14:32:04 <jbair> #startmeeting Fedora CI SIG
14:32:04 <zodbot> Meeting started Wed Jul  1 14:32:04 2020 UTC.
14:32:04 <zodbot> This meeting is logged and archived in a public location.
14:32:04 <zodbot> The chair is jbair. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:32:04 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
14:32:04 <zodbot> The meeting name has been set to 'fedora_ci_sig'
14:32:11 <jbair> .hello jimbair
14:32:11 <zodbot> jbair: jimbair 'Jim Bair' <jbair@redhat.com>
14:32:32 <siddharthvipul> .hello siddharthvipul1
14:32:33 <zodbot> siddharthvipul: siddharthvipul1 'Vipul Siddharth' <siddharthvipul1@gmail.com>
14:32:46 <tflink> .hello2
14:32:47 <zodbot> tflink: tflink 'Tim Flink' <tflink@redhat.com>
14:32:51 <msrb> .hello2
14:32:53 <zodbot> msrb: msrb 'Michal Srb' <msrb@redhat.com>
14:34:02 <jbair> #topic Testing Farm API
14:34:10 <jbair> Per the doc
14:34:17 <jbair> #link https://etherpad.gnome.org/p/fedora-ci
14:34:17 <msrb> so
14:34:24 <jbair> * Testing Farm API prototype is deployed. It can run basic containerised tests. msrb is working on running rpmdeplint there
14:34:55 <msrb> there is a dev instance deployed here: api.dev.testing-farm.io
14:35:08 <msrb> it's super-fresh -- deployed today
14:35:19 <jbair> nice!
14:35:20 <msrb> we should be able to run rpmdeplint there
14:35:25 <msrb> but not installability
14:35:34 <msrb> (only containerized tests are supported atm)
14:35:36 <jbair> So testing to be done in the next few days I assume, time permitting?
14:35:54 <jbair> TIL rpmdeplint is containerized but installability is not :)
14:36:00 <msrb> yes, we will point our pipeline to that instance
14:36:22 <msrb> however, the results probably won't be production-ready yet
14:36:35 <msrb> there is not log streaming for example
14:36:42 <msrb> so people won't know what is going on
14:36:58 <jbair> can they see logs when completed but not in real time?
14:37:18 <msrb> I don't think so
14:37:32 <jbair> yeah, we need some kind of logs at a minimum - realtime like jenkins is nice for sure
14:37:36 <msrb> not even when completed (not 100% sure though)
14:37:38 <jbair> or an API to ping to get state possibly
14:37:51 <jbair> but okay, it sounds like we're testing and learning :)
14:37:58 <msrb> people will only get the final result: pass/fail/error, and xunit
14:38:41 <jbair> xunit?
14:38:46 <jbair> I feel like I should know what that means lol
14:38:51 <msrb> :)
14:39:22 <msrb> that xml file that Jenkins/dashboard and other CI/testing tools can display
14:39:30 <jbair> ah okay
14:39:40 <jbair> Thanks for the clarification :)
14:39:50 <jbair> okay, so testing can start now on API - sounds good - keep us updated!
14:39:55 <msrb> ;)
14:40:18 <jbair> anything else before we move on?
14:40:25 <msrb> not from my side
14:40:35 <jbair> okay onto msrb's next thing :)
14:40:50 <jbair> #topic CI onboarding docs
14:40:53 <jbair> #link https://pagure.io/fedora-ci/docs/pull-request/41#
14:41:26 <jbair> I gave this a quick read before this meeting and it seemed okay - I can read it top to bottom after this meeting if it needs some peer review :)
14:41:28 <msrb> the PR adds new "CI onboarding" section to the Fedora CI wiki
14:41:42 <msrb> jbair, thanks ;)
14:42:06 <jbair> I have a meeting after this one so after that I can read it and hit it with a LGTM (most likely)
14:42:46 <msrb> sounds good :)
14:42:56 <jbair> okay, so that moves us ti siddharthvipul's topic :)
14:43:04 <siddharthvipul> jbair: \o/
14:43:10 <jbair> #topic  Who owns `Testing farm` and `fedora packaging check` CI jobs
14:43:16 <siddharthvipul> so I have a list of say "vip" tenants.. and these 2 are in there
14:43:19 <siddharthvipul> so I wanted to clarify
14:43:38 <siddharthvipul> are you taking care of it as well (in Fedora CI namespace)? or need separately?
14:43:42 <siddharthvipul> cc: mvadkert
14:44:36 <msrb> I think mvadkert is on vacation already
14:45:10 <jbair> so this is one where I would ask bookwar but she's in another meeting (that I was also invited to, but I'm here with you instead) :)
14:45:19 <jbair> msrb: I'm not sure if you know? If not, we can ask after
14:45:27 <siddharthvipul> msrb: jbair.. I see
14:45:45 <tflink> what's fedora packaging check?
14:45:45 <siddharthvipul> would you like to schedule a meeting where we can go over this? or a mail thread perhaps?
14:45:55 <siddharthvipul> tflink: oh, sorry, I had to cc you as well
14:45:56 <msrb> no idea to be honest. are we talking about some jobs on continuous-infra jenkins instance?
14:46:30 <jbair> the fedora packaging checks sound like rpminspect instance to me?
14:46:57 <siddharthvipul> jbair: yes
14:46:59 <siddharthvipul> http://fedora-build-checks.apps.ci.centos.org/
14:47:17 <tflink> I've been working on that, not sure if that answers your question or not
14:47:23 <jbair> The "who" is the CI group at large I think, but as for points of contact
14:47:34 <jbair> tflink for that one, and mvadkert (when not on PTO) for Testing Farm
14:47:57 <siddharthvipul> actually, more than who, I wanted to know what can I do to start the migration process for these instances as well
14:48:13 <siddharthvipul> like give you a testing playground? or will this be included in the existing namespace?
14:49:01 * tflink thought that the new cluster was in more of a testing phase
14:49:02 <jbair> I assume we would want them within the fedora-ci project unless there's a reason to give each app it's own project
14:49:14 <jbair> tflink: I think they announced GA for it this past week
14:49:24 <tflink> having multiple jenkins daemons in the same projects can be fun
14:49:25 <jbair> so now it's open for general testing and migrations
14:49:37 <jbair> tflink: how so? causes some issues?
14:49:39 <tflink> daemons/processes/users whatever you want to call them
14:50:00 <tflink> there are a bunch of names that need to be changed if you want them to co-exist
14:50:01 <jbair> aren't they in different pods? maybe templates with the same named resources perhaps?
14:50:08 <jbair> ah
14:50:26 <tflink> it's not impossible, as much of a PITA than anything else
14:50:35 <siddharthvipul> tflink: the new cluster is ready to be consumed.. fedora CI folks (you all, maybe me included are testing to make sure we don't have hiccups when doing the actual migration
14:51:06 <jbair> I know astepano did some testing recently but he's in said meeting so I can't bother him now :)
14:51:12 <tflink> I thought that the projects weren't going to be migrated, that they would need to be moved/ported
14:51:24 <siddharthvipul> jbair: tflink I don't mind creating separate namespaces if it makes things easier
14:51:55 * tflink shrugs, either way
14:51:59 <msrb> siddharthvipul, I think having separate ns for at least testing farm and fedora ci stuff would be great
14:52:17 <siddharthvipul> msrb: sure thing.. so will the admins be same? or different group?
14:52:22 <tflink> I have a difficult time justifying putting much effort into the rpminspect stuff since it's due to be replaced soon
14:52:57 <tflink> I figure just keep it separate to make the porting easy, delete it all when testing farm takes over
14:53:01 <msrb> siddharthvipul, question for mvadkert, which is not available today :)
14:53:27 <siddharthvipul> msrb: I see.. will drop him an email then :)
14:53:50 <msrb> ;)
14:54:17 <siddharthvipul> I just want to make sure that you have all the things needed
14:54:28 <jbair> yeah, the rpminspect instance future is uncertain, but giving tflink his own space to try and deploy it into seems good to me
14:54:31 <siddharthvipul> and important tenants are given time to test things and migrate in comfort
14:54:32 <jbair> if we keep it, great, if not, we tear it down
14:54:57 <tflink> jbair: where's the uncertainty? I thought that everything was decided and we were just waiting for testing farm to be ready
14:55:25 <jbair> tflink: if I've learned anything in the last year, is that everything is uncertain :)
14:55:32 <jbair> that's more of a jbair thought moreso than any decision I've heard
14:55:41 <siddharthvipul> tflink: we will have the older cluster for atleast 3 months.. if testing farm will be ready to take over by then, we can avoid moving
14:55:42 <siddharthvipul> ?
14:55:44 <jbair> it's running today and useful so good to keep around until TFT can go to go
14:55:48 <tflink> fair enough, I've pretty much given up on pretending I know what's going on
14:55:51 <jbair> can go to town**
14:56:13 <jbair> siddharthvipul: is there a tentative decommission/shutdown date for the old cluster?
14:56:35 <jbair> Sounds like....Oct?
14:56:36 <tflink> the last I heard about testing farm is that they're running behind schedule and it'll be at least october until they're ready to take over rpminspect etc. for production in fedora
14:56:50 <siddharthvipul> jbair: just as I had mentioned earlier.. it's not a hard timeline but we would like to get rid of old cluster by the end of this quarter
14:57:10 <jbair> luckily you're saying this july 1st and not yesterday :D
14:57:34 <msrb> tflink, they are running behind schedule, that I can confirm. but I haven't heard anything about October :)
14:58:26 <tflink> the original date I was told was Flock 2020 which I translated to august. the last I heard, they were running ~2 months behind schedule which I interpret as october-ish
14:58:42 <jbair> siddharthvipul: worst case, I think we can get a good feeling by the next meeting as miro should be back by then
14:58:47 <jbair> tflink: ah, nice detective work :)
14:58:52 <tflink> but things change all the time, I haven't the foggiest idea if any of that is still valid
14:59:13 <jbair> tflink: "things change all the time" which is why I mentioned if there's a need to keep that jenkins around ;) You never know!
14:59:15 <msrb> tflink, I'd bet on the flock date personally
14:59:34 <msrb> October sounds pretty late
14:59:43 <tflink> I'm just going off of what I was told
15:00:20 <siddharthvipul> hmm, I kind of want to propose a meeting with all folks involved but have a feeling that won't exactly solve this issue
15:00:27 <siddharthvipul> so we wait and see?
15:00:41 <jbair> siddharthvipul: I would say by next CI meeting we should know at latest
15:00:42 <tflink> I'm planning to keep the rpminspect stuff running until I'm told to turn it off
15:00:43 <jbair> since miro is out next week
15:00:52 <tflink> and that includes porting to the new cluster
15:01:04 <jbair> I'll make sure it's on the agenda for next time :)
15:01:18 <tflink> in my experience, nothing happens when the original schedule says it will and it's wise to add months to any deadline
15:01:29 <tflink> not just in this case - in general
15:02:15 <siddharthvipul> tflink: that's fair
15:02:15 <tflink> "plan for the worst, hope for the best" :)
15:02:27 <siddharthvipul> let me know if you need a namespace and I will so the very same
15:02:36 <siddharthvipul> if we want to wait for next meeting, that works to
15:02:38 <siddharthvipul> to*
15:02:41 <siddharthvipul> too* argh
15:02:42 <msrb> +1
15:03:21 <tflink> siddharthvipul: how do I start the request? ticket on bugs.centos.org
15:03:23 <tflink> }
15:03:25 <tflink> ?
15:04:14 <jbair> siddharthvipul: I say we make a namespace
15:04:19 <siddharthvipul> tflink: just ping me :)
15:04:23 <siddharthvipul> I guess you did already
15:04:36 <siddharthvipul> can you send me the admin's ACO email addresses
15:04:39 <siddharthvipul> and namespace name
15:04:40 <siddharthvipul> that's all
15:05:07 * tflink will send email
15:05:13 <siddharthvipul> Thank you :)
15:07:59 <tflink> jbair: what's your ACO email? does it match your FAS email?
15:08:18 <jbair> tflink: jbair@redhat.com
15:08:22 <jbair> though we can use yours too for the project
15:08:26 <jbair> my ACO is jbair
15:09:10 <jbair> also I am wrapping this meeting but we can keep talking :)
15:09:14 <jbair> #endmeeting