16:00:05 <priteau> #startmeeting blazar
16:00:05 <opendevmeet> Meeting started Thu Jan 27 16:00:05 2022 UTC and is due to finish in 60 minutes.  The chair is priteau. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:05 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:05 <opendevmeet> The meeting name has been set to 'blazar'
16:00:10 <priteau> #topic Roll call
16:00:33 <mppowers> Hello priteau
16:00:50 <priteau> Hi Mark
16:01:56 <priteau> Waiting for Jason
16:02:10 <mppowers> Yes, he should be here soon
16:04:36 <diurnalist> o7
16:04:50 <priteau> Hi Jason
16:05:18 <priteau> Let's take a look at our fancy dashboard
16:05:22 <priteau> https://etherpad.opendev.org/p/blazar-whiteboard
16:05:29 <priteau> #topic Small patches ready to approve
16:05:36 <priteau> https://review.opendev.org/c/openstack/blazar/+/825255
16:05:36 <priteau> https://review.opendev.org/c/openstack/blazar-tempest-plugin/+/825195
16:05:36 <priteau> https://review.opendev.org/c/openstack/blazar-tempest-plugin/+/825415
16:05:58 <priteau> I've given +2 already
16:07:05 <diurnalist> i'm going through them now
16:07:21 <priteau> Actually I think the tempest one are duplicating a bit
16:08:10 <priteau> Not sure why this wasn't flagged by gerri
16:08:52 <priteau> I'll cancel the merge of 825195
16:09:17 <opendevreview> Pierre Riteau proposed openstack/blazar-tempest-plugin master: Add stable/xena jobs on master gate  https://review.opendev.org/c/openstack/blazar-tempest-plugin/+/825415
16:09:40 <opendevreview> Pierre Riteau proposed openstack/blazar-tempest-plugin master: Add stable/xena jobs on master gate  https://review.opendev.org/c/openstack/blazar-tempest-plugin/+/825415
16:10:02 <priteau> OK, abandoned https://review.opendev.org/c/openstack/blazar-tempest-plugin/+/825415
16:10:37 <priteau> I meant cancel the merge of 825415
16:10:58 <priteau> I think we're good now
16:11:22 <priteau> #topic Review priorities for Yoga
16:13:08 <priteau> Thanks for the review of the blazar-nova preemptible patch, still need to update it. I am working on the service side at the moment
16:13:27 <priteau> Not yet upstream but hopefully soon, tomorrow maybe?
16:13:57 <priteau> We have a lot of other patches ready to review, I have focused on two:
16:14:03 <priteau> 805462: Add a resource availability calendar for hosts | https://review.opendev.org/c/openstack/blazar-dashboard/+/805462
16:14:03 <priteau> 805461: Update list_allocations in hosts plugin to expose start and end dates | https://review.opendev.org/c/openstack/blazar/+/805461
16:14:58 <priteau> I updated both patches, but changes are minor, so Jason's should be able to give his +2 again
16:15:14 <priteau> For blazar-dashboard, I updated the vendored apexcharts JS
16:15:17 <diurnalist> sounds good, will look in to it
16:15:40 <priteau> mppowers: You didn't use a specific version on purpose? I assumed it was the latest at the time you started working on this
16:15:52 <mppowers> Yes, it was just latest at the time
16:16:12 <priteau> I tested the new one, took screenshots of the calendar, couldn't see any difference
16:16:49 <priteau> The overlay box still displays fine
16:16:57 <diurnalist> priteau: when i give my +2 on top of yours, it is ok to do the +1 to approved, yes?
16:17:18 <priteau> Of course, I actually expect you to do so ;-)
16:17:37 <diurnalist> yes, i thought so, just checking in case there was some other process i was stepping over
16:17:40 <priteau> Second +2 can request merge, unless there is a reason not to.
16:18:01 <priteau> For the list allocations patch, I fixed the merge conflict
16:18:08 <priteau> No functional change there
16:18:20 <mppowers> thank you for that
16:18:36 <priteau> Thank YOU
16:18:49 <priteau> As follow up to the dashboard I have these suggestions
16:19:07 <priteau> * docs
16:19:38 <priteau> At the very least we need to document the required policies (calendar doesn't work for non-admin users with default policy)
16:20:04 <priteau> And describe how to customise the host naming using a property
16:20:50 <priteau> That should go in https://docs.openstack.org/blazar-dashboard
16:21:03 <priteau> Does it sound reasonable?
16:21:27 <mppowers> Yes, that sounds good. Did you want this as a follow up patch?
16:21:53 <diurnalist> yes, i just approved the current one \o/
16:22:05 <mppowers> OK, great
16:22:39 <priteau> Yes, let's try to merge big patches when we reach a reasonable stopping point
16:22:56 <priteau> I am afraid trying to get them perfect means they never get merged
16:23:43 <priteau> As long as what we ship at the end of the release cycle is working and makes sense for operators and users, it's fine if it gets contributed in multiple steps
16:24:01 <mppowers> Makes sense
16:24:04 <priteau> My other comment was related to policies again
16:24:14 <diurnalist> i removed the two patches i just approved from the whiteboard
16:24:20 <priteau> Of course you need blazar:oshosts:get_allocations
16:24:27 <priteau> But I discovered you also need blazar:oshosts:get
16:24:35 <priteau> That's for the hypervisor name lookup
16:24:53 <priteau> Do you think we could do without as an option?
16:25:14 <priteau> Either configurable by admins
16:25:39 <priteau> Or dynamically, try a get-host, if it fails with 401, fall back to display host ID
16:25:55 <mppowers> Yes, I was thinking of using the fall back
16:26:25 <priteau> Cool. Do you think you could add this in the next couple of weeks?
16:26:44 <priteau> We have 4 weeks before feature freeze
16:27:35 <mppowers> We have a busy schedule, and my blazar-dashboard environment is not in a good state, so I am unsure if there will be time
16:28:26 <priteau> OK. Maybe for Z then.
16:28:42 <priteau> As long as we document how to make it work, that's fine.
16:28:45 <mppowers> OK, sounds good
16:28:57 <mppowers> I'll be sure to at least get the docs
16:29:01 <priteau> Thanks!
16:29:19 <priteau> Next on my review list is all the others in the main priority list
16:29:32 <priteau> I've not updated the Needs work section, it's possible some are ready to review now
16:30:08 <priteau> resource properties is
16:30:18 <mppowers> Yes, those should be ready
16:31:00 <priteau> RPC layer not passing CI. Maybe this is something we can focus on next release
16:31:39 <priteau> Same for third-party plugins, unless you think it's getting close?
16:32:03 <mppowers> The third party plugins is working, but it is dependent on the RPC refactor patch
16:32:12 <priteau> I see
16:32:27 <priteau> So that's where the CI failures are coming from?
16:32:37 <priteau> I've not looked at them to be honest
16:32:51 <mppowers> Yes. It'll take some time to figure out why the RPC test is failing tempest. I have spent some time debugging without luck
16:33:37 <priteau> 401 auth failures from Keystone
16:34:08 <priteau> Maybe something related to context
16:35:07 <priteau> Related to preemptibles, I submitted this:
16:35:07 <priteau> 825659: Support running periodic tasks from event loop | https://review.opendev.org/c/openstack/blazar/+/825659
16:35:46 <priteau> Fairly small, I use it to run some tasks not related to lease events
16:35:51 <diurnalist> priteau: where did you see the 401s in these logs? i'm not used to looking at tempest output
16:36:42 <opendevreview> Merged openstack/blazar-dashboard master: Add a resource availability calendar for hosts  https://review.opendev.org/c/openstack/blazar-dashboard/+/805462
16:36:52 <priteau> Hum, it was in the third-party plugin patch
16:37:01 <priteau> I cannot see them in the RPC one
16:37:23 <mppowers> Hm, maybe I need to debug both changes then
16:37:27 <mppowers> There is an issue I was seeing in the journal output from tempest, but I could not reproduce
16:37:31 <priteau> https://storage.bhs.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_eba/820255/9/check/blazar-tempest-plugin-base/eba279b/controller/logs/screen-blazar-m.txt
16:37:46 <priteau> https://storage.bhs.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_eba/820255/9/check/blazar-tempest-plugin-base/eba279b/controller/logs/screen-blazar-a.txt
16:37:52 <diurnalist> ah!
16:37:56 <priteau> These are the logs to watch
16:38:09 <diurnalist> it looks like it may actually be an assertionerror in the api?
16:38:20 <diurnalist> this helps, thanks
16:38:30 <priteau> Ah, in api logs
16:38:32 <priteau> assert NOTIFIER is not None
16:38:47 <priteau> Jan 25 18:19:30.372375 ubuntu-focal-ovh-bhs1-0028162968 devstack@blazar-a.service[101462]:   File "/opt/stack/blazar/blazar/notification/notifier.py", line 51, in get_notifier
16:38:47 <priteau> Jan 25 18:19:30.372375 ubuntu-focal-ovh-bhs1-0028162968 devstack@blazar-a.service[101462]:     assert NOTIFIER is not None
16:39:07 <mppowers> Ah that is the issue. I'm not sure why this isn't showing on my dev environment
16:40:35 <priteau> Maybe the config is different
16:41:47 <priteau> Do you use devstack to develop locally?
16:42:38 <diurnalist> can't speak for mark 100%, but we've moved away from devstack and instead work using kolla ansible and dev_mode, b/c it aligns much better w/ our other tooling
16:43:04 <mppowers> Yes, I was using devstack initially for the RPC patch, but then switched as Jason described
16:43:32 <diurnalist> so there is some divergence b/w CI and our local envs unfortunately. in practice it rarely comes up, this is the first time i recall. knowing how to debug the logs is a game changer, i was looking in the wrong places
16:44:11 <priteau> Oh I need to try this dev_mode
16:44:17 <priteau> Thanks for the tip
16:44:35 <priteau> I usually use devstack for development and then deploy with kolla
16:45:04 <diurnalist> not all services support dev_mode and for things like neutron plugins it gets a bit hairy but it's pretty simple to solve in kolla ansible. in general it works pretty well!
16:45:25 <diurnalist> in fact there is probably a patch we have for blazar dashboard and dev mode...
16:45:53 <mppowers> I was unable to get dashboard development working 100% with kolla too
16:46:42 <priteau> Anything else to share today?
16:47:27 <priteau> #topic AOB
16:47:29 <priteau> One thing
16:47:48 <priteau> The Call for Presentations for the next OpenStack Summit is open until February 9
16:48:13 <priteau> But not sure if you will be wanting to travel? It's in Berlin.
16:49:25 <diurnalist> wish i could!
16:50:15 <priteau> I will probably submit a talk on Blazar related to the developments we are doing (various bits not upstream yet, but hopefully all will eventually be)
16:50:33 <priteau> And something else
16:50:42 <priteau> Next PTG is scheduled for April 4 - 8
16:50:49 <priteau> Virtual event this one
16:50:58 <priteau> So please register if you haven't already: https://www.openstack.org/ptg/
16:51:33 <priteau> Will you be OK with a ~2 hour meeting like last time?
16:51:48 <mppowers> That should work for me
16:52:10 <diurnalist> i will actually be on holiday in australia then, i may not be able to attend
16:52:48 <priteau> In that case maybe we can hold a meeting before or after the actual event
16:53:27 <priteau> Let's discuss closer to the date
16:53:36 <diurnalist> could do -- timing is kind of bad, it's right in the middle of an extended trip for me. yes, let's discuss later!
16:53:37 <priteau> That's all I had for today!
16:53:54 <diurnalist> thanks priteau!
16:54:23 <priteau> Sorry it took a while to review these patches. I have more time for Blazar at the moment so I am trying to make the most of it!
16:54:51 <mppowers> Thanks again! Talk to you later
16:55:28 <priteau> #endmeeting