15:00:00 #startmeeting ironic 15:00:00 Meeting started Mon Apr 11 15:00:00 2022 UTC and is due to finish in 60 minutes. The chair is iurygregory. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:00 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:00 The meeting name has been set to 'ironic' 15:00:01 sure :) Thanks :iurygregory 15:00:18 Hello ironicers o/ welcome to our weekly meeting! 15:00:32 the agenda for our meeting can be found in the wiki 15:00:36 o/ 15:00:38 #link https://wiki.openstack.org/wiki/Meetings/Ironic#Agenda_for_next_meeting 15:00:44 o/ 15:00:45 o/ 15:00:48 o/ 15:00:49 o/ 15:01:00 \o 15:01:13 o/ 15:01:28 ok, let's start o/ 15:01:34 #topic Announcements / Reminder 15:01:58 Reminder from the PTG =) 15:02:49 #info Check if you have any action items in the ptg etherpad =) 15:03:08 so we don't miss anything important =) 15:04:01 #info Iury will be sending a summary of the PTG the list til wednesday 15:04:49 does anyone have anything else to add for announcements/reminder? 15:06:17 ok, moving on 15:07:16 skipping the next two topics (since we will have subteam after we have the priorities for Zed) 15:07:35 #topic Deciding on priorities for the coming week 15:07:42 #link https://review.opendev.org/q/status:open+hashtag:ironic-week-prio 15:07:56 Does anyone have any patches that needs review? 15:08:08 https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/826646 15:08:34 can we please review above patch 15:09:10 I don't think I have anything new.. all downstream nowadays. but may produce a couple of patches later this week. 15:09:39 ameya49, ack I will check this one this week 15:09:41 dtantsur, ok 15:09:55 iurygregory : Thanks 15:10:32 I will add a patch later today for the Zed Priorities (FYI) 15:10:40 \o/ 15:12:04 ok, moving on since we have topics for discussion today =) 15:12:11 #topic Open discussion 15:12:32 ameya49, feel free to give context and share some links related to your issue 15:12:46 iurygregory : Yes, Thanks 15:14:04 While running test_introspection_basic, once it verifies the introspected data and tries to boot_instance, it raises an error with nova quota values exceeded 15:14:39 where if we see values of introspected data which are greater than default values of nova quota set 15:15:24 thw workaround I tried by setting default nova quota values equal to introspected values for which it went ahead 15:15:56 So, how can we go further with same approach, updating default quota values while test case is running? 15:16:47 which resources are you hitting the quota with? 15:17:08 cores and ram 15:17:32 hmm.. ironic instances should not consume these any more. are you configuring the flavor correctly? 15:17:58 Ebbex proposed openstack/bifrost master: DNM: Is bindep the thing causing problems? https://review.opendev.org/c/openstack/bifrost/+/837348 15:18:32 humm good point 15:19:00 using default flavor (baremetal) which is getting created while devstack setup 15:19:07 if is using the default flavor it will complain 15:19:41 you will need to override the configs from the flavor (we have some variables that you can set in local.conf I think) 15:19:56 so do you suggest to create a new flavor while running? 15:20:06 I think it's worth asking nova folks 15:20:20 why with resources:MEMORY_MB=0 it still consumes the quota 15:20:27 arne_wiebalck: do you have any experience ^^? 15:21:00 Arne is a bit busy today (he might not answer now) 15:21:10 ah 15:21:20 I have little recent experience with nova, but it sounds like a bug on their side to me 15:22:53 dtantsur : Do you suggest to raise a bug for same? 15:23:06 ameya49: maybe ask on #openstack-nova first? 15:23:12 it's possible that we misunderstand something 15:23:18 https://github.com/openstack/ironic/blob/7ac480412626c38fa3493088dbf49e29303491b6/devstack/lib/ironic#L2606-L2612 15:23:22 show them your flavor and your quotas 15:23:22 because of this right? 15:23:30 dtantsur : sure. 15:23:52 ameya49: yeah. in my view, a flavor configured this way should only consume the (non-existing yet?) custom resource quota 15:24:03 but I remember arne_wiebalck complained a lot about nova quota applied to ironic :) 15:24:26 iurygregory : Yes 15:24:50 * dtantsur is wondering if dansmith is by chance around 15:25:06 in a meeting 15:25:16 k no worries 15:25:46 dtantsur: Do you expect changing those to be non-zero would work around the issue? 15:26:20 rpioso: I would expect changing those to non-zero to break everything :) 15:26:42 it's possible that I don't understand how quota work. but I would expect the stuff set to zero to not count against the quota. 15:26:58 but please do talk to someone from nova who understands these matters 15:27:30 if IRONIC_VM_SPECS is 4GB the flavor only has 4GB in the flavor... I would expect nova to complain since the node will have more than 4GB... 15:27:48 same would go for vcpu IRONIC_VM_SPECS_CPU 15:28:15 when it comes to scheduling, this stuff is overridden by resources 15:28:21 dunno about quota 15:28:45 yup 15:28:46 Please remind me ... Isn't introspection part of our gate? 15:28:52 it is 15:28:55 yes 15:29:21 not totally following here, but nova's quota comes from the actual things in the flavor, and not the resource overrides 15:29:25 to check default nova quota I was running "nova quota-show" which displays table of resources with its default values 15:29:40 dansmith: I guess this answers our question. is there a reason why? 15:29:53 for cores default value is 20 15:30:04 because the quota stuff comes from looong before the resource stuff 15:30:25 so, "nobody has changed that"? 15:30:49 not really, the flavor things are for presentation to the user and the resource stuff is more like instruction to the backends (i.e. placement) 15:31:09 dansmith: mmm, I see. this does not play too well with bare metal. 15:31:13 it would be confusing to a user to see a flavor require X memory and Y be charged against their quota 15:31:18 dtantsur: does anything in nova? :) 15:31:25 haha :D 15:31:29 but yeah, let's talk about options later when I can focus 15:31:47 okie, thanks dansmith! 15:31:56 ty 15:32:03 rloo and zer0c00l likely have input to upstream quota discussions, so I'm pinging them 15:32:04 ameya49: so, yeah, the values in the flavor do count against the quota. 15:32:39 I know they put in quite a bit of work to get quotas working, but I don't think the end result looks/acts very upstream 15:32:41 ameya49: the immediate fix for you is to either increase the quota or decrease the values in the flavor 15:32:53 so one way to workaround their issue is to set IRONIC_VM_SPECS IRONIC_VM_SPECS_CPU in their local.conf with other values 15:33:30 datansur, iurygregory : Thanks. Will check those. 15:34:43 ok, I think this is everything we have for Open Discussion, shall we move on? 15:35:19 #topic Baremetal SIG 15:35:49 #info SIG meeting tomorrow with Manuel Holtgrewe on "Bare Metal for Health - Using OpenStack Ironic for HPC at Berlin Institute of Health" 14:00 UTC (2pm UTC) 15:36:02 * dtantsur would really love to hear that 15:36:11 #link https://etherpad.opendev.org/p/bare-metal-sig 15:36:28 you can find the details to join the meeting tomorrow in the etherpad above 15:36:45 #info arne_wiebalck requested a Forum slot for the Baremetal SIG 15:36:47 \o/ 15:37:48 #topic RFE review 15:38:10 Does anyone have any RFE that we should discuss? 15:39:32 ok, moving on 15:39:36 #topic Who is going to run the next meeting? 15:39:59 do we have any volunteers? 15:40:26 I'll be out - public holiday 15:41:16 * iurygregory is wondering if is holiday in more places... 15:41:28 also on PTO 15:42:26 iurygregory: It is not a US national holiday. 15:42:57 ok, I will run the meeting o/ 15:43:07 in case we don't have quorum we can skip =) 15:43:12 Thanks everyone! 15:43:43 #endmeeting