15:00:33 <krtaylor> #startmeeting third-party
15:00:35 <openstack> Meeting started Wed Apr 29 15:00:33 2015 UTC and is due to finish in 60 minutes.  The chair is krtaylor. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:36 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:39 <openstack> The meeting name has been set to 'third_party'
15:00:49 <asselin> o/
15:00:56 <patrickeast> hi everyone
15:00:58 <krtaylor> Hi asselin
15:01:02 <zz_ja> morning
15:01:03 <krtaylor> hi patrickeast
15:01:08 <krtaylor> hi zz_ja
15:01:12 <marcusvrn_> Hi
15:01:18 <krtaylor> hi marcusvrn
15:01:18 <mmedvede> hi
15:01:22 <rfolco> o/
15:01:23 <krtaylor> hi mmedvede
15:01:28 <krtaylor> hi rfolco
15:01:36 <rhe00_> hi
15:02:20 <krtaylor> marcusvrn, glad you found out about this meeting, its a good point that we should advertise it more
15:02:31 <krtaylor> hi rhe00_
15:02:45 <krtaylor> ok, lets get started
15:02:54 <krtaylor> here's the agenda for today
15:02:55 <nfedotov> hello all
15:03:02 <marcusvrn_> Yeah \o/
15:03:05 <krtaylor> #link https://wiki.openstack.org/wiki/Meetings/ThirdParty#4.2F29.2F15_1500_UTC
15:03:17 <krtaylor> hi nfedotov
15:03:59 <krtaylor> so a quick announcement
15:04:21 <krtaylor> make note of the Gerrit 2.9 upgrade Saturday May 9
15:05:48 <krtaylor> #topic Topics for discussion at Liberty summit
15:06:20 <krtaylor> the proposed cross-project session is getting good remarks, looks lik eit may actually happen
15:06:58 <krtaylor> I would like for it to be more of a discussion, working session or fish bowl
15:07:21 <krtaylor> here is the topics etherpad link:
15:07:43 <krtaylor> #link https://etherpad.openstack.org/p/liberty-third-party-ci-working-group
15:08:33 <krtaylor> asselin, there is a proposed infra session for the downstream puppet work, correct?
15:08:44 <asselin> krtaylor, yes
15:09:07 <asselin> #link https://etherpad.openstack.org/p/infra-liberty-summit-planning
15:09:43 <krtaylor> asselin, ok, so we can have that discussion there then, and move documentation to a work items that will follow that work
15:09:57 <asselin> yes
15:10:57 <krtaylor> So I think #4 is important, but may be able to handle in a QA open session
15:11:42 <krtaylor> so, scanning the list, it looks like the highest priority items to me anyway, are #1 and #3
15:12:44 <krtaylor> although #3 is already happening, so that may just be a spot to get operators to contribute
15:13:06 <asselin> #9 is important to us. If not in 3rd party, we can see about including that in cinder.
15:13:18 <asselin> 10 is covered in -infra session
15:13:24 <patrickeast> asrangne: +1 for #9
15:13:33 <patrickeast> asselin: *
15:13:54 <krtaylor> asselin, I agree, but it does have a smaller audience
15:14:21 <asselin> is that a cinder issues, or do other projects have a similar need?
15:14:37 <asselin> for #9
15:15:23 <patrickeast> i assume other vendors have variable configurations they want to test out, right?
15:15:39 <krtaylor> Is it a common problem in cinder?
15:15:51 <krtaylor> testing multiple configs per patch?
15:16:18 <patrickeast> most backends have multiple switches they can fiddle with for the storage device
15:16:29 <asselin> the alternative is to setup specific jobs for each configuration.
15:16:30 <patrickeast> different authentication, encryption, fabrics, etc
15:16:59 <asselin> yes, there are different backend configurations & openstack configurations.
15:17:16 <asselin> e.g. is multipath installed & enabled in nova.conf?
15:17:50 * ctlaugh_ got in a little late
15:18:07 <krtaylor> I could see neutron potentially having similar
15:19:05 <asselin> it increases the job permutations...and for some it seems they should be able to test within a single job...
15:20:22 <krtaylor> is there requirements for the testing to include the permutations?
15:20:40 <asselin> from cinder no. for customers, yes
15:21:03 <krtaylor> yep, ok, so a "its the right thing to do" test
15:21:27 <krtaylor> we have a few of those requirements as well
15:21:35 <krtaylor> ok, so let's leave that on the short list for summit
15:22:01 <asselin> I think a brainstroming session would be good
15:22:49 <krtaylor> agreed, we need to get broad input for how to improve trust in CI systems
15:23:17 <krtaylor> for #7 nfedotov, want to comment
15:23:49 <krtaylor> for #7 these would be tests also outside the required ?
15:24:32 <nfedotov> Yes. There are tests that could not be merged to tempest
15:25:01 <asselin> my opinion is yes it should be allowed.
15:25:14 <asselin> and should be encouraged, in fact
15:25:50 <krtaylor> nfedotov, what is being asked, how to handle? requirements to do so?
15:25:57 <nfedotov> I think so too. It may be a non voiting job. Somebody who are interested in results may look at it.
15:26:29 <zz_ja> nfedotov, acceptance probably depends on articulating the voting status clearly.
15:26:56 <zz_ja> It's hard to imagine objections to non-voting "informational" results
15:27:14 <krtaylor> this may be related to #5, where there are tests that *should* be run, but are not required, seems like that is a recurring subject here
15:27:46 <nfedotov> Yes it is related to #5.
15:28:07 <zz_ja> krtaylor, "should" vs "must" ... any binary modifier is probably going to annoy someone.  tagging seems to be the panacea nowadays.
15:28:19 <krtaylor> we (powerkvm) are beginning to test outside required projects, we have been planning to post thost log results to our third party system wiki page
15:28:44 <zz_ja> ...I think of should vs must in terms of degree of completeness.  gate is min-required, other levels could be articulated.
15:29:14 <krtaylor> so the issue is that some tests need to be run, but the projects don't necessarily want to see the results of these posted to each patch, too much noise
15:29:37 <krtaylor> so, maybe a community place to store "extra" tests?
15:29:49 <krtaylor> indexed by patch number?
15:30:12 <krtaylor> ok, that would be a good brainstorming subject, and also cross-project
15:30:12 <ctlaugh_> #5 (at least what I was hoping to convey) is that there are test scenarios (like we are running with ARM) where we can't possibly trigger off of (and comment on) every project and every commit.  I'm looking for a place to report test results that are not tied to a specific patch.
15:30:15 <asselin> krtaylor, I prefer to keep it posted with the patch itself
15:30:34 <asselin> krtaylor, perhaps a gerrit UI update can help filter the noise
15:30:38 <krtaylor> asselin, agreed, but the projects don't agree
15:31:02 <krtaylor> hm, so a better CI comments toggle
15:31:24 <asselin> krtaylor, for example
15:32:35 <krtaylor> ok, so do we agree that these can be grouped into a discussion topic?
15:32:47 <asselin> +1
15:32:53 <patrickeast> +1
15:33:57 <zz_ja> +1
15:33:59 <krtaylor> ok, so we have #1 and #5/7/9
15:34:00 <marcusvrn> +1
15:34:26 <ctlaugh_> +1
15:34:48 <krtaylor> #4 may be QA, #2 follows #10 to infra
15:34:56 <krtaylor> that leaves #8
15:35:08 <krtaylor> #8 is a great work topic for liberty
15:35:46 <asselin> perhaps that one should also be proposed in -infra
15:36:01 <patrickeast> yea they would probably have good input for it
15:36:06 <krtaylor> #8 lends itself to good best practices discussions for caching/optimizing too
15:36:31 <patrickeast> ^ that might be something lacking in our documentation
15:36:34 <zz_ja> +1 infra
15:36:47 <krtaylor> that can stay on the active work list for liberty, but do we need to discuss at summit?
15:37:15 <asselin> krtaylor, if not formally, we can discuss informally
15:37:18 <zz_ja> if we think it lands on infra, might be useful to take the temp of the room on it
15:37:38 <krtaylor> #link https://wiki.openstack.org/wiki/ThirdPartyCIWorkingGroup#Development_Priorities
15:37:53 <zz_ja> ...if there's a groundswell of support, harder for others to say it's not important to infra
15:38:02 <krtaylor> just in case some don't know we track CI working group interests  :)
15:38:42 <krtaylor> zz_ja, we haven't had much "groundswell" on anything  :)
15:39:21 <zz_ja> might be how the question is asked.
15:39:40 <zz_ja> if we talk about how to profile *a CI test*, yawn.
15:39:59 <asselin> I'm thinking more like how it's done in disk image builder
15:40:11 <zz_ja> if we talk about how to profile any scenario, well every consumer of the main repo should have that problem.
15:40:16 <asselin> it has built in profiling.
15:40:55 <krtaylor> ok, I'll summarize the topics at the top of that etherpad after this meeting and email for comments
15:40:56 <asselin> zz_ja, this would be profiling all the parts of the ci job itself
15:41:20 <krtaylor> any other comments about summit discussion priorities?
15:41:32 <zz_ja> depends what "all the parts" means asselin
15:42:12 <asselin> so the jobs have timestamps. Would be nice to have a tool that can consume that and give your profile of the entire ci job.
15:42:15 <zz_ja> ... profiling often can only see "so far" down into the stack.  depends on implementation what that means though.
15:42:29 <zz_ja> sure
15:42:41 <asselin> covering all the bash scripts, devstack, tempest, cleanup, uploading, etc.
15:42:46 <zz_ja> as well as its individual components; it's a classic drill down pattern
15:43:23 <krtaylor> I think you are both in agreement that is should be discussed :)
15:43:47 <zz_ja> krtaylor, you and emily will have to proxy of course, no travel for me
15:44:05 <krtaylor> we need to move along in the agenda (time check)
15:44:10 * asselin steps away for a minute
15:44:50 <krtaylor> we can handle clarification of topic discussion points via email and etherpad once I get the summit list
15:44:56 <krtaylor> ok, onward then
15:45:49 <krtaylor> #topic Repo for third party tools
15:46:26 <krtaylor> so, the patch to create the repo got merged, hopefully thats not news to anyone
15:46:35 <krtaylor> here is our shiny new repo
15:46:38 <krtaylor> #link https://github.com/stackforge/third-party-ci-tools
15:47:07 <krtaylor> and I have pushed an initial patch to get basics out there
15:47:27 <krtaylor> but we need approvers/cores to get patches merged
15:47:52 <krtaylor> I have proposed asselin, patrickeast, mmedvede, sweston as cores
15:48:22 <krtaylor> I have not heard back from sweston yet, but can remove him later if he does not want to take it on
15:48:54 <krtaylor> oh and me
15:49:18 <krtaylor> I have not heard anything from the email sent
15:49:35 <krtaylor> so I am assuming that there is overwhelming support for those proposed
15:49:40 <krtaylor> comments?
15:49:52 <patrickeast> sounds good to me
15:50:11 <asselin> from a diversity point of view, do we have anyone from neutron?
15:50:16 <patrickeast> i assume we can always adjust as we go as needed
15:50:18 <zz_ja> +1
15:50:34 <krtaylor> lets vote, I love using the meeting tools
15:51:09 <krtaylor> #startvote initial cores as proposed be added to third-party-ci-repo
15:51:09 <openstack> Unable to parse vote topic and options.
15:51:34 <krtaylor> hm, interesting
15:51:45 <marcusvrn> how can I use this tool to vote?
15:52:09 <asselin> #help ?
15:52:10 <krtaylor> hm, it should have given us instructions to use yes or no
15:52:17 <zz_ja> stop reading the pictures in krtaylor's mind ;-)
15:52:31 <krtaylor> #startvote
15:52:33 <openstack> Unable to parse vote topic and options.
15:52:49 <krtaylor> ok, whatever a simple +1 -1 will do
15:52:53 <zz_ja> krtaylor, try your original without the hyphens
15:52:53 <mmedvede> #startvote agree on initial cores for new repo?
15:52:54 <openstack> Only the meeting chair may start a vote.
15:52:58 <mmedvede> haha sorry
15:53:09 <zz_ja> +1
15:53:12 <mmedvede> +1
15:53:13 <krtaylor> #startvote agree on initial cores for new repo
15:53:16 <asselin> +1
15:53:24 <patrickeast> +1
15:53:32 <krtaylor> #startvote agree on initial cores for new repo
15:53:32 <zz_ja> +1
15:53:33 <openstack> Unable to parse vote topic and options.
15:53:34 <marcusvrn> #vote yes
15:53:37 <marcusvrn> :P
15:53:41 <marcusvrn> +1
15:53:45 <ctlaugh_> +!
15:53:46 <krtaylor> lol, its broke real good
15:53:54 <krtaylor> anyway
15:53:58 <krtaylor> +1 from me
15:54:11 <krtaylor> any against?
15:55:23 <krtaylor> last call for votes
15:55:38 <krtaylor> #agreed asselin, patrickeast, mmedvede, sweston, krtaylor as initial cores for third party ci tools repo
15:55:47 <krtaylor> ok, 5 mins
15:56:02 <krtaylor> #topic Monitoring dashboard status
15:56:14 <krtaylor> I am assuming sweston didn't join us
15:56:48 <krtaylor> ok, then I will contact Timothy Chavez, he is not addressing the resolution of the comment on the patch
15:57:08 <krtaylor> I believe that is all that is holding that up
15:57:42 <marcusvrn> just a question: patrickeast and asselin are from cinder, correct? what about  you, krtaylor and the other 2 folks?
15:58:12 <patrickeast> yep i’m from cinder
15:58:40 <krtaylor> marcusvrn, ctlaugh_ and I are cross project, but mainly focused on nova
15:58:55 <krtaylor> sorry if I answered for you ctlaugh_
15:59:00 <krtaylor> asselin, is it ok if we defer the downstream puppet status, since you gave an update at the office hours meeting on Monday?
15:59:09 <asselin> krtaylor, yes
15:59:12 <egon> question for #3 in the etherpad: In terms of getting operators involved, is the suggestion that operators use their own infra, common infra, or just common tools?
15:59:24 <krtaylor> ok, good, 'cause we are out of time :)
15:59:28 <marcusvrn> patrickeast: krtaylor nice! thanks
15:59:35 <egon> I'll comment in the etherpad...
15:59:43 <krtaylor> egon, yes please
15:59:59 <krtaylor> thanks everyone, really good meeting as usual
16:00:38 <krtaylor> #endmeeting