02:35:59 <ekcs> #startmeeting congressteammeeting
02:36:01 <openstack> Meeting started Fri Jan  5 02:35:59 2018 UTC and is due to finish in 60 minutes.  The chair is ekcs. Information about MeetBot at http://wiki.debian.org/MeetBot.
02:36:02 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
02:36:04 <openstack> The meeting name has been set to 'congressteammeeting'
02:36:34 <ekcs> hi all welcome back and happy new year!
02:36:37 <ekcs> topics here as usual. please read/comment/add =) https://etherpad.openstack.org/p/congress-meeting-topics
02:37:19 <ramineni_> ekcs: hi .. happy new year !! :)
02:37:40 <ekcs> hi ramineni_ ! happy new year to you too!
02:37:53 <ekcs> how have you been? hope you had a nice time away =)
02:38:33 <ramineni_> Ya .. had a great vacation .. catching up things now :)
02:38:42 <ekcs> awesome
02:39:02 <ekcs> i had a nice little break too.
02:39:22 <ekcs> let’s get started then.
02:39:31 <ekcs> #topic tempest plugin
02:40:00 <ekcs> great to see all the progress on tempest plugin! ramineni_ would you like to give a quick update on where we stand?
02:40:48 <ramineni_> Required infra patches are merged yesterday .. so we are good to merge required tempest patches now
02:41:24 <ramineni_> I'll look into it ..and merge required patches .. please have a look and let me know any comments
02:42:03 <ekcs> great! yup I see these patches are now passing. https://review.openstack.org/#/q/project:openstack/congress-tempest-plugin+status:open
02:42:18 <ekcs> I also added one for the port attach/detach
02:42:50 <ramineni_> ekcs : in plugin repo .. Should we enable all gate jobs ? Or only mysql  jobs enough
02:44:04 <ekcs> I’m not totally sure how it’s all supposed to work. when it’s all done, the gate tests on congress server repo would pull from tempest plugin to run right?
02:44:37 <ramineni_> Yes ..
02:45:23 <ramineni_> I'm yet to do that change .. Will propose that by next week
02:46:03 <ekcs> got it.
02:46:30 <ekcs> so the purpose of the dsvm tests in tempest is to make sure we don’t break things when we modify tempest tests.
02:46:44 <ekcs> I’m not sure which way is best.
02:47:15 <ekcs> why don’t we try just mysql for now?
02:47:36 <ekcs> if we break something in tempest repo, we’ll see it next time a congress patch runs.
02:48:17 <ramineni_> Ya .. Little confusing for me too .. need to look into it
02:48:58 <ekcs> I think mysql + replicated would be good. but we can always see how thinghs go and adjust.
02:49:18 <ramineni_> Ok .. sounds good
02:49:24 <ekcs> i’m still not sure how to do things when I make a congress change with an accompanying tempest change.
02:49:42 <ekcs> that way the congress change depends on the tempest change to pass gate, but the tempest change also depends on the congress change.
02:50:05 <ramineni_> Ya ..I'll look into that .. and document the procedure everyone follow
02:50:18 <ekcs> great thanks a lot!
02:51:06 <ekcs> just to wrap this up, all the tempest changes lgtm. I left +1/+2’s. but leaving it to you for merging is that cool?
02:51:24 <ramineni_> Ya .. thanks
02:52:46 <ekcs> cool great to have so much progress there!
02:52:51 <ekcs> moving on then?
02:53:13 <ramineni_> Ok
02:53:33 <ekcs> ok.
02:53:34 <ekcs> #topic patches
02:53:52 <ekcs> I merged all the patches from orange.
02:54:06 <ekcs> and also merging the neutron attach/detach patch finally.
02:54:19 <ekcs> anything we want to discuss on the patches?
02:54:34 <ramineni_> Ok ..
02:54:34 <ekcs> open patches:
02:54:34 <ekcs> server: https://review.openstack.org/#/q/project:openstack/congress+status:open
02:54:35 <ekcs> tempest: https://review.openstack.org/#/q/project:openstack/congress-tempest-plugin+status:open
02:54:36 <ekcs> dashboard: https://review.openstack.org/#/q/project:openstack/congress-dashboard+status:open
02:54:37 <ekcs> client: https://review.openstack.org/#/q/project:openstack/python-congressclient+status:open
02:56:18 <ramineni_> When is queens 3?
02:56:44 <ramineni_> Month-end right
02:57:12 <ekcs> 25 January 2018
02:57:19 <ekcs> oh wow a bit earlier than I thought haha.
02:57:29 <ekcs> lots I’m trying to get done still =p
02:57:59 <ramineni_> Ok :)
02:58:50 <ekcs> do you think you’ll get to monasco push driver?
02:59:20 <ekcs> that’d be great to have in for queens. ready in time for self-healing discussions at PTG.
02:59:53 <ramineni_> I'll try ..but for monasca push driver ..we need some changes in monasca
03:00:17 <ramineni_> Monasca notification webhook doesn't support keystone auth
03:00:36 <ramineni_> So unable to call the congress api as it needs auth
03:00:57 <ekcs> oh i see. hmmm.
03:01:14 <ekcs> I wonder how masahito did it with zabbix.
03:01:21 <ramineni_> I need to discuss the same with monasca team of they have any plans
03:01:26 <ekcs> maybe zabbix is more flexible.
03:01:36 <ramineni_> For doctor?
03:01:40 <ekcs> yea.
03:01:58 <ramineni_> Their doctor script sends authenticatrd request only
03:02:11 <ramineni_> I have checked their script
03:02:19 <ramineni_> In opnfv
03:02:37 <ekcs> oh do you know how they get updates from zabbix?
03:02:57 <ekcs> does zabbix make the api call directly or does it get passed to someone else who then makes api call to congress?
03:03:08 <ramineni_> Yes
03:03:36 <ramineni_> There is no direct integration from zabbix to Congress as far as I know
03:04:26 <ramineni_> If we add monasca push driver we want to make that direct integration
03:04:28 <ekcs> so there is something doctor team wrote that listens to zabbix and forwards to congress?
03:04:49 <ramineni_> ekcs: yes ,AFAIK
03:05:24 <ekcs> got it. yea that sounds reasonable. ok maybe i’ll verify with masahito.
03:05:43 <ekcs> seems like an issue we can think about with generic push driver.
03:05:59 <ekcs> may not make sense to require keystone auth.
03:06:09 <ramineni_> ekcs: ok , ya
03:06:41 <ekcs> do you know what alternative way of security verification monasca or zabbix do when they do webhook?
03:07:10 <ekcs> can’d be nothing right?
03:07:57 <ramineni_> Right now its unauthenticated ..I checked monasca code
03:08:22 <ramineni_> But should be easy to implement in monasca I guess ..
03:09:24 <ekcs> hmm got it. i wonder if congress can also expose an unauthenticated api method.
03:09:52 <ekcs> and just document that you need to set that up in properly protected network.
03:10:05 <ekcs> that only allows traffic from specific senders.
03:14:21 <ekcs> anyway. if you have thoughts on how best to proceed let me know. would be great to have something in for queens-3 even if limited.
03:14:54 <ekcs> but we can probably move on for now?
03:16:05 <ekcs> oh btw the mistral driver patch is ready too. it’s in merge conflict but I can deal with that after review.
03:16:20 <ekcs> https://review.openstack.org/#/c/529427/
03:16:54 <ramineni_> ekcs: ok , will check .. sorry got disconnected
03:17:23 <ekcs> oh np.
03:18:38 <ekcs> let’s move on then
03:18:53 <ramineni_> Ok
03:18:57 <ekcs> #topic congress data type
03:19:22 <ekcs> noticed this while working on mistral.
03:19:44 <ekcs> right now congress converts everything to strings in datasource translation.
03:19:49 <ekcs> for example, {'id': '12-34', value: 1} gets converted to row ('12-34', '1')
03:20:13 <ramineni_> ekcs: yes, I observed too
03:20:31 <ekcs> seems like an old issue that never got fixed. arosen left a fixme there: https://github.com/openstack/congress/blob/master/congress_tempest_tests/tests/scenario/congress_datasources/test_glancev2.py#L86
03:21:02 <ekcs> so I’m thinking maybe we should fix it. shouldn’t be a hard fix. really just fixing the value translator.
03:21:03 <ramineni_> Ohok
03:21:11 <ekcs> problem is, it’s not backward compatible behavior.
03:21:35 <ekcs> so I’m thinking we can change the default behavior but add an optional flag someone can set to preserve current bad behavior.
03:21:37 <ramineni_> But are there any datasources columns have int column
03:21:49 <ramineni_> Integer column
03:22:19 <ramineni_> Otherwise should not be problem
03:22:42 <ekcs> not many I don’t think. but I tihnk there are some, thus the fixme.
03:22:48 <ekcs> like cinder volume size
03:22:53 <ekcs> i’m guessing is int.
03:22:57 <ekcs> but I haven’t verified.
03:23:42 <ekcs> maybe it’s not a big deal though we could leave it for next cycle.
03:24:08 <ekcs> I am kind of torn because if it’s something bad we need to break compatibility to fix, earlier is always better.
03:24:44 <ekcs> on the other hand it’s not necessarily super urgent now because no one totally requires it.
03:24:45 <ekcs> any thoughts?
03:25:01 <ramineni_> Ya ..may be new datasources could use new translator and later we can check how to deal with old ones
03:25:21 <ramineni_> Like mistral etc ..
03:26:11 <ekcs> hmm
03:26:21 <ekcs> that could make sense.
03:26:40 <ekcs> well. last 4 minutes. anything else we wanna talk about either on data type or anything else?
03:27:44 <ramineni_> No .. nothing from my side
03:28:00 <ekcs> ok then.
03:28:39 <ekcs> urgency aside, what do you think about the idea of changing default behavior but adding a config flag or even a datasource config to preserve old behavior?
03:29:48 <ramineni_> Hmm .. like while creating datasource
03:29:50 <ramineni_> ?
03:30:12 <ekcs> config flag would be set on congress conf.
03:30:23 <ekcs> but we could also have new datasource config that get’s set while creating DS.
03:30:28 <ekcs> both are options.
03:31:30 <ramineni_> Hmm ..not sure ..then we couldn't deprecate anytime
03:32:06 <ramineni_> Ok ..I need to leave .. Should we add this topic to next meeting
03:32:21 <ekcs> great later then!
03:33:35 <ekcs> #endmeeting