08:01:02 <licanwei> #startmeeting watcher 08:01:03 <openstack> Meeting started Wed May 8 08:01:02 2019 UTC and is due to finish in 60 minutes. The chair is licanwei. Information about MeetBot at http://wiki.debian.org/MeetBot. 08:01:04 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 08:01:06 <openstack> The meeting name has been set to 'watcher' 08:01:19 <Dantalion> hello 08:01:33 <licanwei> hi 08:01:40 <licanwei> Agenda for today: https://wiki.openstack.org/wiki/Watcher_Meeting_Agenda#05.2F08.2F2019 08:03:16 <licanwei> #topic Announcements 08:03:24 <chenker> hello 08:03:38 <chenker> sorry i come later. 08:04:47 <licanwei> I update Watcher blueprints and set Train cycle for bp I think we should do 08:05:19 <chenker> i plan to use base method add_action_migrate() to replace the migration method in specific strategies 08:05:45 <licanwei> Dantalion: https://blueprints.launchpad.net/watcher you can check again 08:06:16 <Dantalion> licanwei: Yes I saw the blueprints were updated I think this looks good 08:06:28 <licanwei> and update what you are interesting 08:07:22 <licanwei> #topic specs 08:07:48 <licanwei> https://review.opendev.org/#/c/656448/ Add Placement helper needs review 08:08:32 <licanwei> I'll reply matt's comments 08:09:14 <Dantalion> Yes I'll have time to review some open patches including this one this week as well 08:09:38 <licanwei> Dantalion: Thank you! 08:09:55 <chenker> Dantalion: Objectively speaking, the speed which you modify the code according others' advice is so quick 08:10:11 <licanwei> Dantalion: I just plan to add placement helper in this spec 08:10:31 <Dantalion> Yes with placement helper new strategies can start using it if desired 08:10:52 <licanwei> and there are more jobs need to do about strategy using placement 08:10:56 <Dantalion> chenker: thank you, I try to be quick because my term at CERN ends in 3 months 08:11:01 <licanwei> Dantalion: yes 08:12:03 <licanwei> Dantalion: then you will no time for Watcher? 08:12:24 <chenker> watcher need you, haha. 08:12:53 <Dantalion> licanwei: I'll start a new masters degree at the free university in Amsterdam so i'll have way less time 08:13:47 <Dantalion> But i'll still have some time but I will lose access to resources such as the CERN openstack cloud 08:14:13 <chenker> that's cool to have a new master degree 08:14:13 <licanwei> Dantalion: congrats! 08:14:50 <Dantalion> Thank you :) 08:15:00 <licanwei> OK, let's move 08:15:19 <Dantalion> yes 08:15:24 <licanwei> #topic review action 08:16:07 <licanwei> https://review.opendev.org/#/c/645294/ Allow for global datasource preference from config needs review 08:16:39 <licanwei> It's look good for me 08:16:56 <licanwei> and I'll review again 08:17:04 <Dantalion> thanks 08:17:56 <licanwei> chenker: I wish you also have time review 08:18:29 <licanwei> https://review.opendev.org/#/c/651746/ Using node replace resource_id in method add_action_disable_node() needs review 08:18:40 <chenker> yep, i will. i think it looks good to me. 08:19:34 <Dantalion> I think node replacement looks good, in general passing the entire Instance or ComputeNode objects to methods I think is better as it gives more flexibility 08:20:16 <licanwei> +1 08:20:45 <chenker> ye. 08:20:57 <licanwei> https://review.opendev.org/#/c/656457/ Add tempest voting needs review 08:22:13 <Dantalion> Maybe we can follow Matt's recommendation of combining the job before we merge. I do not know how difficult that is to do as I have little experience with zuul 08:23:44 <licanwei> Add gating will spend more time 08:24:58 <Dantalion> Yes tempest tests take long time. I think trying to improve tempest test time before gating is better 08:24:58 <chenker> i think Zuul check is enough 08:25:03 <licanwei> maybe we can reduce the install devstack before adding gate 08:25:23 <licanwei> Dantalion: yes 08:25:24 <Dantalion> licanwei: Yes I agree reduce then add gate 08:26:02 <licanwei> I think we should combine tempest tests 08:26:27 <licanwei> but it needs time to do 08:27:06 <licanwei> https://review.opendev.org/#/c/657218/ update wsme types needs review 08:27:52 <licanwei> there are some mix using wsme and wsme.types, 08:28:51 <Dantalion> I need more time to look at this I am unfamiliar with wsme / wtypes but in general using the same everywhere makes sense 08:28:55 <licanwei> this patch fix that 08:29:36 <licanwei> https://review.opendev.org/#/c/657211/ Add force field to Audit needs review 08:30:48 <licanwei> Dantalion, chenker: I wish you can review this patch 08:31:07 <Dantalion> This one I can do this week 08:31:22 <licanwei> Dantalion: thanks 08:31:40 <chenker> ok, i will do this week. 08:31:53 <licanwei> #topic Discussion 08:32:30 <licanwei> Documentation on devstack with different datasources 08:32:56 <licanwei> Dantalion: You mean Watcher no such doc? 08:33:28 <Dantalion> Watcher has no documentation to setup devstack with each possible datasource 08:33:38 <Dantalion> I think it would be good addition 08:35:20 <licanwei> https://docs.openstack.org/watcher/latest/ 08:36:06 <licanwei> This is Watcher docs 08:36:52 <licanwei> Dantalion: Which chapter need to add ? 08:37:10 <licanwei> or a new chapter? 08:37:47 <Dantalion> for example: devstack with gnocchi, devstack with ceilometer and devstack with monasca can be added to setup watcher on devstack 08:38:20 <Dantalion> we could specify a local.conf which we know for sure that works 08:38:59 <licanwei> Dantalion: OK, but need to test 08:40:08 <Dantalion> Yes requires thorough testing we need to make sure that it works for everyone who tries 08:41:02 <licanwei> Dantalion: I'll do some tests if I have time 08:41:45 <Dantalion> Great I'll also try to get started 08:42:09 <licanwei> another topic 08:42:14 <licanwei> Documentation on watcher tempest tests 08:42:55 <chenker> this part docs need to add 08:43:12 <Dantalion> I tried find tempest documentation but urls lead to source code not documentation. I do not know how to configure watcher tempest and debug. It makes developing tempest tests hard 08:44:10 <licanwei> Dantalion: there are many logs 08:44:45 <chenker> now, we test tempest just use the on-line openstack to help us. 08:44:56 <licanwei> such as :http://logs.openstack.org/57/656457/1/check/watcher-tempest-basic_optim/0a721bb/ 08:46:17 <licanwei> and there are devstack running log 08:46:35 <Dantalion> Yes for now online is sufficient but offline in local devstack would be better. Maybe I am just missing how to do it and its easy 08:46:35 <chenker> If we want to test in local env, it needs not only one compute node. 08:47:12 <licanwei> Dantalion: :) 08:47:14 <Dantalion> Ah teststack with two compute nodes is indeed pretty hard to setup 08:47:58 <chenker> some strategies need at least three compute nodes. 08:48:42 <licanwei> Dantalion: If you want you can write that doc :) 08:49:10 <Dantalion> I can try and see how far I get would be good excersize to learn understand tempest. 08:50:49 <licanwei> Dantalion: Great! 08:51:10 <chenker> wow 08:52:08 <licanwei> the last one topic 08:52:16 <licanwei> There are too many individual tempest tests, maybe we can merge them 08:52:37 <licanwei> We have discussed that 08:53:06 <licanwei> Now tempest tests spend more time, 08:53:41 <licanwei> It's good if we can combine them 08:54:24 <Dantalion> Yes afterwards we can look into adding gating when time is lower 08:54:39 <licanwei> Dantalion: I'll try to do it 08:56:28 <licanwei> Dantalion: If there are trouble, we discuss next time 08:56:35 <Dantalion> sound good 08:57:24 <licanwei> there are nothing for me :) 08:57:38 <chenker> me too 08:57:44 <licanwei> I'll end the meeting if no questions 08:58:01 <Dantalion> I briefly want to say we might see new developers soon 08:58:11 <licanwei> Dantalion: yes 08:58:21 <licanwei> Dantalion: I wish~ 08:58:38 <licanwei> Dantalion: There are too many thing need to do 08:58:46 <chenker> watcher is a good project , IMO 08:58:56 <Dantalion> It be good to have extra support we will have to wait for now lets hope 08:59:38 <licanwei> Dantalion: See you next meeting :) 08:59:55 <Dantalion> See you all next meeting :) 09:00:00 <chenker> Bye~ 09:00:02 <licanwei> #endmeeting