15:00:03 <anteaya> #startmeeting third-party
15:00:04 <openstack> Meeting started Mon Sep 21 15:00:03 2015 UTC and is due to finish in 60 minutes.  The chair is anteaya. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:06 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:09 <openstack> The meeting name has been set to 'third_party'
15:00:23 <anteaya> hello
15:00:39 <mmedvede> hi
15:00:44 <anteaya> hey mmedvede
15:00:56 <anteaya> how is the day treaing you so far?
15:01:03 <lennyb> hi
15:01:09 <mmedvede> it is good :) thank you
15:01:15 <anteaya> mmedvede: glad to hear it
15:01:18 <anteaya> hi lennyb
15:01:30 <anteaya> any thoughts on what we should discuss today?
15:01:49 <lennyb> all I have is https://review.openstack.org/#/c/206513/
15:02:41 <anteaya> lennyb: yup you need another review from a system-config core reviewer
15:02:49 <asselin__> o/
15:02:51 <anteaya> #link https://review.openstack.org/#/c/206513/
15:02:54 <anteaya> hey asselin__
15:03:04 <anteaya> lennyb: keep your chin up :)
15:03:09 <asselin__> hi
15:03:10 <anteaya> I have something
15:03:21 <anteaya> this patch
15:03:25 <anteaya> #link https://review.openstack.org/#/c/186684/
15:03:32 <anteaya> it merged and broke some things
15:03:39 <anteaya> and now will be merging again soon
15:03:44 <anteaya> so take a look at it
15:03:57 <anteaya> it changes the default identity api version in devstack
15:03:57 <asselin__> lennyb, we should get that merged soon
15:04:13 <anteaya> and take steps if you think it will break you
15:04:19 <lennyb> asselin_: thanks
15:05:16 <asselin_> anteaya, yeah, we got caught by that patch the first time around and fixed our ci. I didn't revert it, so we should still be good.
15:05:24 <anteaya> asselin_: okay great
15:05:36 <asselin_> big difference is before we were using cinder client calls, and now we use openstack client calls
15:05:46 <anteaya> ah good to know
15:05:53 <anteaya> anyone else in this situation?
15:06:24 <lennyb> we are always working with devstack upstream, so if it's merged I guess we are ok.
15:06:40 <anteaya> lennyb: my point is it was merged and then reverted
15:06:42 <asselin_> I think it also impacted manila folks, and they fixed their base scripts
15:06:45 <anteaya> and will be merging again
15:06:50 <anteaya> asselin_: good to know
15:07:02 <lennyb> anteaya: I will check this, thanks
15:07:08 <anteaya> lennyb: great thank you
15:07:22 <anteaya> anyone else with any thoughts on the identity api change?
15:08:10 <anteaya> an email will be going out soon to the -dev list
15:08:16 <anteaya> informing people of this change
15:08:23 <anteaya> just wanted to give you folks a heads up
15:08:33 <anteaya> anything more on this topic?
15:08:55 <anteaya> does anyone have a different topic they would like to discuss?
15:09:42 <asselin_> #link manila identify patch fix: https://review.openstack.org/#/c/220601/
15:09:59 <anteaya> thank you
15:11:32 <anteaya> anything more here?
15:12:04 * anteaya is uncertain if we are moving on or not
15:12:40 <asselin_> I think we can movw on
15:12:45 <anteaya> oh good
15:12:48 <asselin_> move*
15:12:59 <anteaya> does anyone have any other topic they wish to discuss today?
15:13:30 <mmedvede> does anybody explicitly test devstack on their CI?
15:13:43 <asselin_> we don't
15:14:01 <lennyb> not us, but maybe I will start listening to it.
15:14:07 <anteaya> mmedvede: do you mean do a test run of devstack without running the tempest test suite?
15:14:38 <mmedvede> anteaya: I mean to run the full tempest. But feed it off patches to devstack
15:14:59 <mmedvede> so it is easier to see exactly which one breaks our CI
15:15:07 <mmedvede> we were thinking to start doing it
15:15:41 <anteaya> feed it off patches to devstack <- I'm uncertain what you mean by this phrase
15:15:45 <mmedvede> so I wondered if anybody else is doing it, but not reporting
15:15:45 <asselin_> I had that idea too but wasn't sure devstack would want out ci commenting on it.
15:16:10 <anteaya> oh have your ci report on devstack patches?
15:16:15 <mmedvede> asselin_: yes, that is what I am afraid of
15:16:25 <anteaya> well you can ask
15:16:36 <anteaya> mtreinish: has given permission to other systems in the past
15:16:46 <asselin_> and if we don't comment, we don't have a good way to monitor the results and make them useful
15:16:52 <mmedvede> anteaya: yes. They might complain that it would be noisy if every third-party CI starts reporting
15:17:06 <anteaya> you see citrix xenserver ci commenting: https://review.openstack.org/#/c/186684/
15:17:07 <mmedvede> anteaya: yes, we would ask
15:17:15 <lennyb> we were broken few days ago due to pbr version issue, that maybe could have been saved
15:17:21 <anteaya> mmedvede: right you don't know until you ask
15:17:40 <anteaya> and you have to ensure you pay close attention to your system so it doens't post noise
15:17:48 <anteaya> shutting off commenting if it goes sideways
15:18:00 <eantyshev> lennyb: +1 good example
15:18:05 <anteaya> and correct, not very ci would be allowed
15:18:22 <anteaya> you would have to enusre you take maintenance of your system very seriously
15:18:23 <asselin_> perhaps it would be good to get some volunteers reporting from various projects
15:18:39 <anteaya> well actually I think it is best if individuals just asked
15:18:51 <anteaya> and then mtreinish can decide on a case by case basis
15:19:06 <anteaya> and if you want to use someone else devstack results that is up to you
15:20:46 <mtreinish> anteaya: what am I deciding?
15:20:58 <anteaya> hi there
15:21:15 <anteaya> well the ci operators are talking about devstack patches
15:21:30 <mmedvede> mtreinish: I was wandering if it would be ok for a third-party CI to report on devstack
15:21:35 <asselin_> mtreinish, there's a question whether there's any value to devstack if select 3rd party ci systems post comments
15:21:36 <anteaya> and are discussing whether you would allow any additional cis to report on devstack kpatches
15:21:46 <anteaya> mtreinish: so I said it is your decision
15:21:47 <mmedvede> I noticed that devstack is most frequent culprit of us breaking
15:22:18 <anteaya> mtreinish: and I'm encouraging individuals to ask permisison for their system
15:22:24 <mtreinish> mmedvede: heh, not the weird architecture you run things on :)
15:22:36 <anteaya> rather than asking for a general broadsweeping permission for all cis
15:22:48 <mmedvede> mtreinish: hehe, true
15:23:02 <mtreinish> but, sure I don't think there is anything wrong with adding additionally cis to devstack or tempest on a case by case basis
15:23:03 <asselin_> anteaya, I don't think anyone's asking for broadsweeping permission
15:23:20 <anteaya> asselin_: great, glad we have that in the logs
15:23:25 <mmedvede> mtreinish: but if we were exactly the same, there would not be a case for testing at all :)
15:24:37 <mmedvede> thank you, that answers my inquiry
15:25:06 <anteaya> so just offer mtreinish the name of your ci and a link so he can see the operational history
15:25:19 <anteaya> mtreinish: would that be sufficient for you to make a decision?
15:25:57 <mtreinish> anteaya: sure should be
15:26:03 <anteaya> great thank you
15:26:10 <anteaya> anything more on this topic?
15:27:00 <anteaya> does anyone have any other topic they would like to discuss today?
15:27:42 <anteaya> asselin_: have you made any progress on selecting a third party dashboard for infra to host?
15:28:03 <asselin_> anteaya, we had very light participation last week
15:28:27 <anteaya> what needs to happen to make progress?
15:28:43 <asselin_> anteaya, we discussed the current solutions and compared the various solutions
15:28:50 <anteaya> that is helpful
15:28:54 <anteaya> any conclusions?
15:29:17 <lennyb> btw, jogo's server is no longer avaliable #link http://52.27.62.150/lastcomment.json
15:29:22 <asselin_> anteaya, well I have my preference, but the meeting lacked the 2 main developers
15:29:42 <mmedvede> I also had conflict, so could not attend
15:29:46 <anteaya> lennyb: that is not surprising as jogo is not working on openstack currently
15:29:47 <asselin_> sweston's solution is still not up and running
15:30:02 <anteaya> asselin_: ah yes having developer input is important
15:30:06 <anteaya> asselin_:  :(
15:30:21 <asselin_> lennyb, I was able to launch his solution locally.
15:30:22 <anteaya> lennyb: the code should be available though
15:30:23 <mmedvede> I actually wrote initial puppet deployment module for patrickeast 's scoreboard
15:30:42 <anteaya> okay so sounds like we are no closer to agreement
15:30:52 <asselin_> honestly, they'll all bring value. Personally, I'd like to see if we can merge the solutions into one
15:30:56 <anteaya> okay thanks for keeping on the forefront of discussion
15:31:05 <anteaya> asselin_: I'm not against that
15:31:24 <anteaya> but infra needs to know how the puppet module should look so it can be merged
15:31:37 <anteaya> so one solution to start with and adding later after it is up
15:31:39 <lennyb> asselin_: what do you mean ( how did you do it locally ) ?
15:31:43 <anteaya> would be what I envision
15:31:50 <mmedvede> anteaya: +1
15:32:00 <asselin_> lennyb, I just followed the instructions in his readme
15:32:43 <anteaya> so I'm hearing no consensus yet
15:32:50 <asselin_> we also lacked inputs from mmedvede rfolco and other consumer stakeholders
15:32:59 <anteaya> yup that is fine
15:33:06 <anteaya> let me know how it goes
15:33:10 <anteaya> thank you
15:33:10 <asselin_> anteaya, not enough people to form a concensus
15:33:15 <anteaya> okay
15:33:43 <asselin_> but perhaps we can discuss a bit more now if there's time?
15:33:59 <anteaya> oh sure
15:34:01 <lennyb> asselin_: could you drop me a link to the readme, pls ?
15:34:11 <anteaya> if y'all want to discuss here, that's great
15:34:22 <anteaya> #topic third-party dashboard
15:34:23 <mmedvede> scoreboard can be deployed the fastest, I am a bit subjective here :), here is the module
15:34:28 <mmedvede> #link https://review.openstack.org/214823
15:34:36 <asselin_> #link 3rd party ci dashboard comparison: https://etherpad.openstack.org/p/third-party-ci-dashboard-comp
15:35:09 <asselin_> #link scoreboard (patrickeast) http://ec2-54-67-102-119.us-west-1.compute.amazonaws.com:5000/?project=openstack%2Fnova&user=&timeframe=24
15:35:38 <asselin_> #link ci-watch (skyler) http://ci-watch.tintri.com/project?project=nova&time=24+hours
15:36:38 <asselin_> #link last-commend (jogo) https://github.com/stackforge/third-party-ci-tools/tree/master/monitoring/lastcomment-scoreboard
15:37:17 <mmedvede> is ci-watch source code available already?
15:37:26 <asselin_> mmedvede, yes, it's merged
15:37:43 <asselin_> https://github.com/stackforge/third-party-ci-tools/tree/master/monitoring/ciwatch
15:37:45 <asselin_> 20 days ago
15:39:24 <asselin_> please speak openly as to what you like or not
15:39:47 <mmedvede> asselin_: I like that ci-watch has per-job view
15:40:04 <anteaya> ci-watch loads faster (refreshes faster) than scoreboard
15:40:16 <anteaya> other than that, I like them both the same
15:42:08 <anteaya> does anyone have any other comments?
15:42:12 <mmedvede> I think ci-watch collects more information to its db compared to scoreboard. I like how you can filter the scoreboard per project, and that scoreboard collects all events, while ci-watch seems to only have a few projects at the moment
15:42:30 <lennyb> what is the main goal of this board? Seeing this constantly or having some internal scripts ( REST ) to check our CI against the others ?
15:42:32 <asselin_> for our ci, the per-job view is really valuable
15:43:02 <mmedvede> superficially, ci-watch might be a better starting point, but I did not look at source code, neither tried to automate its deployment
15:43:38 <mmedvede> +1 on per-job view
15:44:04 <asselin_> lennyb, well, for example, you can see which patches your ci commented on, and which it didn't along with the result.
15:44:41 <asselin_> lennyb, see which jobs are having trouble
15:45:05 <asselin_> lennyb, see if upstream jenkins is ok or not
15:45:47 <asselin_> lennyb, as a developer get a quick indication of a 3rd party ci past history to know if you should even care about it's results in your super-awesome-patch that it says doesn't work
15:46:45 <asselin_> in short, for operators: gage how well your system is doing so you can improve it and fix issues
15:47:28 <asselin_> for developers: establish a level of trust so you can better focus your time on code
15:47:46 <anteaya> very important
15:48:18 <asselin_> for ptls: see who's got ci and who doesn't
15:48:43 <krtaylor> o/
15:48:45 <anteaya> all reasons infra would like to host one
15:48:58 <anteaya> so folks know where to go and have similar information
15:49:05 <krtaylor> sorry I'm late, good to see dashboard discussion
15:50:02 <krtaylor> I'd ask a different way, is there a problem with using scoreboard short-term, and then focusing on a merged solution?
15:50:47 <anteaya> yes
15:51:01 <anteaya> that approach has been a blocker for about 18 months
15:51:12 <anteaya> infra wants to host _one_
15:51:19 <asselin_> krtaylor, honestly I don't use scoreboard b/c it doesn't have the per-job stats
15:51:20 <anteaya> not a different one as the mood changes
15:51:32 <asselin_> krtaylor, if we need to pick one of the two today, I'd pick ci-watch
15:51:32 <krtaylor> yes, one, that is the proposal
15:51:48 <krtaylor> fine, everybody agree?
15:52:22 <asselin_> krtaylor, agree to what?
15:52:22 <mmedvede> +1 on ci-watch functionality
15:52:31 <krtaylor> ci-watch
15:52:31 <asselin_> perhaps we should vote?
15:52:53 <krtaylor> s/scoreboard/ci-watch - easy fix to spec
15:53:16 <asselin_> krtaylor, and s/temporary//g
15:53:17 <krtaylor> +1 to whatever gets a hosted solution approved
15:53:30 <krtaylor> asselin, fair point
15:53:34 <asselin_> or s/temporary/permanent/g
15:53:48 <anteaya> just removing temporary should be fine
15:53:51 <asselin_> we'll iterate from there
15:54:10 <krtaylor> I'm fine with that, but it was jeblair 's idea to have a quick fix, I agreed to take the pen
15:54:27 <anteaya> well jim thought it would be quick
15:54:30 <anteaya> and it wasn't
15:54:31 <krtaylor> if it is a permanent, thats fine, we can work on merging punctionality into it
15:54:37 <krtaylor> whatever it is
15:55:01 <krtaylor> anteaya, yes, too many -1 for typos  :)
15:55:18 * asselin_ thinks we should vote
15:55:37 <anteaya> asselin_: what do you want to vote on?
15:55:56 <krtaylor> y, 5 mins left
15:56:09 <anteaya> asselin_: what question and what options
15:56:15 <anteaya> I can start a vote
15:56:27 <asselin_> #vote which solutions should we start with to monitor 3rd party ci. scoreboard, ci-watch, need-more-info
15:56:37 <asselin_> as infra-hosted
15:57:07 <mmedvede> asselin_: quick question - you have been running ci-watch internally?
15:57:14 <asselin_> mmedvede, no
15:57:30 <krtaylor> I like CI-watch's functionality, but scoreboard's layout
15:57:48 <mmedvede> my hope is to be able to allow flexibility in layout later
15:57:53 <mmedvede> *some
15:57:56 <krtaylor> agreed
15:58:08 <krtaylor> not worth hanging up the progress
15:58:14 <asselin_> krtaylor, +1
15:58:28 <krtaylor> vote? 2 mins :)
15:58:46 <krtaylor> +1 ci-watch
15:58:54 <anteaya> #startvote chose the infra-hosted third party ci monitoring solution #options scoreboard, ci-watch, need-more-info
15:58:55 <openstack> Unable to parse vote topic and options.
15:59:01 <mmedvede> +1 ci-watch
15:59:07 <anteaya> I don't have the syntax right
15:59:15 <asselin_> +1 ci-watch
15:59:22 <anteaya> that will do I think
15:59:28 <anteaya> as it is in the logs
15:59:32 <anteaya> time for the meeting to end
15:59:42 <anteaya> thanks all for your participation and attendance
15:59:45 <asselin_> thanks
15:59:48 <anteaya> see you next week
15:59:52 <anteaya> #endmeeting