15:00:39 #startmeeting ironic 15:00:39 o/ 15:00:39 Meeting started Mon Feb 17 15:00:39 2020 UTC and is due to finish in 60 minutes. The chair is TheJulia. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:40 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:42 The meeting name has been set to 'ironic' 15:00:44 Good morning everyone! 15:00:47 o/ 15:00:48 o/ 15:00:49 o/ 15:00:50 o/ 15:01:07 \o 15:01:10 My apologies in advance, I'm on a call at the same time this morning 15:01:21 o/ 15:01:27 o/ 15:01:30 Our agenda can be found on the wiki! 15:01:32 #link https://wiki.openstack.org/wiki/Meetings/Ironic#Agenda_for_next_meeting 15:01:35 o/ 15:01:38 TheJulia: I can drive this if you #chair me 15:01:50 o/ 15:01:56 dtantsur: I think I'll be listen only,but I'll chair you anyway in case I go idle 15:02:01 #chair dtantsur 15:02:01 Current chairs: TheJulia dtantsur 15:02:06 o/ 15:02:09 (spoon'o'meter in orange zone, but I'm coping) 15:02:30 #topic Announcements / Reminders 15:02:57 #info Next week is the ironic mid-cycle at CERN. Thanks goes to arne_wiebalck and everyone else at CERN for hosting us next week. 15:03:09 #info The meeting next week is cancelled as a result of the mid-cycle. 15:04:18 #info mgoddard notes that stable releases have occured, and only ironic, ironic-inspector, and IPA have yet to be released on master. 15:04:42 and sushy-cli I suspect 15:04:51 which we promised to release around M2 for it to be included in Ussuri 15:05:20 #info One other item, regarding the PTG, I registered our team's interest with approximately 15 possible participants. 15:05:49 dtantsur: good point 15:05:59 I can request it since mgoddard is out 15:06:09 Does anyone else have anything to announce? 15:06:12 dtantsur: That would be awesome! 15:06:31 * dtantsur still haz powerrz 15:06:32 we still have a couple of awesome sushy-cli patches to merge 15:06:43 etingof: is there an explicit list? 15:06:44 etingof: yeah, but there is no need to hurry them in the release 15:06:52 there'll be more releases in ussuri 15:07:01 (we can release it every week, if you want to) 15:07:01 ++ 15:07:12 The fact there is a release is the important part 15:07:14 how often does oslo release, every week, every couple of weeks? 15:07:21 I also think wwe should release it now 15:07:42 dtantsur: every couple of weeks or as-needed 15:07:43 I can compose a list, however if one just search by repo name - all open patches can be merged 15:08:31 Merged openstack/ironic-python-agent master: "fix" GenericHardwareManger unit test failure https://review.opendev.org/707730 15:08:32 Merged openstack/ironic master: Refactoring: finish splitting do_node_deploy https://review.opendev.org/707394 15:08:37 etingof: if there is nothing blocking an initial release, we should release 15:08:53 but a prioritized list is always important to help reviewers 15:09:02 it's 0.1, we can still rewrite it a few times :) 15:09:11 * dtantsur has rewritten ironic-inspector after 0.1, I think :D 15:09:12 nothing blocking, but it would be good to merge pending patches as they bring more value to the tool 15:09:34 Anyway, I believe we had no action items last week, so we can skip forward to subteam status reports 15:09:52 dtantsur: ship it! 15:09:54 :) 15:10:05 Well, seems there are no other annoucements, so I think we can proceed 15:10:33 #topic Review subteam status reports 15:10:37 #link https://etherpad.openstack.org/p/IronicWhiteBoard 15:10:57 Starting at line 276 15:12:19 I do't know if anyone has started to look at mgoddard's POC patches for deploy step decompoosition and agent deploy steps 15:12:26 It might be good to get some early reviews of that 15:12:49 * dtantsur doesn't remember if he did 15:12:52 tzumainn: just pending reviews right now? 15:13:11 TheJulia, yep - I need to work on that neutron port thing, and then some documentation after that 15:13:27 tzumainn: FYI I'm getting some good progress with the deployment API work 15:13:37 mostly blocked by trying not to cause conflicts for TheJulia 15:13:44 dtantsur, awesome - hopefully I'll be able to help next week 15:14:05 neat! I may have finished the deployment part by then, you can help with undeployment and with policies 15:14:15 sounds good! 15:14:41 dtantsur: <3 15:15:02 dtantsur: on a plus side, looks like your out of territory of conflicts with my agent token changes 15:15:17 TheJulia: that's because I haven't started refactoring tear down ;) 15:15:25 heh 15:15:28 lol 15:16:35 I'd like to encourage everyone to trim the history off their items that may be out of date :) 15:18:38 I think we're good to move on to weekly priorities 15:18:48 ++ 15:19:02 #topic Deciding on priorities for the coming week 15:19:11 #link https://etherpad.openstack.org/p/IronicWhiteBoard 15:19:14 Line 164 15:19:27 I didn't have a chance to go through each item to strike out merged items :\ 15:19:58 So someone wants to start at the bottom, I can start from the top 15:20:49 * dtantsur does 15:21:49 iurygregory: https://review.opendev.org/#/c/706346/ needs rebasing 15:21:49 patch 706346 - ironic-prometheus-exporter - Fix None for instance_uuid label - 1 patch set 15:22:02 dtantsur, yeah, i'm working on it o/ 15:22:44 Thank you whoever deleted the striked out lines :) 15:23:02 oh I was updating it and didn't see dtantsur also was :D 15:23:36 the faster - the better 15:23:38 Is there anything that anyone is aware of that needs to be added to the list? 15:24:00 We did some fixes last week to stabilize ironic CI, but I think the needful merged 15:24:09 I'd like a couple of eyes on https://review.opendev.org/672179 but it's not super urgent 15:24:09 patch 672179 - ironic-inspector - Use openstacksdk for ironic module - 53 patch sets 15:24:45 xXraphXx: I can have a look at rebasing the UEFI RAID patches if you have too many other things on your plate atm, let me know 15:25:07 https://review.opendev.org/706632 should reduce failures further with the iscsi raid test 15:25:08 patch 706632 - ironic-python-agent - Rescan after restarting the md device - 4 patch sets 15:25:12 but it is already on the list 15:26:39 arne_wiebalck: I suspect that may be good to do 15:26:45 rpittau: added 15:27:07 thanks 15:27:10 TheJulia: ok 15:27:49 ohh, the ilo item merged 15:28:36 I think the list is good to go if there is nothing else to add ? 15:28:59 there are also a bunch of patches for IPA-builder to fix different jobs, thwy should be quick to review 15:29:07 * dtantsur slaps himself on hand to stop reordering things 15:29:32 dtantsur: heh 15:29:50 rpittau: Those should be quick to review after the meetings 15:30:01 concurrent meetings should be banned 15:30:13 ok :) 15:30:17 I'm good with the list as-is 15:30:23 We have no discussion topics 15:30:32 arne_wiebalck: anything Baremetal SIG related? 15:30:51 no 15:30:59 after the mid cycle 15:31:04 Okay then, if there are no objections, we can proceed to RFE review 15:31:27 Everyone good to proceed 15:31:35 let's 15:31:37 yes 15:31:41 #topic RFE Review 15:31:51 * TheJulia backs a truck into the irc channel with an unpainted bike shed 15:32:02 \o/ 15:32:10 dtantsur has brought us https://storyboard.openstack.org/#!/story/2006910 15:32:25 not the first time :) this is the one-stop deployment API 15:32:36 I primarily need opinions on the API design, field names, etc 15:33:11 almost like if I wanted to use the ansible module, at least that is the analogy in my mind 15:33:48 sort of, yeah 15:34:02 it's also another chunk of functionality that is currently implemented in metalsmith 15:34:58 This seems logical... will it be done in flask? *ducks* 15:35:15 * kaifeng_ is still opening the story... 15:35:19 TheJulia: I'm actually considering at least not using wsme 15:35:31 I don't think we can bolt a flask blueprint on top of a pecan app 15:35:33 that is a start 15:35:34 * dtantsur can try 15:35:46 I don't even know if we will use blueprints :) 15:36:04 what makes sense to you in making something that works with your time, don't let me drive you to do tons of extra work 15:36:18 heh, yeah, so it's probably pecan+jsonschema, then we see 15:37:11 works for me, tbh 15:37:18 * dtantsur hopes everyone's silence implies unconditional agreement :D 15:37:22 but I also suspect some thought is required 15:37:32 this may also be the case! 15:38:04 etingof: "no release job specified for openstack/sushy-cli, one of ['publish-to-pypi'] needs to be included" 15:38:16 we need to fix project-config to be able to release, could you check it? 15:38:21 sure 15:38:23 I was wondering if we could wait for the big flask migration, but I guess we can just modify the implementation afterwards once we have a clear workflow 15:38:23 ++ 15:38:39 rpittau: I'm afraid the big flask migration can span a few cycles 15:38:46 dtantsur: yeah 15:38:51 agreed :( 15:39:07 I hope we can un-wsme-fy ironic this cycle though 15:39:33 okay, if anybody gets opinions about this RFE, lemme know at any moment 15:40:44 I've also found this one while triaging: 15:40:47 #link https://storyboard.openstack.org/#!/story/2006421 validate image partition at early stage 15:43:04 I find it controversial to make assumptions about user images.. but probably "does it have a partition table" is safe enough? 15:43:06 dtantsur: interesting 15:43:11 there are a number of facets there 15:43:59 that is a fairly safe thing to check 15:44:20 I'd make it optional maybe, and we might have to do it in two spearate plaes, so it would be an ironic-lib thing maybe 15:44:34 is that a check that should be done/checked in ironic? 15:44:34 maybe we can have a "force" flag to override this check? 15:44:48 arne_wiebalck: ideally so, because failure to do it is quite hard to debug 15:44:54 Well, the problem also is that the only time we can check this is way down the deploy path 15:45:17 we can use guestfish to dump partitions 15:45:18 dtantsur: what I meant is that this would be the same for a VM, no? 15:45:39 arne_wiebalck: probably? but it's us how make the decision based on the presence of kernel/ramdisk (for standalone case) 15:47:03 is there a way to check partition table on qcow instead of decompressing it to raw? 15:47:20 loopback attachment 15:47:30 instead of guestfish we can also use lsblk on loopback 15:47:31 but we also have streaming deployment cases 15:47:31 questfish works as well, I think 15:47:35 We need to map things out 15:47:38 yeah 15:47:38 guestfish willalso mask stuff 15:47:55 the biggest problem with guestfish is that it's resource-greedy 15:47:56 we first need to map everything out case wise and then consider where makes sense because there are overlaps 15:48:09 guestfish takes its time 15:48:25 I guess I agree with needs-spec on this RFE 15:48:42 ++ 15:49:07 okay, I have no more RFEs today 15:50:31 ++ 15:52:51 Moving to Open Discussion 15:52:54 #topic Open Discussion 15:52:58 with 7 minutes remaining! 15:53:08 Does anyone have anything to discuss? 15:53:40 BTW, the week after the mid-cycle, would anyone be available to run the meeting in the event that I'm unable to? 15:54:15 likely me 15:54:40 Okay, I volunteered to give a talk that week and I've not heard back fro the organizers yet 15:54:49 so *shrug* 15:55:15 heh 15:55:57 s/fro/from/ 15:56:15 dtantsur: re combining flask and pecan, you can probably do it with something like this: https://flask.palletsprojects.com/en/1.1.x/patterns/appdispatch/#combining-applications 15:57:36 * TheJulia opens that in tab to take a look after all of the morning meetingsd 15:57:44 Thanks everyone! 15:57:56 tl;dr werkzeug has a thing to mount multiple wsgi apps and dispatch them :) 15:58:18 thanks, TheJulia 15:58:59 thanks! 15:59:04 #endmeeting