20:00:33 <sgallagh> #startmeeting Server Working Group Weekly Meeting (2016-09-27)
20:00:33 <zodbot> Meeting started Tue Sep 27 20:00:33 2016 UTC.  The chair is sgallagh. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:33 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
20:00:33 <zodbot> The meeting name has been set to 'server_working_group_weekly_meeting_(2016-09-27)'
20:00:33 <sgallagh> #meetingname ServerSIG
20:00:33 <zodbot> The meeting name has been set to 'serversig'
20:00:33 <sgallagh> #chair nirik adamw mhayden jds2001 mjwolf sgallagh dperpeet smooge vvaldez
20:00:33 <zodbot> Current chairs: adamw dperpeet jds2001 mhayden mjwolf nirik sgallagh smooge vvaldez
20:00:33 <sgallagh> #topic Roll Call
20:00:33 <sgallagh> .hello sgallagh
20:00:34 <zodbot> sgallagh: sgallagh 'Stephen Gallagher' <sgallagh@redhat.com>
20:00:43 <mhayden> .hello mhayden
20:00:44 <zodbot> mhayden: mhayden 'Major Hayden' <major@mhtx.net>
20:00:45 <smooge> .hello smooge
20:00:47 <zodbot> smooge: smooge 'Stephen J Smoogen' <smooge@gmail.com>
20:01:12 <adamw> hi
20:01:19 <adamw> .hello adamwill
20:01:20 <zodbot> adamw: adamwill 'Adam Williamson' <awilliam@redhat.com>
20:01:38 <nirik> .hello kevin
20:01:39 <zodbot> nirik: kevin 'Kevin Fenzi' <kevin@scrye.com>
20:01:53 <jds2001> .hello jstanley
20:01:54 <zodbot> jds2001: jstanley 'Jon Stanley' <jonstanley@gmail.com>
20:02:37 <jds2001> multitasking - someone scheduled a conference call even when they asked me "how flexible is your 4pm?" and I said "not" :/
20:02:45 <sgallagh> jds2001: Naturally
20:03:19 <sgallagh> dperpeet is having trouble connecting to Freenode
20:03:47 <nirik> freenode is not having a fun day today
20:04:02 <sgallagh> nirik: Got a known-good mirror I can point him to?
20:04:20 <nirik> no idea what server(s) are working currently. ;)
20:05:27 <cmurf> it's a dos attack according to their twitter
20:05:34 <sgallagh> OK, he can't get in, so he's not going to make it today.
20:05:38 <sgallagh> Unfortunate
20:06:21 <sgallagh> OK, we have quorum at least (though barely more than minimal). I guess we'll get started
20:06:56 <sgallagh> #topic Agenda
20:07:09 <sgallagh> #info Agenda Item: Cloud/Server positioning
20:07:17 <sgallagh> #info Review the Kellogg Logic Model and discuss prioritization/assignment
20:07:26 <sgallagh> Other topics to discuss?
20:07:31 <sgallagh> Oh, actually...
20:07:37 <sgallagh> #topic Beta Freeze and Testing
20:07:42 <sgallagh> #undo
20:07:42 <zodbot> Removing item from minutes: <MeetBot.items.Topic object at 0x3563b790>
20:07:48 <sgallagh> #info Agenda Item: Beta Freeze and Testing
20:08:00 <sgallagh> Maybe we should hit that one first.
20:08:21 <sgallagh> Any other topics?
20:09:00 <sgallagh> (Someone please reply, as since I'm on a bouncer, I have no way to tell if stuff is getting through)
20:09:45 <nirik> yep. sounds fine
20:09:48 <adamw> i'm listening
20:10:09 <sgallagh> OK, let's do that one first
20:10:14 <sgallagh> #topic Beta Freeze and Testing
20:10:23 <adamw> everything's working fine in openqa atm
20:10:31 <adamw> and the only tests openqa doesn't hit are the AD ones
20:10:54 <adamw> it'd be great for people to 'check my work', btw - review the openqa tests, make sure they're actually testing everything we should be testing...i wrote 'em off the manual test cases
20:11:09 <sgallagh> adamw: Did you have a chance to re-test the AD domain controller I have set up?
20:11:09 <adamw> you can watch videos of the tests and i can point you at the code too if you like
20:11:14 <adamw> sgallagh: not yet, sorry
20:11:43 <sgallagh> adamw: Do you recall who the other folks were that were volunteering to do that manual testing?
20:11:56 <sgallagh> Ideally I'd like to avoid either of us getting stuck with it yet again if it can be helped
20:12:27 <adamw> er....no. should be in the minutes, right?
20:12:39 <sgallagh> adamw: No, that discussion happened in #fedora-qa
20:12:42 <nirik> There was someone who had a AD setup to test with...
20:12:46 <nirik> but I cannot recall who now
20:12:49 <adamw> oh right. well, i'll have logs
20:12:53 * adamw logs everything
20:13:04 <smooge> adamw, as I am wanting to make sure.. where do I look at the openqa work and check its work?
20:13:09 <sgallagh> #action adamw to figure out who volunteered to test AD support and politely encourage them to do so
20:13:31 <adamw> smooge: welp, go to http://openqa.fedoraproject.org/
20:13:43 <adamw> click on the latest f25 nightly link, so https://openqa.fedoraproject.org/tests/overview?distri=fedora&version=25&build=Fedora-25-20160927.n.0&groupid=1 right now
20:13:58 <adamw> you'll find all the server tests under 'Server-dvd-iso'
20:14:13 <adamw> for each test, click on the colored blob (green/red/orange, green is pass, red is fail, orange is softfail) to see test details
20:14:34 <nirik> It was puffi who was able to test alpha against AD.
20:14:39 <smooge> and look for the lines like server_realmd_join_kickstart 	- 	_boot_to_anaconda
20:14:48 <adamw> then you can see thumbnails of screenshots taken throughout the test process, on the 'Logs & Assets' tab you can download a (somewhat frame skipped) video of the test
20:15:25 <adamw> and you can click the test module names (that's what _boot_to_anaconda etc. are) and read the actual code for the test, though you might need me to help you know exactly what they mean (and what some of the functions they get from libraries do)
20:15:49 <smooge> An unknown error has occurred. :) I love that error
20:16:19 <adamw> smooge: ah, there's a funny thing right now...i actually just pushed a change which kinda re-arranged some of the test code
20:16:31 <adamw> so some of the test modules used for that last nightly run don't exist any more under the same names
20:16:40 <adamw> sorry about that :) after the next nightly it'll be ok.
20:16:43 <smooge> no problem. I can look at the next run
20:16:49 <smooge> now I know where to look
20:16:54 <sgallagh> Thanks adamw, smooge
20:17:17 <adamw> smooge: you can check out the entire test repo from https://bitbucket.org/rajcze/openqa_fedora
20:17:45 <adamw> smooge: you'll find the fedora-specific shared functions in the files in lib/ in that checkout; openqa upstream standard ones are in https://github.com/os-autoinst/os-autoinst , mainly in testapi.pm
20:18:49 <sgallagh> #info adamw requests that anyone available should review the openQA tests for Server and confirm they are testing a right and sufficient amount of the required functionality
20:19:33 <sgallagh> OK, shall we discuss the Kellogg Logic Model updates next, the Cloud/Server positioning or roll them together?
20:20:14 <jds2001> sgallagh: i can test with AD....reading scrollback
20:20:34 <sgallagh> jds2001: That would be very helpful
20:20:54 <sgallagh> Keep an eye out for one of adamw's automated "Please test this compose" messages and... test that compose, please :)
20:21:38 <sgallagh> So, next topic?
20:22:43 <cmurf> uhoh
20:23:12 <sgallagh> Quick, assign everything to adamw`!
20:23:32 <adamw> jds2001: current test compose link: https://fedoraproject.org/wiki/Test_Results:Current_Server_Test
20:23:53 <adamw> sure, go ahead, that guy deserves it all ;)
20:23:58 <sgallagh> heh
20:24:14 <sgallagh> #topic Kellogg Logic Model Discussion
20:24:29 <jds2001> adamw: do you care about install method? typically it's virt-install for me....with or without kickstart.
20:24:45 <jds2001> i.e. no media involved.....virtual or otherwise :)
20:25:10 <sgallagh> jds2001: Some of the tests are kickstart tests, so you will need to use media for those at least
20:25:55 <sgallagh> I'm trying to get a link to the logic model I created, but Heroku is being super slow
20:25:58 * mhayden will fire up some media on some servers :)
20:26:16 <sgallagh> (I'll be moving the server over to fedora infra this week, but I haven't yet had time)
20:26:29 <adamw> jds2001: there's basically three enrolment methods to test, ideally we'd test all three with both AD and freeipa, but we can kinda fudge a bit if necessary, it's quite unlikely that kickstart enrol with freeipa would work, and post-install enrol with ad would work, but kickstart enrol with ad would fail, for e.g.
20:26:32 <sgallagh> #link http://kolinahr.herokuapp.com/edit/57d05a6984338834000515c9
20:27:24 <sgallagh> adamw, jds2001: Mind if we take that outside the meeting? It's in the weeds.
20:27:38 <sgallagh> I have a hard stop in 33 minutes today
20:27:45 <adamw> sure
20:27:52 <adamw> was just filling time till you got your logic models
20:28:07 <sgallagh> adamw: Yeah, the server is fired up now, so it should be visible
20:28:22 <sgallagh> At the moment, it requires Google account creds to sign in, but that will be FAS when I move it over.
20:28:40 <sgallagh> (A piece of software so new, it doesn't yet have spam bots created for it!)
20:28:48 * nirik is looking now.
20:28:52 <nirik> this is pretty slick
20:29:40 <sgallagh> Yeah, it's fairly impressive for about three weeks' work from a single person.
20:30:13 <sgallagh> I didn't color in the ones I added since last week, partly so it would be obvious what was new
20:30:17 <sgallagh> but mostly because I didn't have the time
20:30:24 <jds2001> so i've looked at this before and gotten very confused. For example, on the Inputs in PXE
20:30:39 <jds2001> we don't need PXE to *do* anything for this activity=
20:30:58 <jds2001> but another input is Fedora Infrastructure
20:31:05 <sgallagh> jds2001: I was interpreting the inputs as "projects that will contribute to this activity"
20:31:05 <jds2001> and there is a speicific ask of them.
20:31:10 <sgallagh> Even if not directly
20:31:25 * nirik doesn't see pxe...
20:31:35 <nirik> oh, there it is
20:31:36 <sgallagh> nirik: Preboot eXecution Environment
20:31:37 <jds2001> fourth from the bottom :)
20:31:56 <nirik> I can't see where it's arrow is going really.
20:32:15 <sgallagh> nirik: You can always click on it to see a detail view
20:32:26 <sgallagh> Just don't click the hyperlink part
20:32:30 <sgallagh> Or that will take you... elsewhere
20:32:38 <nirik> it seems to take me to wikipedia
20:32:44 <smooge> can we drop this model and focus on the Enterprise 5 year onbe?
20:32:44 <smooge> seriously though I found this very useful
20:33:17 <sgallagh> nirik: Maybe I should have omitted the hyperlinks. I can see how that might be confusing.
20:33:28 <nirik> no biggie. I got it.
20:33:29 <sgallagh> Just click on the box anywhere but the hyperlinked text and you'll get the detail view
20:34:56 <sgallagh> So to tie this in with cmurf's request, we should talk about whether there's an Outcome that feeds our Mission and/or Vision that would be helped by the Cloud stuff.
20:35:10 <smooge> so I realized to put in say "SMTP mail server" as an output I need to make an outcome which matches that which ties into a ...
20:35:47 <sgallagh> smooge: Right, the idea is that in order to add an Output, you have to justify its value by adding an Outcome that serves the Mission
20:35:52 <smooge> sgallagh, so the cloud items fit the Vison/Mission easily
20:36:01 <sgallagh> Or tie it to an existing Outcome, of course
20:36:56 <sgallagh> (Aside: Kolinahr doesn't have live-updating or concurrent editing yet, so if anyone makes edits, let us know to refresh after you hit save)
20:37:28 <smooge> I am going to hold off on any edits til after this
20:37:33 <sgallagh> ack
20:37:45 * nirik too.
20:38:11 <sgallagh> smooge: OK, so you started to say "the cloud items fit the Vison/Mission easily". Want to take a stab at an Outcome?
20:39:17 <smooge> Users of Fedora Server Cloud will have a small base set of packages which they can grow their infrastructure from?
20:39:35 <cmurf> Well my take is Cloud WG may not even remain Cloud WG, may become Container WG or Atomic WG, or get rolled into to Server - that's all unclear. But the existing Cloud WG members are primarily interested in focusing on atomic host stuff.
20:39:46 <smooge> since pretty much the cloud request is "make a small install.. really small."
20:40:30 <cmurf> It sounded to me like cloud folks were asking if Server WG would take on cloud non-atomic deliverables. Is that... correct?
20:40:36 <sgallagh> smooge: I'd argue there are two reasons for that, neither of which is clearly mentioned in that sentence
20:41:16 <sgallagh> I think one of them is "Users of the Fedora Server Edition have a very limited attack surface" and the other is "Users of Fedora Server will consume the smallest possible amount of storage for the operating system"
20:41:28 <sgallagh> One of those is about security/manageability and the other is cost-savings
20:41:48 <sgallagh> cmurf: That is definitely on the table
20:42:08 <sgallagh> cmurf: However, we asserted previously that we *won't* be taking on a "super-minimal grow-your-own-os" image.
20:42:19 <sgallagh> If we take on the AMIs and such, they'll be Fedora Server based
20:42:30 <sgallagh> Though we *are* planning to trim Server down more than we have in the past.
20:43:21 <sgallagh> We haven't talked at all about a container base image, though.
20:43:45 <sgallagh> That may in fact be owned by the Modularity WG and/or the Base Runtime team and not be technically a Server Edition offering
20:45:00 <sgallagh> Am I talking to myself again, or is Freenode fritzing?
20:45:14 <cmurf> Ok so maybe there needs to be a list of questions, as I have a lot more questions than answers.
20:45:22 <nirik> it is, but I'm still here...
20:45:22 <jds2001> i can see you and only you
20:45:52 * nirik has a personal interest in openstack images... minimal or server based either one is fine.
20:46:02 <sgallagh> jds2001: I'm flattered, but I'm going to have to politely turn you down ;-)
20:46:09 <ssmoogen> ugh sorry how much did I miss?
20:46:52 <sgallagh> Catch-up: https://paste.fedoraproject.org/436757/09191147/
20:46:53 <cmurf> And then that big list can just flushed out on devel@, and see if it's something Modularity WG is going to take on in all or part.
20:46:58 <sgallagh> (since the last time smooge spoke)
20:47:01 <jds2001> sgallagh: lol
20:47:26 <sgallagh> nirik: Yes, OpenStack images should absolutely be on the table
20:47:58 <sgallagh> cmurf: Could you please itemize those questions and send them out? I'll answer whatever I can on devel@
20:48:18 * nirik knows he's at least one paradigm shift behind, but such is the way it goes. ;)
20:48:31 <sgallagh> nirik, smooge: I'd be very appreciative if you could put your thoughts into Kolinahr RE: openstack, SMTP, AMIs, etc.
20:48:47 <nirik> can try
20:49:03 <sgallagh> OK, if you're more comfortable sending them to server@, I'll handle updating the model
20:49:49 <sgallagh> Ideally, I'd like for us to lock down the things we want to do by next week, after which I'll write a draft of the new PRD based on it and send it out for comment
20:50:11 <nirik> and all this is starting with f26 right?
20:51:16 <jds2001> the f25 ship has sailed, no????
20:51:39 * jds2001 thinks it's pretty far from the port :)
20:51:45 <nirik> yes. I think so
20:51:58 <sgallagh> nirik: Yes, absolutely this is all F26 work
20:52:16 <sgallagh> F25 is mostly a holding pattern with just standard FreeIPA and Cockpit updates
20:52:30 <cmurf> ship it!
20:52:35 <sgallagh> And probably a little trimming down as Base Runtime gets into gear and starts chopping stuff out of the low-level
20:53:30 <sgallagh> #action cmurf to start a discussion around concrete plans for cloud images et. al. on devel@
20:54:15 <sgallagh> #action nirik and smooge to email server@ with additional Activities, Outputs and Outcomes related to cloud images and SMTP.
20:54:45 <sgallagh> #info Next week we plan to finalize the logic model. sgallagh will then create a draft of a PRD from it
20:54:55 <sgallagh> Does that about sum it up?
20:54:57 <smooge> sgallagh, I will do so and look at the logs after this meeting as I keep getting bursts of traffic from my bouncer
20:55:09 <smooge> sorry this is like typing through 14.4 lag
20:55:18 <sgallagh> smooge: Understood. Thank you (everyone) for attending even through this mess.
20:55:56 <nirik> thanks sgallagh
20:56:21 <sgallagh> #topic Open Floor
20:56:33 <sgallagh> Five minutes for Open Floor, then I turn into a pumpkin
20:59:14 <sgallagh> OK, thanks for coming folks.
20:59:24 <sgallagh> Talk to you on the lists and in #fedora-server
20:59:35 <sgallagh> (If Freenode survives the day, anyway)
21:01:38 <nirik> sgallagh: endmeeting? or did you and I was split?
21:01:45 <sgallagh> #endmeeting