08:01:41 <ifat_afek> #startmeeting vitrage
08:01:42 <openstack> Meeting started Wed Oct 24 08:01:41 2018 UTC and is due to finish in 60 minutes.  The chair is ifat_afek. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:01:43 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
08:01:45 <openstack> The meeting name has been set to 'vitrage'
08:01:48 <ifat_afek> Hi :-)
08:01:49 <e0ne> hi
08:04:50 <ifat_afek> Let’s start
08:04:52 <ifat_afek> #topic Status and updates
08:05:04 <eyalb> \o/
08:05:14 <ifat_afek> Tomorrow is Stein milestone 1, so according to the new policy we need to tag python-vitrageclient. I will tag today also vitrage and vitrage-dashboard, so they will be aligned.
08:05:23 <ifat_afek> Unless someone has a reason not to do it
08:05:59 <ifat_afek> I also want to tag stable/rocky in the coming days. I’m waiting for najjar’s bug fix for Prometheus
08:06:56 <ifat_afek> Besides that, I’m still working on the Nova versioned notifications issue. It requires a small refactoring in nova.instance datasource, so we will be able to accept both old and new notification format.
08:07:01 <ifat_afek> That’s it for me
08:07:07 <mnajjar> can you wait little bit with vitrage-dashboard
08:07:18 <ifat_afek> With Rocky or Stein?
08:07:41 <mnajjar> both or with the more important
08:07:57 <ifat_afek> With Rocky I can wait for sure
08:08:11 <mnajjar> ok
08:08:17 <e0ne> mnajjar: we have to release stein-2 tomorrow
08:08:38 <ifat_afek> Regarding Stein, I need to make sure that the existing vitrage-dashboard works with the new python-vitrageclinet. I assume there won’t be any problem becuase I don’t remember we changed the client recently
08:09:03 <ifat_afek> e0ne: we are working by cycle-with-intermediary mode, so we don’t need a stable release tomorrow
08:09:13 <ifat_afek> We just need to release the client, if I understand correclty
08:09:22 <ifat_afek> So it makes sense to release the backend as well
08:09:31 <ifat_afek> But I’m not sure if we have to release the dashboard
08:09:49 <ifat_afek> BTW, is it stein-2 or stein-1?
08:10:10 <mnajjar> i'm planning to look at the timestamp bung on template list page (vitrage-dashboard) today..
08:10:19 <mnajjar> that's why i'm asking
08:10:21 <e0ne> stein-1
08:10:42 <eyalb> i am preparing the lab for the openstack summit in berlin
08:10:45 <ifat_afek> mnajjar: if you fix it today then there is no problem at all, I can wait with the release until tomorrow
08:11:20 <mnajjar> great...yes i'm planning to fix it today
08:11:25 <ifat_afek> Great
08:12:06 <mnajjar> more than that, i worked on prometheus datasource
08:12:38 <ifat_afek> BTW, it might be a good idea to block the other approvals until the release
08:12:59 <ifat_afek> mnajjar: Are you talking about the new change with the Prometheus conf file?
08:13:12 <mnajjar> and added the mapping values of the resources from conf file
08:13:17 <e0ne> IMO, it's only first milestone, we'll have enough time to fix everything if we break it
08:13:30 <ifat_afek> e0ne: I guess you are right :-)
08:13:37 <e0ne> ifat_afek: :)
08:14:52 <ifat_afek> eyalb: Thanks for the update, it somehow got lost in the discussion…
08:16:39 <ifat_afek> Anything else? seems like we are very few today
08:16:56 <e0ne> I've got update about tempest plugin refactoring
08:17:28 <e0ne> I dig into it and it makes me sad:(
08:17:41 <mnajjar> why '=D
08:18:18 <e0ne> #link https://storyboard.openstack.org/#!/story/2004053
08:18:46 <e0ne> we use tempest in other way as it supposed to be
08:19:12 <ifat_afek> I believe we knew that… but what is the problem with python-*client exactly? why not call it?
08:19:36 <e0ne> we've got all kind of tests in tempest plugin
08:19:48 <ifat_afek> Some of them do not belong there, I guess
08:20:01 <e0ne> ifat_afek: originally, tempest created to verify API
08:20:08 <ifat_afek> Right
08:20:20 <e0ne> but we've got a good number of functional test there too
08:20:44 <e0ne> ifat_afek: tempest has own wrappers around clients
08:21:34 <ifat_afek> e0ne: Yes… and we are not using it. But fixing it might be a big effort. If we are using python-*client, does it cause a real problem?
08:21:35 <e0ne> #link https://github.com/openstack/tempest/blob/b53b9238181077bf20c75c2b6c1edc00204ca7e7/tempest/clients.py
08:22:14 <eyalb> it should not be a problem
08:22:14 <e0ne> ifat_afek: honestly, it doesn't
08:22:46 <ifat_afek> I mean, we have several problems in our tempests. One is that we use vitrage. We must fix it. Another is that we don’t work in the recommended way. I say let’s do it stage by stage
08:22:48 <eyalb> when it was written we didn't know about the tempest clients
08:22:48 <e0ne> the bigger problem is that we use some "internal" vitrage things
08:23:06 <eyalb> eOne: that is correct
08:23:06 <ifat_afek> e0ne: Of course, this is a real problem
08:23:45 <ifat_afek> e0ne: I noticed you pushed a fix for that, but I didn’t get to see the details yet
08:23:53 <e0ne> eyalb: I'm not saying that somebody did something wrong. I just note that we've got several issues now
08:24:32 <eyalb> eOne: yes thats why it is important to do the refactoring
08:24:40 <e0ne> ifat_afek: I removed vitrage constants usage only. we've got all that constants in tempest  plugin
08:24:52 <ifat_afek> e0ne: I guess someone made a mistake once, and then everybody copied&pasted…
08:24:53 <e0ne> eyalb: +1
08:25:02 <eyalb> it should be done step by step
08:25:08 <e0ne> eyalb: +1
08:25:25 <eyalb> first try to decouple tempest from vitrage
08:25:35 <ifat_afek> Do we want to create vitrage-lib? now? later? or just copy the constants? are there utilities that are common too?
08:25:40 <eyalb> then change the tests to be more api oriented
08:25:40 <e0ne> earlier we discussed that it could  be an option to introduce vitrage-lig
08:25:53 <e0ne> *vitrage-lib
08:26:18 <ifat_afek> If it’s needed, let’s go for it. If it’s just a matter of a few constants, maybe it’s a waste
08:26:32 <ifat_afek> We can also decide later I guess
08:26:38 <e0ne> IMO, we shouldn't do it. we can cleanup vitrage usage in tempest and move some tests to vitrage.tests.functional
08:26:49 <ifat_afek> OK
08:27:13 <e0ne> but it's only my opinion :). I'm open to any other proposals
08:27:54 <eyalb> I think we sould start cleaning the tempest
08:28:07 <ifat_afek> I say let’s start without it and we can change our minds later. Most of the work is around the decoupling anyhow
08:28:11 <e0ne> I created "Get rid of python-*client usages from vitrage-tempest-plugin" story because it's a small step to not use vitrage.os_clients
08:28:33 <ifat_afek> Oh, right. We shouldn’t use this file from vitrage
08:29:10 <eyalb> there is also the issue of configuration
08:29:21 <e0ne> eyalb: +1
08:29:22 <eyalb> we should do it the tempest way
08:29:46 <eyalb> currantly we use a vitrage method
08:29:52 <eyalb> prepare_service
08:30:00 <e0ne> I can finish removing vitrage.os_clients next next week and start work on configuration issues with tempest plugin
08:30:06 <eyalb> +1
08:30:16 <ifat_afek> +1
08:30:22 <ifat_afek> Sounds great
08:30:42 <e0ne> I want to focus on vitrage dashboard in stein-2 too
08:30:56 <e0ne> to remove vendorized scripts and use xstatic
08:31:06 <ifat_afek> This would help us a lot
08:31:12 <eyalb> eOne: go for it :-)
08:31:27 <e0ne> I think, that's all what can I do during this release
08:31:52 <ifat_afek> Thanks!
08:31:52 <e0ne> of course, I'm not sure. I need to work on some other projects too:(
08:32:02 <e0ne> so any help with tempest  is welcome :)
08:32:27 <ifat_afek> Of course, let us know when you need help
08:33:02 <e0ne> ifat_afek: will do, thanks
08:33:42 <ifat_afek> Anything else for today?
08:33:51 <e0ne> a short note
08:34:23 <e0ne> during creation new xstatic packages for vitrage-dashboard, I'll give +2 for them both to horizon and vitrage teams
08:34:38 <ifat_afek> What do you mean?
08:34:47 <ifat_afek> Is the change in vitrage-dashboard or in Horizon?
08:35:06 <e0ne> new xstatic packages will require new repos in github.com/openstack/ nam,espace
08:35:31 <e0ne> we did the same for heat-dashboard in the previous release
08:35:53 <ifat_afek> Ok, if you know how to do it it will save us a lot of time
08:36:19 <e0ne> so if any fixes will be needed to vitrage-dashboard specific xstatic packages, vitrage team will be able to do it withon horizon team
08:36:37 <e0ne> that's all from me
08:36:45 <ifat_afek> Ok, thanks
08:37:51 <ifat_afek> Let’s close the meeting
08:37:56 <ifat_afek> Bye :-)
08:38:09 <e0ne> see you next weel
08:38:11 <e0ne> see you next week
08:38:29 <ifat_afek> #endmeeting