15:00:08 #startmeeting ironic 15:00:08 Meeting started Mon Nov 29 15:00:08 2021 UTC and is due to finish in 60 minutes. The chair is dtantsur. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:08 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:08 The meeting name has been set to 'ironic' 15:00:12 o/ 15:00:16 hi folks 15:00:30 o/ 15:00:39 \o 15:00:44 o/ 15:00:57 o/ 15:01:17 #link https://wiki.openstack.org/wiki/Meetings/Ironic our agenda for today 15:01:24 * TheJulia wonders how light our december meetings may be 15:01:30 very light, I suspect :) 15:01:35 #topic Announcements / Reminder 15:02:10 #info Tested Runtimes for Yoga - Python 3.8 and 3.9 (https://governance.openstack.org/tc/reference/runtimes/yoga.html) 15:02:27 This did cause a heated discussion on the ML since 3.6 is what CentOS/RHEL 8 has 15:02:36 I have a discussion topic to return to this issue 15:02:48 #info First bugfix branch should be created by the end of next week 15:02:56 Who doesn't like a fresh warm release? :) 15:03:08 let's make sure to finish whatever can be finished by EOW (this week) 15:03:25 o/ 15:04:05 #info PTL on PTO during whole December 15:04:13 let's wish iurygregory great time :) 15:04:38 ++ I'm glad he made it home 15:04:44 yep! 15:04:47 That's all we had, does anyone have any other announcements? 15:05:44 I've got nothing annoucement wise 15:06:02 we *should* try and see when we want to stop holding meetings for the rest of the year, during open discussion 15:06:14 yep 15:06:22 #topic Review action items from previous meeting 15:06:27 iurygregory to create a doodle for the new time so people can vote + email to openstack-discuss 15:06:32 I don't think this has actually happened? 15:06:38 I did not see it 15:06:54 He seemed excepptionally busy last week 15:06:58 I can imagine 15:07:20 was it about moving the meeting time? maybe rpittau could drive it since you're the most affected? 15:07:39 Yeah 15:09:36 * dtantsur pokes rpittau 15:09:48 sounds good, I will take care of it (sorry, double meeting) 15:09:48 ah, well, conflicting meetings. we'll get back to it. 15:10:02 no worries, I obviously forgot *why* you were affected :D 15:10:10 #topic Review subteam status reports 15:10:15 hmm, first 15:10:27 #action rpittau to create a doodle for the new meeting time so people can vote + email to openstack-discuss 15:10:32 now, whiteboard 15:10:36 #link https://etherpad.opendev.org/p/IronicWhiteBoard white board 15:10:47 I don't think we've merged the Yoga topics 15:11:01 We have not 15:11:03 #link https://review.opendev.org/c/openstack/ironic-specs/+/815308/ Yoga themes 15:11:15 still, I'd suggest we start reshaping the whiteboard according to it 15:11:21 any volunteers? 15:11:29 (I can do it, but I'm happy to delegate :) 15:12:06 I might be able to this week. I've got a few past-due things to take care of today 15:12:21 I can do it as well, just checking if anyone is really eager 15:12:26 ++ 15:12:44 #action dtantsur to update the whiteboard with the (draft) yoga themes 15:13:03 #topic Deciding on priorities for the coming week 15:13:10 first and foremost, please review the yoga themes 15:13:31 I'd love to see https://review.opendev.org/c/openstack/sushy/+/817792 merged, but it's on the list already 15:13:35 anything else? 15:13:58 If I can crank out the client patch for node history events tomorrow, any objection for me to add it tomorrow? 15:14:05 none from me 15:15:27 okay, let's move on to discussions, we have a couple of things 15:15:34 #topic Situation around Python 3.6 support 15:15:41 I Added a couple of patche sfor tempest 15:15:45 too late :( 15:15:46 thanks rpittau 15:15:59 rpittau: ack, I'll try to review them soon() 15:16:04 ack 15:16:10 So, as I mentioned earlier, the tested runtimes no longer include python 3.6 15:16:28 which causes issues for Bifrost which supports Stream 8, which has 3.6 15:16:47 (it also supports Bionic, openSUSE Leap and Debian 10, all of which have <3.8 and will need to be deprecated) 15:17:07 but Ubuntu and Debian have new versions which support 3.8+ 15:17:25 CentOS has Stream 9, but I don't know how many folks will feel easy to use it for their production deployments 15:17:36 any opinions or suggestions? 15:18:15 I have this bad feeling it doesn't end up with anyone happy 15:18:39 about centos stream 9, I definitely do not suggest it in the current status 15:18:39 for the rest, if we want to keep testing on py3.6 we can, we "just" to add a py3.6 test 15:18:57 rpittau: do you have any practical experience that makes you say so? re 9? 15:19:15 we have a discussion on the ML where some people are basically going "9 is just as good as 8, what's the problem?" 15:20:56 I feel like those saying it didn't get to feel pain during centos stream 8 15:21:56 I'd like us to keep 3.6 support for at least Yoga; similar voices are coming from Kolla and RDO teams 15:22:07 but it's seems like the matter is already decided on the TC side 15:22:09 I'm on board with that for now 15:22:28 ++ keep 3.6 support 15:22:35 my experience so far is restricted to bifrost and metal3 15:22:35 if it was an overlap phase where we had to temporarily support both centos8 and centos9 I would say no problem, but abruptly going from one to another is a bit tough 15:23:02 at the very least we can add add 3.6 unit tests on Ironic and other our projects to make sure *we* don't break and serve as canary for broken dependencies 15:23:09 This does feel like the community is trying to make a very hard cut 15:23:26 but I think upfront we knew we would only be on 3.6 for 3-4 years tops. 15:23:57 but it is hard for us to really comprehend or remember that in active memory from day to day when distros do long lived packaging 15:24:01 I'm ok to add unit tests for py3.6 for yoga, and work on centos9 at the same time 15:24:05 yep 15:24:07 and yep 15:24:33 voting or non-voting? 15:24:49 voting... at least for the time being 15:24:53 voting 15:25:36 okay, sounds good, will do (help welcome) 15:25:51 moving on? or any other opinions? 15:26:07 I suspect we're safe to move on 15:26:22 I will help 15:26:22 one concern I have is for dependencies removing py3.6 support (oslo libraries for example), need to keep an eye on those 15:26:30 yeah 15:26:40 yeah, that's why I'm saying that the job will work as a canary 15:26:52 if they start slipping in a minimum requirement of 3.8, that is pretty much shooting eveyrone in the foot 15:26:52 at least we'll know when it's over :) 15:26:56 :) 15:26:57 ++ 15:27:03 #topic Baremetal SIG 15:27:07 arne_wiebalck: anything from you? 15:27:40 * TheJulia wonders if the sig is on holiday until January 15:28:13 well, I have a talk on the 14th :) 15:28:18 * dtantsur is curious if anyone shows up 15:30:18 okay, I assume nothing new from the SIG at this point 15:30:30 #topic RFE review 15:30:39 dtantsur: what is the talk on? I've been under a rock 15:30:40 #link https://storyboard.openstack.org/#!/story/2008723 Enhance Redfish firmware update user experience 15:30:53 TheJulia: I'm going to talk about Metal3, maybe even show it 15:31:20 meanwhile, ajya, you around? 15:31:21 about RFE - it is updated after previous review - zip support removed and added download_source similar to images. Any feedback? 15:32:08 Was zip support added by the ilo folks? 15:33:13 I think we tried to figure that out and decided that rather no? 15:33:15 Unfortunately, removing support would cause me to be worried about compatibility and messaging. 15:33:29 I think ajya means it was removed from the RFE, not from the code 15:33:32 as far as I recall, no it was not aded by ilo, they had different packages handled server side 15:33:34 oh! 15:33:35 okay 15:33:41 yes, removed from RFE 15:33:46 for redfish 15:33:48 nevermind me! 15:34:07 the RFE seems good to me 15:34:51 thanks 15:34:54 anyone else? 15:35:29 * TheJulia looks 15:36:59 lgtm 15:37:13 any objections? 15:37:23 none at this time 15:38:03 okay, marked approved 15:38:10 #topic Open discussion 15:38:17 the floor is open 15:38:52 So, Typically we've taken a moment in these meetings in past years to identify when people were generally going to go *poof* for the year 15:39:24 I'll be more or less present until ~ 22nd 15:39:42 I personally can disappear after the 15th, but I can hang out until say the 22nd or just keep an eye on things 15:39:55 I will be present until the 24th 15:40:23 so the next question, is how many more times do we want to meet this year? 15:40:28 Typically we stop mid-month 15:40:54 definitely not 27th :) unlikely Jan, 3rd 15:41:06 we could skip the 21st ? 15:41:09 I'll probably be here on the 20th still 15:41:20 sorry, the 20th, yeah 15:41:35 o/ 15:41:44 * arne_wiebalck was in another meeting 15:41:57 arne_wiebalck: anything from the SIG standpoint? 15:42:13 arne_wiebalck: I wonder how many people will still be here on the 14th for the meeting.. 15:42:13 nothing for the SIG, except dtantsur's presentation on the 14th! 15:42:18 hmm 15:42:31 I don't really have any strong opinions either way, If we want to skip the 27th and the 3rd, I think that works but I suspect the 20th will lack critical mass. Just speaking from experience 15:42:41 we can postpone to Jan if you prefer, but mid december should be ok, I guess 15:42:46 as you prefer 15:43:13 I'll be here anyway 15:43:17 me, too 15:43:25 maybe a pre-holiday presentation is actually nice :) 15:43:33 :) 15:43:37 TheJulia: I'm cool with skipping 20th 15:43:41 need some x-mas filter! 15:44:00 yeah 15:44:13 rudolph circling the screen or sth 15:44:28 then lets skip the 20th, 27th 15:44:33 what about the third? 15:45:02 I'll probably be still out 15:45:04 still holiday here 15:45:25 Then lets resume January 10th 15:45:30 ++ 15:45:34 yep 15:45:37 and hope all Ironic contributors are ready for plenty of code reviews :) 15:45:40 who wants to send the announcement? 15:45:42 yeah 15:45:43 I'll send the email 15:45:46 thanks! 15:45:59 I guess I have a couple other minor items 15:46:45 It is board nominations right now. You should have an email from the Open Infrastructure Foundation if you'd like to nominate anyone for board of directors. As a reminder, individuals need 10 nominees to appear on the ballot. Please use firefox since it is known to work. 15:47:32 ++ 15:47:36 anything else? 15:47:44 Additionally, the nova patche to try and fix the lack of actual rebalance needs eyes 15:47:45 https://review.opendev.org/c/openstack/nova/+/813897 15:47:49 i've been workin on the implementation of the feature described here and would super appreciate some eyes on my work: https://review.opendev.org/c/openstack/ironic-specs/+/808804/ 15:47:57 it changes the behavior, I'd <3 for operators to review it 15:48:15 * TheJulia looks at rloo, zer0c00l, mnaser, arne_wiebalck ;) 15:48:42 sam_z: ah, redfish proxy. yeah, it fell off my radar, sorry 15:48:44 TheJulia: ++ 15:49:04 I wonder how it plays together with the allinone (sorry TheJulia) ironic process.. will need to think 15:49:23 dtantsur: sam_z ? 15:49:26 or the nova stuffs? 15:49:33 the redfish stuffs 15:49:40 k, good question 15:49:53 also i remember TheJulia saying she had interest in migrating the ironic api to flask and since i'm workin' with flask here i just wanted to put my name in for that lol 15:49:54 it's probably not a concern, unless sam_z wants to have a new executable 15:50:04 yeah 15:50:09 allinone process? 15:50:18 I think the consensus kind of lead to a new process for the proxy anyhow, at least for now 15:50:23 sam_z: easier to show than to explain: https://review.opendev.org/c/openstack/ironic/+/819620 15:50:34 basically API and conductor in a single executable 15:50:37 sam_z: the idea has been kicking around for the last two years 15:51:25 I'd definitely want to see a possibility to use an all-in-one process 15:51:39 even if the default is to use a new ironic-redfish-api executable 15:51:43 one last item from me. It seems I'm doomed to do a lot more in tempest this cycle than I expected. I've been informed more comprehensive testing is expected, basically to ultimately remodel the plugin to be rbac aware at the level of tests above and beyond what client to use 15:51:48 would this feature working with allinone ironic imply it being integrated into the allinone executable? 15:52:08 cuz i *think* that actually might not be too bad actually 15:52:11 sam_z: likely so. it's probably going to be trivial, just making sure you're aware that it's happening 15:52:12 I've got a basic wip up https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/819165, since the desire is a mix of positive and negative testing, I'd appreciate any thoughts/feedback since it is going to be a large pile of work 15:52:22 TheJulia: my condolences (re tempest) 15:52:27 dtantsur: thanks 15:52:57 not what I expected to spend december on 15:53:23 I guess tempest is not the future we wanted, but the future we deserved 15:53:34 maybe 15:54:12 Our plugin has somewhat languished as well since it is often an after thought 15:54:15 i have already been brainstorming ways to launch the ironic api and this redfish proxy in the same service anyway and i think because my thingy is a separate wsgi app, in theory we could just use something like the werkzeug dispatcher to serve both 15:54:29 or a minor thing creeps into being a more major thing 15:55:10 sam_z: might work, although I think rpittau had an issue with werkzeug directly 15:55:16 or I may be remembering something else 15:55:23 sam_z: there are numerous possibilities, depending on whether you want to reuse the same port or not 15:55:47 * TheJulia wonders if we should just let sam_z do what he wants and see what the outcome is 15:55:49 mmmm not me or it was not recently 15:55:57 rpittau: this felt recent to me 15:56:02 I have some experience with werkzeug, but it mostly positive 15:56:15 I think we use it in IPA directly? 15:56:19 yea, i've been meaning to look into trying to implement it and just seeing what happens lol 15:56:29 hjensas: was there something w/r/t werkzeug recently? 15:56:43 dtantsur: we do 15:57:17 * TheJulia wonders if we're in the multiverse of python madness 15:57:26 (w/r/t my thinking of recent werkzeug issues 15:57:28 ) 15:57:56 werkzeug is the thing behind flask, it's not the worst piece of the puzzle 15:58:03 true 16:00:18 Anyway, I guess we're done for today? 16:00:44 yep 16:00:47 thanks folks! 16:00:48 Dmitry Tantsur proposed openstack/bifrost master: Stop exposing JSON RPC to the whole network https://review.opendev.org/c/openstack/bifrost/+/819649 16:00:53 #endmeeting