22:02:10 <jeblair> #startmeeting zuul
22:02:11 <dmsimard> I don't have any particular topic in mind this week.
22:02:12 <openstack> Meeting started Mon Dec  4 22:02:10 2017 UTC and is due to finish in 60 minutes.  The chair is jeblair. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:02:13 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
22:02:15 <openstack> The meeting name has been set to 'zuul'
22:02:20 <jeblair> #link agenda https://wiki.openstack.org/wiki/Meetings/Zuul
22:03:04 <jeblair> #link previous meeting http://eavesdrop.openstack.org/meetings/zuul/2017/zuul.2017-11-27-22.06.html
22:03:18 <jeblair> #topic action items from previous meeting
22:03:19 * fungi wants to chat about zuul too, just had evening chores which ran a couple minutes over
22:03:21 <mrhillsman> bah, i'm not on that ML :(
22:03:22 <mordred> o/
22:03:54 <jeblair> mrhillsman: i have something to add about that in just a minute :)
22:03:59 <jeblair> jeblair start email thread about merging into master
22:04:12 <mrhillsman> cool
22:04:21 <jeblair> i did that and mmedvede has volunteered to help get puppet-openstackci ready for this
22:04:26 * Shrews is around-sih
22:04:28 <Shrews> ish
22:04:50 <jeblair> there's a first patch already up -- needs some revision.  hopefully we'll have that ready soon, then we can announce that we're going to do it and then do it.  (!!!)
22:04:56 <jeblair> clarkb fix https://review.openstack.org/516740 to call pip directly
22:04:59 <mordred> \o/
22:05:06 <rcarrillocruz> hola
22:05:11 <jeblair> i believe that's ready for review now
22:05:12 <clarkb> jeblair: that hapepned after much job debugging, was fun
22:05:17 <clarkb> ya should be ready
22:05:28 <jeblair> that'll get ara updated on openstack's install
22:05:31 <jeblair> clarkb: thanks!
22:05:49 <jeblair> #topic Roadmap
22:05:50 <clarkb> fungi should scrutinize it again as the original scrutinizer
22:06:13 * fungi is "the scrutinizer"
22:06:15 <jeblair> we talked about this last week, but i'll pretend i didn't just forget to remove it from the agenda
22:06:16 <fungi> coming this fall
22:06:42 <jeblair> #link https://storyboard.openstack.org/#!/board/53 now encodes the roadmap in storyboard
22:06:51 <jeblair> we should update our docs / readme to point to that
22:07:38 <pabelanger> I have a few patchs up for command socket work, but will finish rebasing them today
22:07:44 <jeblair> tristanC brought up that we should add the static nodepool driver to the release 3.0 list -- it's mostly written and will improve the experience for new users (simplifying what they need to do with nodepool before using zuul)
22:07:57 <fungi> seems like a reasonable exception
22:08:01 <dmsimard> +1
22:08:03 <rcarrillocruz> yes
22:08:19 <jeblair> cool; and there was some support on the ml too, so i'll add that.
22:08:22 <rcarrillocruz> and for non-openstack shops, is a great first step
22:08:36 <mrhillsman> ++
22:08:53 <jeblair> also, i realized that we should add 'get releases of github3 and gitpython' to the list -- those deps are currently pointing at unreleased git repos
22:08:59 <Shrews> the finger multiplexor/gateway thing doesn't seem to be captured there. but i'm on it
22:09:20 <pabelanger> +1 to deps
22:09:35 <jeblair> Shrews: cool.  i'd like to add that too, for much the same reason -- executors starting with root and dropping privs has been a major roadblock for new users, so it'll be great to get that out of the critical path.
22:09:48 <clarkb> jeblair: any idea if upsteram will be receptive to that? eg do we just ask nicely?
22:10:02 <jeblair> clarkb: they both have different issues
22:10:24 <jeblair> afaik, all we need from github3 is a release.  so we should ask nicely if we haven't.  let's find out from jlk and spamaps if they've tried that.
22:11:14 <jeblair> for git-python, basically they merged some major performance regressions.  i submitted a pr to mitigate one of them, and i think there's been a release, so first step there is to verify that and update the dep and see if our unit tests slow down.
22:11:27 <jeblair> if they still do, there may still be more work to do.
22:11:53 <clarkb> gotcha
22:12:14 <jeblair> (and in that case, we may need to evaluate whether we should continue throwing fixes at it, or look at switching to a different project that more aligns to what we need)
22:12:21 <jeblair> (because git performance in zuul is... critical.)
22:13:09 <jeblair> if those things all sound reasonable to folks here, i'll send an ml post about them
22:14:10 <clarkb> wfm
22:14:30 <jeblair> i'm going to abuse my chairpersonship and insert a topic not on agenda now
22:14:41 <jeblair> #topic zuul project hosting
22:15:00 <jeblair> i've posted this infra-spec: https://review.openstack.org/524024
22:15:29 <jeblair> it's a general purpose spec that will apply to other projects, but i also want to take advantage of it for zuul
22:16:13 <jeblair> in conversations with folks, i think there's widespread support for the idea that it'll be easier to communicate about what zuul is if it isn't buried in the middle of openstack/openstack-infra
22:16:43 <jeblair> in particular, i'd like for zuul to have its own mailing list, to help separate traffic from openstack-infra / third-party-ci / etc that we have on the openstack-infra list
22:16:53 * jeblair waves at mrhillsman
22:17:02 <fungi> and... we already have a non-confusing irc channel!
22:17:21 <jeblair> fungi: indeed!  i think that's proven very useful
22:17:22 <mrhillsman> nice
22:17:45 <dmsimard> +1, there's often negative perception/stigma when people from outside the larger community see "OpenStack" associated with something.. experienced it first hand with ara :(
22:17:51 * mrhillsman puts on his glasses
22:18:10 <jeblair> i bring that up here because i'd like to get at least two things related to that in use before we release 3.0:  mailing list, and git hosting aliases
22:18:57 <fungi> on the mailing list front, we've already booted up a (temporary, until we can merge them all into a single server with different virtual domains) mailman server for another similar sort of non-openstack community/effort under the openstack foundation umbrella
22:19:01 <jeblair> i think that's important for the 3.0 release so that in conjunction with that announcement, we can tell people "here is the mailing list to use" and "here is the location of the zuul-jobs repo we think you will want to add to your zuul" and neither of those will have openstack in the name, nor will have to change immediately after the release.
22:19:11 <fungi> we can pretty easily do the same for the zuul ci suite effort
22:19:48 <mrhillsman> awesome gents
22:19:57 <jeblair> fungi: ya.  i'd like to take a stab at just getting the multi-domain-on-one-server hosting that we would probably prefer done real quick.  if that drags on, i might advocate for a new, temp server.
22:20:09 <fungi> sgtm
22:20:18 <dmsimard> Well darn, the github zuul namespace is taken https://github.com/zuul
22:20:25 <clarkb> dmsimard: netflix right?
22:20:25 <jeblair> yeah, we could ask him nicely.  :)
22:20:30 <dmsimard> clarkb: some random person
22:20:31 <jeblair> nope, an individual
22:20:34 <clarkb> oh as a root name
22:20:47 <mrhillsman> there is only one repo, and it is old
22:20:51 <mrhillsman> unless there are private ones
22:21:05 <clarkb> there are already other name collisions though
22:21:06 <jeblair> i've registered zuul-ci.org for this purpose
22:21:14 <mrhillsman> i have reached out to github before and took over an account
22:21:30 <mrhillsman> well, got the name transferred
22:21:35 <dmsimard> Worth a try I guess
22:21:39 <mrhillsman> it is pretty simple to do and worth a try
22:21:40 <jeblair> obviously zuul.com/net/org are already taken :|
22:21:55 <jeblair> zuul-ci.org seemed like a reasonable compromise
22:21:57 <dmsimard> zuul.io !? :D
22:22:01 <mrhillsman> ^
22:22:01 <dmsimard> (probably taken too)
22:22:11 <jeblair> yes it is
22:22:12 <mrhillsman> ffs
22:22:13 <mrhillsman> yes
22:22:23 <pabelanger> so, if we are moving zuul, into its own domain. Have we consider renaming zuul to avoid the whole netflix, ghostbusters, naming issues?
22:22:27 <dmsimard> Good luck trying to find a 4 letter domain name that isn't taken :(
22:22:39 <SamYaple> zuul.io is pointed to rackspace vm somewhere lol
22:22:57 <mrhillsman> ^
22:22:59 <fungi> on the gh org name front, other than potential confusion there's probably not a strong need to mirror it to gh anyway right?
22:23:05 <fungi> or do we need it for hosting the app?
22:23:05 <mrhillsman> possibly can get :)
22:23:40 <jeblair> fungi: not sure.  i grabbed zuul-ci/ on github in case we decide we need something like that.
22:23:47 <fungi> wfm
22:23:51 <jeblair> but i don't have immediate plans to use it.
22:23:52 <mrhillsman> i think it would be mirrored for sure
22:24:21 <mrhillsman> but good to grab as you have jeblair just in case
22:24:29 <pabelanger> https://trademarks.justia.com/863/83/zuul-86383510.html
22:24:42 <pabelanger> quick google of zuul trademark, have no idea of that is legit
22:25:08 <dmsimard> huh, netflix would've trademarked it ?
22:25:17 <dmsimard> cause that says computer software, not ghostbusters :)
22:25:24 <mrhillsman> probably could discuss with board/staff on that front?
22:25:51 <dmsimard> might be worth just asking legal-discuss perhaps.
22:25:52 <mrhillsman> working out all those intricate details and just use what jeblair you have setup until it is required to change if at all?
22:26:49 <fungi> talking with lawyers... i think i have a root canal i need to schedule for... whenever that is
22:27:40 <fungi> i've heard lawyers say discussing trademark searches in public is risky business anyway
22:27:47 <mrhillsman> lol
22:28:03 <jeblair> yes, i don't think this is the right venue
22:28:21 <jeblair> does anyone else have any topics?
22:28:23 <fungi> because if you mention that you're aware of a mark the claimant can cite that as evidence of wilful infringement or whatever
22:29:17 * mrhillsman , "is no aware of anything"...default statement in 99.9% of situations :)
22:30:26 <ianw> wait, isn't this public?
22:30:32 <ianw> :)
22:30:39 <fungi> was more or less my point
22:31:04 <dmsimard> Not a Zuul-specific topic but an update regarding ARA.. There's progress with 1.0 and I will soon try my hand at running a zuul v3 integration job in ARA's gate
22:31:09 <clarkb> zookeeper cluster transition is happening
22:31:15 <jeblair> #topic open discussion
22:31:26 <clarkb> could provide new and exciting real world ops data for zuul
22:31:36 <clarkb> (not sure how much that will affect timing of operations for example)
22:31:43 <fungi> clarkb: i left you some feedback on the ara upgrading change. i think the regex insanity there can be substantially simplified, but overall looks fine to me
22:31:47 <dmsimard> new? exciting? zookeeper ?
22:32:14 <SamYaple> zuul.cloud is available....
22:32:18 <pabelanger> I'd like to ask for more reviews on https://review.openstack.org/521324/ to determine support for shared ansible_host in inventory
22:32:48 <dmsimard> SamYaple: I refuse to give in to all these weird TLDs :)
22:33:07 <dmsimard> I'm sure zuul.ninja is available but yeah
22:33:09 <clarkb> fungi: and responded. Thanks for the review
22:33:25 <fungi> dmsimard: we should get a .zuul tld then, obviously?
22:33:35 <SamYaple> dmsimard: you getting some old man syndrome going there? CHANGE OR DIE
22:33:42 <dmsimard> If you have a few 100k$ to spare
22:33:44 <SamYaple> fungi: com.zuul
22:34:00 <dmsimard> SamYaple: that sounds too java
22:34:27 <SamYaple> com.google existed asa mirror of google.com for a bit
22:34:30 <SamYaple> that was funny
22:34:49 <SamYaple> s/mirror/mirror image/
22:34:57 <dmsimard> Any other topics for the meeting ? :)
22:35:01 * SpamapS is dealing with priority issues, hopefully you can manage w/o me. :-/
22:35:31 <jeblair> SpamapS: i'll make sure there's an email about important stuff related to you
22:35:32 <pabelanger> I'd also like to finish off github.com testing this week, if possible. Right now we have a PR working on zuuv3.o.o, but want to confirm we are not leaking request limits or any other things we need to mark it off roadmap
22:35:46 <SpamapS> I see something about github3, was there something I could help with?
22:35:55 * SpamapS is waiting for a 5 minute process to complete... have some time to see here. ;)
22:35:59 <jlk> we need them to make a release
22:36:09 <jlk> or... if that doesnt' happen, perhaps identify what we need from the new release, and work around it?
22:36:12 <jeblair> SpamapS: we wondered if you or jlk have asked sigmavirus24 nicely for a release
22:36:21 <SpamapS> ah, k, no.
22:36:23 <SpamapS> not i
22:36:28 <jlk> We have a couple of times in the PR, but not outside of that
22:36:38 <jeblair> cool, so we should get a volunteer to send a nice email with chocolates
22:37:18 <jlk> I could do that I suppose.
22:37:41 <pabelanger> tip4commit
22:38:18 <jeblair> #action jlk ask sigmavirus24 about github3 release
22:38:20 <jeblair> jlk: thanks!
22:39:13 <mordred> pabelanger: I think we need to get jlk's move-gh-to-zuul-web patch finished up and rolled out- I think tristanC was doing some work on it
22:39:41 <fungi> jlk: hopefully you're able to expense the chocolates ;)
22:40:01 <pabelanger> mordred: okay
22:41:18 <mordred> pabelanger: mainly because I thnk that shift is going to be at least slightly disruptive - what with needing to swap rewrite rules as well ... I totally agree with the things you said
22:41:59 <jeblair> mordred, pabelanger: yeah.  i mean, ideally it shouldn't matter -- it should work the same both ways.  but it's a substantial change that could benefit from a close eye when we roll it out.
22:42:24 <rcarrillocruz> is it safe to assume that the connection/port plumbing from nodepool to zuul is a 3.0 deliverable? i.e. https://review.openstack.org/#/c/501976/
22:42:50 <ianw> "Puppet::Parser::AST::Resource failed with error ArgumentError: Invalid resource type vcsrepo " ... http://logs.openstack.org/48/525048/2/gate/legacy-infra-puppet-apply-3-ubuntu-trusty/887fbf6/openstack-infra/system-config/applytest/
22:42:54 <ianw> is this something we know about? ^
22:43:52 <fungi> missing the vcsrepo module?
22:44:35 <jeblair> rcarrillocruz: i don't think it's a v3.0 blocker, but my guess is that since it's mostly written and looks like something we can easily iterate on after the release without worrying too much about backwards compat, we'll probably be able to land it before release.
22:44:55 <rcarrillocruz> cool thx
22:47:13 <ianw> fungi: indeed ... the root cause is actually "fatal: unable to access 'https://git.openstack.org/openstack-infra/puppet-vcsrepo/': gnutls_handshake() failed: A TLS packet with unexpected length was received."
22:48:02 <jeblair> anything else, or should we wrap it up?
22:48:17 <fungi> i've got nuthin
22:48:41 <clarkb> EOF
22:49:10 <jeblair> thanks all!
22:49:12 <jeblair> #endmeeting