blame the conflicting area | 18:31 |
| JayF | the conflicts are all in tests | 18:31 |
| JayF | but when they are fixed; the tests fail | 18:31 |
| TheJulia | hmm :\ | 18:31 |
| JayF | aha | 18:32 |
| JayF | https://github.com/openstack/ironic/commit/33bb2c248a20d6a2a0af570655124cbc86d58b6a | 18:32 |
| JayF | it's that, almost certainly | 18:32 |
| JayF | yeah, and I need to rebase out the test that is needed for that, maybe | 18:33 |
| JayF | okay this makes sense | 18:33 |
| JayF | unit tests failing because the branch doesn't have the feature, who woulda thought :| | 18:34 |
| opendevreview | Jay Faulkner proposed openstack/ironic bugfix/20.2: Fix pxe image lookups https://review.opendev.org/c/openstack/ironic/+/858572 | 18:38 |
| JayF | that also means that https://review.opendev.org/c/openstack/ironic/+/858459 shouldn't be backported beyond 21.0, since https://github.com/openstack/ironic/commit/33bb2c248a20d6a2a0af570655124cbc86d58b6a didn't exist in 20.2 | 18:39 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/ussuri: CI: fix ussuri branch https://review.opendev.org/c/openstack/ironic/+/843515 | 18:40 |
| opendevreview | Jay Faulkner proposed openstack/ironic master: Document existence of non-production "fake" driver https://review.opendev.org/c/openstack/ironic/+/857984 | 18:43 |
| opendevreview | Jay Faulkner proposed openstack/ironic bugfix/18.1: Do not reboot into nowhere after BIOS settings with fast-track https://review.opendev.org/c/openstack/ironic/+/858468 | 18:44 |
| TheJulia | JayF: https://github.com/openstack/ironic-specs/blob/master/specs/approved/new-release-model.rst <-- background | 18:54 |
| TheJulia | https://review.opendev.org/c/openstack/ironic/+/838592/2/ironic/conf/deploy.py <-- easy win?! | 19:01 |
| TheJulia | err | 19:01 |
| TheJulia | https://review.opendev.org/c/openstack/ironic/+/838592 | 19:01 |
| TheJulia | NobodyCam: can you chime in on https://review.opendev.org/c/openstack/ironic-python-agent/+/566544 w/r/t step name | 19:20 |
| JayF | TheJulia: sounds like my interpretation: bugfix/[] branches should die when it's named-release comes out seems right? | 19:20 |
| TheJulia | I believe that was the original intent/hope | 19:21 |
| JayF | ack; in that case I'll still say something in meeting and we should probably cut final releases for those bugfix branches when zed is cut | 19:21 |
| opendevreview | Merged openstack/ironic master: Document existence of non-production "fake" driver https://review.opendev.org/c/openstack/ironic/+/857984 | 19:32 |
| iurygregory | TheJulia, sorry for some reason I missed your ping looking at other things, for bugfix branches we added to the etherpad the ones that are in use cc JayF | 20:00 |
| JayF | Yeah; it looks like the written policy says we should be retiring them once the letter-release they relate to is out | 20:01 |
| iurygregory | yeah | 20:01 |
| JayF | right now we violate that in a couple of places ... I'm just trying to ensure the extra churn we pay for stable branch maintenance is worth it | 20:01 |
| JayF | afaict, for ironic, we need to retire 19.0 and 18.0, and for IPA, 8.3 and 8.1 -- I haven't looked at the other repos yet | 20:02 |
| JayF | I'm happy to do this, but was going to wait until we said it at a meeting | 20:02 |
| JayF | (and hopefully there won't be objection to retiring the zed-related bugfix branches fairly quickly after zed is cut) | 20:03 |
| JayF | honestly, I don't care how many of these we have as long as we actually think people are using them and that people *should* use them | 20:03 |
| iurygregory | bugfix/19.0 we do use | 20:03 |
| JayF | for that second one -- I don't see any reason someone would run 19.0 instead of yoga | 20:03 |
| JayF | lol curious why not yoga in that case :D | 20:03 |
| JayF | I'm glad you are the counterexample, I wanna understand the motivation | 20:04 |
| iurygregory | yeah, basically is to avoid shipping new features | 20:05 |
| JayF | Like, let me put it this way: if you don't want new features; why not xena? If you do, why yoga? | 20:06 |
| JayF | in fact, afaict, our release doc as written specifically discourages people from running a bugfix release when the named one exists | 20:06 |
| iurygregory | in this case we landed a feature after xena that we claim we support downstream | 20:07 |
| JayF | but you can't put people on that release into yoga b/c it's not a "safe" upgrade per-se | 20:07 |
| JayF | honestly this written policy doesn't reflect reality at all | 20:08 |
| JayF | and we should fix it | 20:08 |
| NobodyCam | Will look in a bit | 20:09 |
| JayF | we support named branches **much** longer than 12 months, and we support bugfix releases for much longer than that oo | 20:09 |
| iurygregory | they are not really aware of "what version of Ironic" is in use | 20:09 |
| JayF | **too | 20:09 |
| JayF | I mean sure, but I have an upstream hat on, I don't have a red one to wear :D | 20:09 |
| JayF | so I'm trying to find a way to make this ... understandable without including RH shipped software in the explanation | 20:09 |
| TheJulia | NobodyCam: thanks | 20:09 |
| JayF | (and I'm very OK with making that explanation "work" for what RH ships, you all are the gateway to a huge amount of Ironic users; I just want our written things to make sense :D) | 20:10 |
| JayF | and bluntly, I've maintained downstream forks before; it's no fun and no good for anyone upstream or down | 20:11 |
| TheJulia | do metal3 folks, public folks that is, make use of our bugfix branches? | 20:12 |
| iurygregory | yeah we need to update https://github.com/openstack/ironic-specs/blob/2f4edc8320a0594402c41d44ebe4368ec43a5f16/specs/approved/new-release-model.rst | 20:13 |
| JayF | Honestly this conversation is me fishing for input so I can go update that :D | 20:14 |
| iurygregory | TheJulia, afaik nope | 20:14 |
| JayF | So bugfix/* branches are for: 1) standalone ironic users and 2) occassionaly they map directly to a downstream RH release, so you all want to keep it supported | 20:14 |
| TheJulia | iurygregory: not update/amend, but draft new perhaps? | 20:15 |
| JayF | that's reasonable in terms of, how the world actually works; I just wonder if 1 and 2 are the same set (e.g. do we have any users, not on rh, using bugfix releases?) | 20:15 |
| iurygregory | TheJulia, agree | 20:15 |
| * JayF realizes these answers are probably impossibleish to get | 20:16 | |
| TheJulia | JayF: yup :( | 20:16 |
| TheJulia | we *have* had issues in the past where we were blocked on getting a x.y version that was intermediate to a branch released because infra wouldn't let us create new intermediate tags/branches to fix it | 20:17 |
| TheJulia | and those were largely driven by folks who looked at our versions and went "I'll use that" | 20:17 |
| TheJulia | I *suspect* active bugfix branch maintenance has largely made those sorts of complaints disappear | 20:17 |
| TheJulia | because now they are not really orpahned like they were | 20:17 |
| TheJulia | and we actually are able to backport some fixes.... | 20:18 |
| JayF | honestly, I'm tempted to suggest we say bugfix releases go away when the named-version of those are out (e.g. 21.0/20.2 would go away when zed is released) | 20:18 |
| JayF | EXCEPTING if someone in the community has adopted it | 20:18 |
| * TheJulia wonders if we're just over-thinking and should just trim out the old bugfix branches | 20:18 | |
| JayF | or something along those lines | 20:18 |
| JayF | then we can sorta pick and choose but still have a policy that's reasonably consistent | 20:18 |
| TheJulia | JayF: ++ | 20:18 |
| TheJulia | I think that is a good way to put it | 20:18 |
| TheJulia | a short update that would be, fwiw | 20:18 |
| JayF | I mean, hell, when I worked places that didn't track master Ironic, I usually told them to use the stable/* release that RDO was using | 20:18 |
| JayF | because you were nearly-guaranteed to be on the long-end-border of support | 20:19 |
| JayF | I'm going to leave that as a meeting agenda item, and reword it a little, but as long as we're all on the same page in the meeting I'll push a new spec | 20:20 |
| JayF | > TheJulia wonders if we're just over-thinking and should just trim out the old bugfix branches | 20:21 |
| JayF | fwiw, I don't like this approach because as a project, we should like kinda consider ourselves an API? | 20:21 |
| iurygregory | I think rpittau dtantsur and I would live in paradise if we could move things to stable/zed XD | 20:21 |
| JayF | and not having policies, and following them, that other people can read means we have a bad api lol | 20:21 |
| * iurygregory brb coffee | 20:21 | |
| JayF | these are things I think about, even though they might not matter quite so much | 20:21 |
| TheJulia | JayF: I was more trying to convey that maybe what we are doing now does the needful and solves a lot of the problems, the issue by in large part is we've created a lot more backporting work which may not be needful | 20:22 |
| TheJulia | and going back/fixing some of these things when someone discovers a really bad bug is sometimes impossible without the pre-existing branchpoint | 20:23 |
| JayF | TheJulia: that's mainly why I brought it up as a topic, tbh. I feel like I do useful stuff when I backport to stable/[abc] but I feel like I'm throwing things into a black hole when they go into bugfix/[0-9] | 20:23 |
| TheJulia | i agree, it is nebulous until someone complains :( | 20:23 |
| JayF | I'm going to post to the mailing list and ask. More information never hurt anyone :D | 20:24 |
| TheJulia | good plan | 20:25 |
| TheJulia | now where did my brain go | 20:25 |
| iurygregory | I think we asked already if there were people using it | 20:29 |
| iurygregory | [ironic] Thoughts on closing some branches | 20:30 |
| iurygregory | 27 de jun | 20:30 |
| JayF | ah, and that's how we got to the currently-pruned list | 20:32 |
| iurygregory | yeah | 20:33 |
| opendevreview | Jay Faulkner proposed openstack/ironic bugfix/20.2: Fix pxe image lookups https://review.opendev.org/c/openstack/ironic/+/858572 | 20:37 |
| opendevreview | Julia Kreger proposed openstack/ironic-tempest-plugin master: Tempest test for anaconda deploy https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/854031 | 20:56 |
| opendevreview | Merged openstack/ironic master: Fix nodes stuck at cleaning on Network Service issues https://review.opendev.org/c/openstack/ironic/+/857615 | 20:56 |
| opendevreview | Merged openstack/ironic master: increase disk_erasure_coconcurrency https://review.opendev.org/c/openstack/ironic/+/838592 | 20:56 |
| JayF | iurygregory: TheJulia: rpittau: dtantsur: lolsob. First TC meeting? Thursday, right when our PTG topics meeting is | 20:58 |
| JayF | I could probably do both given it's an IRC meeting, but that's not ideal | 20:58 |
| iurygregory | JayF, ops XD | 21:00 |
| NobodyCam | oh I've never seen a comment post like that... wonder what I did | 21:00 |
| iurygregory | yeah, totally not ideal (you would need to switch context a lot I think) | 21:00 |
| JayF | I mean, it is what it is. I think I can swing it, and don't want us to have to move it (again) | 21:01 |
| opendevreview | Merged openstack/ironic-python-agent master: add openstack-python3-zed-jobs-arm64 job https://review.opendev.org/c/openstack/ironic-python-agent/+/854937 | 21:05 |
| iurygregory | WHAT? :eyes: | 21:16 |
| iurygregory | oh ok, zed-jobs-arm64 | 21:17 |
| iurygregory | jesus I was like "we didn't cut the release why we have this merged" LOL | 21:17 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/ussuri: CI: fix ussuri branch https://review.opendev.org/c/openstack/ironic/+/843515 | 21:29 |
| TheJulia | heh, better to have the jobs now than later :) | 21:29 |
| TheJulia | JayF: is the tc meeting going to be a call or a irc chat? | 21:31 |
| JayF | TheJulia: ... I assumed it's IRC?? | 21:34 |
| JayF | is that a bad assumption? | 21:34 |
| JayF | > Location: IRC OFTC network in the #openstack-tc channel | 21:34 |
| JayF | so yeah, it's IRC | 21:34 |
| TheJulia | have we written a prelude for Zed?! | 22:07 |
| TheJulia | no, we have no prelude | 22:07 |
| opendevreview | Julia Kreger proposed openstack/ironic master: add default conductor group capability https://review.opendev.org/c/openstack/ironic/+/855705 | 22:13 |
| * TheJulia gets out the words and a text editor | 22:19 | |
| JayF | TheJulia: "Now that we have re-leased Z, next time will be A you see" :P | 22:20 |
| * TheJulia groans | 22:20 | |
| JayF | (to the tune of the ending of the alphabet song, appropriately) | 22:20 |
| JayF | TheJulia: I think you should add a task for the 2nd thing to 2010267 so if/when that merges, it won't close yet | 22:21 |
| JayF | (re: your recently posted patch) | 22:21 |
| JayF | also might wanna put it up for RFE-approval next meeting, although I'm fine with it and it's obvious | 22:22 |
| TheJulia | I was like "surely we've done this" and nope | 22:22 |
| TheJulia | so "that was easy" | 22:22 |
| JayF | assuming current behaior is if you create a new node w/o specifying group but with groups configured, it goes kaboom? (API error/bad req) | 22:23 |
| TheJulia | JayF: current behavior is it gets created with none, and you need a conductor in the none group to change the node after the fact. | 22:28 |
| JayF | That smells like there's a validation bug in there too -- e.g. should I be able to add nodes to *any* conductor group that doesn't have conductors from that group registered in it? | 22:28 |
| JayF | like I would've hoped creating a "None" group'd node would be impossible if there were no "None" group'd conductors | 22:28 |
| JayF | can I create a "banana" conductor group node w/o having a "banana" conductor? | 22:29 |
| TheJulia | we don't keep a central list of them, and they can be changed | 22:29 |
| TheJulia | its a hash ring division key, really | 22:29 |
| JayF | ooh | 22:29 |
| JayF | that is ... unfortunate | 22:29 |
| TheJulia | yeah | 22:29 |
| JayF | there's no good way to fix it if that's the design | 22:30 |
| TheJulia | but... powerful too | 22:30 |
| JayF | Oh, believe me, I was already thinking of ways that you could (ab)use that | 22:30 |
| JayF | TBH I'm just thinking of the poor operator "oh hell, I added these to the producktion group by accident" and having no way to fix but touching DB or adjusting conductor config | 22:31 |
| TheJulia | I think arne hit that once | 22:31 |
| TheJulia | I *think* we likely need to have a fallback/retry on node_update because of that | 22:31 |
| TheJulia | and drivers | 22:31 |
| TheJulia | anything that is a hashring key should trigger a fallback imho | 22:32 |
| TheJulia | "oh! you wan todo x and that would have changed the conductor, lets fallback if not locked, try to lock and update it" | 22:32 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Zed: Add a prelude for the release notes https://review.opendev.org/c/openstack/ironic/+/858582 | 22:33 |
| JayF | I was thikning that too, but couldn't wrap my brain around how | 22:33 |
| JayF | I am ... less optimistic about that type of change after the work I did for heartbeat-only agent downstream at yahoo | 22:33 |
| TheJulia | ahh | 22:33 |
| TheJulia | I can see how it works mechanics wise in my mind, but I'm approaching the end of my day after getting started so early | 22:33 |
| JayF | of course, we don't have to solve anything now | 22:34 |
| JayF | just food for thought | 22:34 |
| JayF | go have an eod, see you tomorrow o/ | 22:34 |
| TheJulia | tomorrow we likely need to update the release mappings | 22:34 |
| TheJulia | and check release note content | 22:35 |
| JayF | btw, if I haven't said it "Zed" is like, my favorite Ironic release name so far | 22:36 |
| JayF | I have a thing for three-letter-words-pronounced-like-a-letter | 22:36 |
| TheJulia | I dislike it because the "zed is dead" joking that immediately started | 22:36 |
| TheJulia | but that is just me | 22:37 |
| JayF | I've not heard that except just now, from you | 22:37 |
| JayF | you have proliferated your hated meme! oh no :( | 22:37 |
| TheJulia | nooooooooooooooooooo | 22:37 |
| JayF | honestly, outside of openstack folks, the releases are all letters | 22:38 |
| JayF | you start throwing wallaby and victoria and all that around, people get confused, letters are understandable | 22:38 |
| JayF | I actually am looking very forward to the 2023.x naming convention starting up | 22:38 |
| * TheJulia wonders if the board is *actually* aware of that | 22:40 | |
| JayF | TheJulia: the new versioning stuff? The TC rez said we keep names for marketing (e.g. Antelope = 2023.1) | 22:42 |
| JayF | so like, does things change from a foundation standpoint? | 22:42 |
| * JayF probably would've referred us just use actual version numbers, but that's less fun anyway | 22:42 | |
| JayF | actually, I wonder how that impacts our bugfix releases | 22:42 |
| TheJulia | unlikely, since the name is the marketing unifier/delineator | 22:43 |
| JayF | if policy says releases should be versioned 2023.1 | 22:43 |
| JayF | then how do we make a bugfix off that? | 22:43 |
| TheJulia | we don't | 22:43 |
| TheJulia | oh wait | 22:43 |
| JayF | https://governance.openstack.org/tc/resolutions/20220524-release-identification-process.html it's unclear | 22:43 |
| TheJulia | are they mandating everyone tanks their versioning | 22:44 |
| TheJulia | ? | 22:44 |
| JayF | I'm saying it's unclear. | 22:44 |
| JayF | And we need to proactively be aware of how things are lining up w/release team for A | 22:44 |
| JayF | so that's not implemented as an implied requirement when it's not required by the resolution | 22:44 |
| TheJulia | it is the unified release | 22:44 |
| TheJulia | each project together, not each individual project | 22:45 |
| JayF | So I'm saying, right now, yoga Ironic is uploaded to pypi as like, probably 20.0 or something /me didn't look it up | 22:45 |
| TheJulia | and it should continue to be so | 22:45 |
| JayF | I'm wondering if this implies -- or if releaqse team will try to require us -- to publish antelope as 2023.1 | 22:45 |
| TheJulia | and if they try to change that, they are going to really screw up pypi | 22:45 |
| JayF | well, not really, as long as versions go up | 22:45 |
| JayF | which they would, moving to year-first in the release name | 22:45 |
| TheJulia | but it breaks versioning logic | 22:45 |
| JayF | the only thing that would break in a technical sense is our workflow (and other projects that might use it, if they exist) | 22:45 |
| TheJulia | we had to nuke the cirica 2014 versions for similar reasons | 22:46 |
| JayF | Hm. I don't think it'd be a break in this case (because existing versions are all going to be <2023) | 22:46 |
| JayF | but hopefully that experience will be in folks' mind if they consider doing that | 22:46 |
| JayF | and I'll keep an eye on it | 22:46 |
| TheJulia | OpenStack Release (as in called openstcak) != nova 23.0, ironic 22.0, etc | 22:47 |
| TheJulia | afaik, there is no longer a single OpenStack package on things like pypi... | 22:47 |
| TheJulia | (... there was once... it was addressed if memory serves) | 22:47 |
| JayF | Your comments all match my understanding. It's just weird to have a numerical name that is not the same as the numerical version | 22:49 |
| JayF | and I had a concern; possibly unfounded, that folks would try to unify them | 22:49 |
| TheJulia | This is... a worry | 22:49 |
| TheJulia | anyhow, time will tlel | 22:49 |
| JayF | just something to keep an eye on :) | 22:49 |
| TheJulia | I'm going to plugin the work lappy and start thinking about dinner | 22:49 |
| JayF | Obligitory "go make pizza for Summer" :D | 22:50 |
| JayF | o/ | 22:50 |
| TheJulia | not tonight, but I should make some dough | 22:50 |
| JayF | Vanessa has been making some pizza dough out of sourdough disard | 22:51 |
| JayF | *discard | 22:51 |
| JayF | I stretched one out real thin and made myself a bad NY-style pizza lol | 22:51 |
| TheJulia | nice | 22:52 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!