15:00:03 #startmeeting ironic 15:00:03 Meeting started Mon Dec 14 15:00:03 2020 UTC and is due to finish in 60 minutes. The chair is TheJulia. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:04 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:06 The meeting name has been set to 'ironic' 15:00:08 Good morning everyone! 15:00:17 o/ 15:00:22 Welcome to our last meeting of this year! 15:00:24 o/ 15:00:34 Our agenda can eb found on the wiki, as always. 15:00:37 #link https://wiki.openstack.org/wiki/Meetings/Ironic#Agenda_for_next_meeting 15:00:51 #topic Announcements / Reminders 15:00:51 o/ 15:00:55 o/ 15:00:57 o/ 15:00:59 o/ 15:01:10 #info No ironic meetings the next three weeks - See mailing list for details. 15:01:20 \o 15:02:11 tl;dr we're in our end of year autopilot mode. If your a core and you see a gate fix go up, please don't wait for another core reviewer to approve during this time. 15:02:58 #info According to our release plan, we should be shipping our first sprint work this week. I know there is lots of review activity, so please keep it up so we can hopefully get the release changes all posted this week. 15:03:30 #info PTL is off Friday of this week and will be around if needed the next two weeks. 15:03:32 rpittau and I have already started pre-release activities 15:03:46 Thanks! I had noticed this morning :) 15:04:01 * dtantsur is semi-around now 15:04:32 o/ 15:04:34 I'm personally going to be writing words. I put a 50k word goal on myself for this month and I'm like 1700 words in. 15:05:01 Does anyone else have anything to announce or remind us of? 15:05:24 We have no action items, so we will proceed directly to reviewing subteam status reports. 15:06:41 * TheJulia hears crickets 15:08:09 I guess we can proceed then! 15:08:13 #topic Review subteam status reports 15:08:22 #link https://etherpad.opendev.org/p/IronicWhiteBoard 15:08:37 Starting at line 274 15:08:56 dtantsur: will there be another spuc this friday? 15:11:32 kaifeng: arne_wiebalck I guess no updates on the node history beyond I think that the spec has merged? 15:11:56 no updates from me 15:13:00 bdodd: ajya|afk: Looks like ajya|afk updated the patch on the 9th, so I guess still in development for redfish raid? 15:13:37 hmm, no iury, I guess he is on pto now? 15:13:46 TheJulia Yes, I'm working on the comments from ajya|afk 15:14:36 rpioso: for the interop profiles, is there anything major missing? apart from more people trying? the change is still marked WIP 15:14:46 (and I still owe you the logs) 15:15:11 zer0c00l: Last I heard you were working on a psec udpate? 15:15:13 update 15:15:18 w/r/t anaconda 15:15:43 TheJulia: yes, not start yet 15:15:50 kaifeng: okay 15:16:44 TheJulia: don't know if zer0c001 will show up but yes, he said he was working on an update to that 15:16:53 rloo: ack 15:17:22 rpioso: ajya|afk: anything on configuration molds? 15:17:23 arne_wiebalck: More review of the management (sensor data) and boot (vmedia requirements) profiles is needed. Also, there are questions mraineri and I need to review. Following that, all need to be aggregated into one. 15:17:49 TheJulia: Yes, we updated https://review.opendev.org/c/openstack/ironic-specs/+/740721. Reviews before the holidays would be much appreciated. 15:17:52 rpioso: thanks! 15:18:01 arne_wiebalck: yw 15:18:10 rpioso: Ack, looks like it was updated Friday. 15:19:17 rpioso: I don't know how much review bandwidth is really left since people need to start dropping off to use PTO or for shutdown. I'll try to take a look this week though. 15:19:21 TheJulia: The most recent patch set was uploaded a week ago. 15:19:37 oh, I see it was a comment on the 11th 15:19:40 * TheJulia needs more coffee 15:19:43 TheJulia: Thank you! It should look very familiar :) 15:19:54 Was it just me, or was last week just fly by> 15:19:55 ? 15:20:14 Week, what week? 15:20:20 Do we want to review priorities for the week, or skip ahead to discussion. I ask because we're going into end of year autopilot mode 15:20:22 rpioso: exactly. 15:22:10 * rpioso belatedly updates Redfish Interop and config mold statuses on the White Board. 15:22:15 priorities for the rest of the year? 15:22:40 TheJulia: I'm planning on a SPUC, yeah 15:22:41 rloo: It goes into an anyone can update/add mode 15:22:47 although I cannot 100% guarantee that I'll be present 15:22:59 dtantsur: okay 15:23:02 rloo: fix l-c with new pip in at least victoria 15:23:06 TheJulia: Yes. It is on my plate 15:23:12 Will be doing that this week. 15:23:17 although i guess there are some PRs for this week, wrt release this week? 15:23:54 I guess they can be added to the top of the list? 15:24:23 dtantsur: rpittau: you indicated you were starting pre-release activities, any changes outside of last minute feature/fix merges expected? 15:24:46 IPA has a bunch of patches approved and slowly making it through the gate 15:24:56 ironic seems fine, I think (your IPMI patch is on teh way) 15:25:05 TheJulia: not on inspector, I've already proposed the release 15:25:29 rpittau: ack. 15:25:31 dtantsur: \o/ 15:25:58 Yeah, IPA has a bunch of things queued up, some of that is intertwined too :\ 15:26:37 so I think we're kind of good then on that front? 15:26:49 Which I guess means Discussion is next? 15:27:53 * TheJulia wonders if it time to switch the in-irc channel coffee machine to more seasonal beverages 15:28:12 vin brule? 15:28:44 * TheJulia shrugs 15:29:16 * TheJulia presses onwards 15:29:20 #topic Open Discussion 15:29:27 #undo 15:29:28 Removing item from minutes: #topic Open Discussion 15:29:29 #topic Discussion 15:30:18 We have one discussion item, regarding the lower-constraints issues. A discussion on the mailing list has lead towards the beginnings of a consensus that they provide no actual value to downstream packagers and have become nothing but a headache upstream. 15:30:38 I'm curious how people are feeling about the idea of removing lower-constraints jobs. 15:30:40 ? 15:31:23 minimum required package == lower constraints? 15:31:28 i guess we actually more care about upper-connstraits :) 15:31:33 yes 15:31:41 it's fun that now that the pip-resolver actually enforce version dependencies, we vote to remove lower-constraints :D 15:31:43 kaifeng: I do think so as well 15:31:46 i was wondering about non-voting jobs. but that might basically be as bad as no jobs 15:32:27 i think lower-constraints could be useful as the lower-bound. but... 15:33:47 I'm kind of the opinion that while yes it is useful as a lower bounds, we shouldn't be encouraging people to run older versions of libraries or updating things. Many libraries get fixes, and we test with those fixes. We interact with them upstream. We see resulting configuration differences say in what the config generator puts together based on that 15:34:20 dtantsur: any thoughts on this? 15:34:27 I just looked it up in our internal build system, we don't use lower constraints either. 15:34:35 I'm not against removing lower-constraints, it's true that the suggested min version kind of lose meaning in this way 15:36:36 I know sometimes packagers don't get newer version in, but if it still falls inside the bounds, then it doesn't get noticed. I guess it is useful for breaking upstream, but we're just updating the requirements.txt and lower-constraints at the same time then 15:37:09 i think if community is discussing, and there doesn't seem to be any great reason for keeping lc, we should do what makes things easier (so remove from what I can tell wrt discussion) 15:37:35 rloo: I think that is a good summary 15:37:55 could always be added if someone shows up with a great reason for it... 15:38:15 yup 15:38:17 mmm but we won't run tests with the min version anymore, it becomes more a started suggested version than a "we're sure it works at least with that" 15:38:26 I think that l-c is too much of a burden 15:38:40 this sort of feature should be built into pip to be usablwe 15:38:57 l-c's existance enourages us not to increase the main version, to find a version we think works. It creates a burden and yeah. :( 15:39:02 wonder if we can actually remove it NOW or if we need to inform users about deprecation/removal. will wait to see :) 15:39:06 dtantsur: ++ 15:39:17 I think we can remove it now 15:39:24 also if we remove it, we should remove it from stable branches too ? 15:39:26 we can just drop release notes in to the other field 15:39:35 rpittau: I think we will need to make them functional again 15:39:45 I was afraid of that 15:40:48 so the general idea is train,ussuri,victoria need working l-c, but we drop from wallaby ? 15:40:48 I feel like we have consensus, and can go to the Baremetal SIG 15:41:05 rpittau: I don't see why they need a working l-c 15:42:57 errr I'm confused now 15:43:07 how so? 15:44:26 I'm probably misunderstanding the intention here, so we drop l-c from all the branches? 15:44:48 yes, all branches 15:45:04 ok, thanks, then yes, I misunderstood :) 15:45:42 I think we maybe want to leave the file 15:45:47 on the branch, but just kill the job 15:46:52 Does that sound good with everyone? 15:47:00 yep 15:47:03 Okay then! 15:47:09 Onward to the SIG 15:47:12 #topic Baremetal SIG 15:47:18 arne_wiebalck: anything to announce today? 15:47:23 Next meeting scheduled for Tue January 12th at 2pm UTC - Topic is multi-tenancy (tzumainn) 15:47:26 or bring up? 15:47:28 \o/ 15:47:56 thanks once more to rpioso for the interop discussion 15:48:21 Verification of a change to openstack/ironic-python-agent failed: Correctly decode error messages from ironic API https://review.opendev.org/c/openstack/ironic-python-agent/+/765838 15:48:29 arne_wiebalck: My pleasure. Thanks to you for organizing and those who participated. 15:48:37 TheJulia: we will need to see about the videos at some point. 15:48:51 arne_wiebalck: yes, after the first of the year? 15:49:00 TheJulia: sounds good, thanks! 15:49:10 Dmitry Tantsur proposed openstack/ironic-python-agent master: Avoid a full install in tox environments that do not need it https://review.opendev.org/c/openstack/ironic-python-agent/+/766951 15:49:16 I think that's it for the SIG. 15:49:31 Okay then! 15:49:37 Onward to Open Discussion 15:49:54 #topic Open Discussion 15:50:14 So everyone working on taking over the world this holiday season? 15:51:06 I have a question ... Has anyone gotten DHCP-less Redfish vmedia boot working recently? 15:51:26 I couldn't get it to work at all 15:51:30 the ramdisk side, mostly 15:51:42 dtantsur: I believe we're encountering the same :-( 15:51:48 mmm, actually, yes, I was going to mark it as non-working in the docs unless somebody can tell me how it can be made working 15:51:54 probably before we do a release.. 15:52:17 rpioso: you're seeing the network_data passed correctly but not actually applied? 15:52:35 dtantsur: Yes, and DHCP is being used. 15:52:41 I guess we need to get a few people digging into it after the new year :\ 15:52:48 yeah 15:53:16 I'll put some vague wording about "the feature is there but the ramdisk side is not" 15:53:34 dtantsur: Thank you! 15:53:38 or "unanticipated issue in the ramdisk side" 15:53:58 Anyway, sounds good 15:54:17 * TheJulia adds a note to the whiteboard team status area so we circle back to it 15:54:30 if there is nothing else, i want to get some insight about the boot process on sw-raid 15:54:39 kaifeng: sure 15:55:06 from the original spec, it said only whole disk is supported 15:56:07 yes 15:56:08 and i see we also install bootloader for it, what makes me confused it appears like a chainloading, but can't figure out how it works for a partition table inside a partition 15:57:01 Oddly it is a thing that is actually supported by the bootloader if memory serves as well as the kernel 15:57:16 since the md device is seen as a whole block device on it's own 15:57:24 which can be subdivided 15:58:19 Maybe that provided the necessary insight? 15:59:05 TheJulia: thanks, this is helpful :) 15:59:34 so, I _think_ a partition image may just work with my uefi partition image changes, but it is not something I'm in a position to test... at all. 15:59:49 That being said, it is worth carefully reviewing our software raid logs as well. 15:59:56 and this is why I put only wdi into the spec 16:00:08 (since I could not easily test) 16:00:24 ahh, good to know 16:00:39 I think our CI jobs got it working, so we likely just need full standalone full uefi or something like that 16:00:42 * arne_wiebalck has no idea id partition images work :) 16:00:47 *if 16:00:50 which, afaik, is not a current scenario 16:01:12 Riccardo Pittau proposed openstack/python-ironic-inspector-client master: Remove lower-constraints job https://review.opendev.org/c/openstack/python-ironic-inspector-client/+/766985 16:01:23 I'll leave uefi as "complicated" 16:01:32 Do we have anything else to discuss this week? 16:02:20 i dont' feel there is a blocker for partition image on bios boot mode 16:02:32 kaifeng: probably not 16:02:46 kaifeng: I think we ?accidently? got it working in CI 16:03:14 I guess the docs may also say no support. 16:03:22 For the same reason :) 16:03:25 yeah 16:03:32 I think they do 16:03:44 yeap, i image for uefi it would be a problem 16:03:53 kaifeng: hopefully not much longer :) 16:04:12 Anyway, Thanks everyone! Have a wonderful rest of the year! See you in a few weeks! 16:04:17 o/ 16:04:21 \o 16:04:29 o/ thanks TheJulia 16:04:39 thanks, and happy new year o/ 16:04:41 thanks! o/ 16:04:48 Happy new year ++ 16:05:00 #endmeeting