15:00:38 <andreasn> #startmeeting Cockpit public meeting 2014-10-06 15:00:38 <zodbot> Meeting started Mon Oct 6 15:00:38 2014 UTC. The chair is andreasn. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:38 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:45 <andreasn> #meetingname Cockpit 15:00:45 <zodbot> The meeting name has been set to 'cockpit' 15:00:53 <andreasn> #chair andreasn mvollmer stefw sgallagh 15:00:53 <zodbot> Current chairs: andreasn mvollmer sgallagh stefw 15:00:59 <andreasn> #topic Welcome 15:01:07 <andreasn> .hellomynameis andreasn 15:01:08 <zodbot> andreasn: andreasn 'Andreas Nilsson' <anilsson@redhat.com> 15:01:15 <mvollmer> .hellomyname is mvo 15:01:19 * stefw is here 15:01:20 <mvollmer> .hellomynameis mvo 15:01:21 <zodbot> mvollmer: mvo 'Marius Vollmer' <marius.vollmer@gmail.com> 15:01:34 <andreasn> .hellomynameis stefw 15:01:35 <zodbot> andreasn: stefw 'Stef Walter' <stefw@redhat.com> 15:01:45 <mvollmer> heh! :-) 15:01:55 <andreasn> #topic Agenda 15:02:20 <andreasn> * as mvollmer said, Fedora 21 15:02:22 <mvollmer> Fedora 21, GSSAPI, Modular arch, roadmap? 15:02:39 <andreasn> yeah 15:03:06 <andreasn> #info agenda is Fedora 21, GSSAPI, Modular arch, Roadmap 15:03:17 <andreasn> #topic Fedora 21 15:03:56 <mvollmer> I am running out of things to fix, stef is merging like a hero. 15:04:08 <stefw> still have a ways to go ... 15:04:11 <stefw> on merging 15:04:31 <mvollmer> yes, and there will be things to fix about the pull requests 15:04:42 <andreasn> there is quite a bunch of pull requests 15:04:55 <mvollmer> so, question: are we forgetting something important? 15:05:06 <andreasn> 4 of them are f21-proposed 15:05:23 <mvollmer> I would maybe look at #1176, and not much else 15:05:26 <stefw> sgallagh, do you have any feedback on anything missing for f21? 15:05:51 <mvollmer> andreasn, I think the pull request labels and milestones are not very consistent. 15:05:59 <andreasn> ah, ok 15:07:10 <mvollmer> what do we need to do outside of cockpit? 15:07:21 <mvollmer> NetworkManaer testing and release 15:07:34 <sgallagh> stefw: Sorry, wasn't following. 15:07:37 <sgallagh> One moment to read scrollback, please 15:07:40 <stefw> sure 15:09:16 <sgallagh> Ideally, I'd like the modular capabilities in sufficient shape that it might be feasible to add some domain-controller UI post-F21, but that's firmly a nice-to-have. 15:09:43 <mvollmer> sgallagh, how much time do we have? 15:10:06 <stefw> sgallagh, there's a slight chance these modular capabilities would land 15:10:09 <stefw> but more importantly 15:10:11 <stefw> they would not be stable 15:10:13 <mvollmer> sgallagh, I was thinking to wrap up Fedora 21 from our point of view in this week. 15:10:23 <stefw> and therefore any such domai-controller UI would be prototypical 15:10:28 <sgallagh> mvollmer: Oct 14th is Beta Freeze 15:10:38 <sgallagh> Like I said, firmly in the nice-to-have category. 15:10:39 <stefw> let me reword that 15:10:49 <stefw> there's a slight chance one branch towards the modular capabilities might land 15:11:04 <stefw> so unfortunately, although it would have been nice, we're likely going to have to skip on this 15:11:11 <mvollmer> sgallagh, so we'd need to have that by the end of this week, right? 15:11:52 <sgallagh> Yes 15:12:01 <mvollmer> right 15:12:24 <mvollmer> so, let's see. :-) 15:12:35 <mvollmer> everyone looks at stefw... 15:12:38 <stefw> heh 15:13:04 <andreasn> so if we can't land the modulization stuff for f21 anyway, might be better to focus on polish stuff 15:13:12 <andreasn> or start moving on to the next stage 15:13:42 <stefw> yup 15:14:27 <mvollmer> stefw, what do you think, first prio is merging/rejecting the pull requests we ave? 15:14:38 <stefw> yes 15:14:53 <mvollmer> yes, very good. 15:15:31 <mvollmer> then we make a release for F21 beta, and then we concentrate on the module stuff. 15:15:43 <andreasn> sounds good 15:16:09 <mvollmer> if we can make it happen without regressions, we can then see whether we feel comfortable about putting it into F21 stable. 15:16:16 <mvollmer> or just the copr. 15:16:43 <stefw> i would be uncomfortable merging even the first modular code branch into f21 15:16:57 <stefw> i would suggest we essentially stabilize f21 before we merge it 15:16:59 <stefw> even if we do have more time 15:17:01 <mvollmer> ok, much easier that way. 15:18:17 <andreasn> oh, right, the bot 15:18:43 <andreasn> #info no modularization work for f21 15:19:16 <andreasn> #info first priority is to merge the pull requests we have at https://github.com/cockpit-project/cockpit/pulls 15:19:33 <andreasn> cool 15:19:38 <andreasn> next topic? 15:19:40 <mvollmer> #info Cockpit release for Fedora 21 will be cut at the end of this week. 15:20:01 <andreasn> #topic GSSAPI 15:20:05 <mvollmer> #action everyone help stef with pull request review 15:20:19 <andreasn> sounds good 15:20:52 <mvollmer> I'll rebase everything and run VERIFY 15:21:13 <mvollmer> ok, GSSAPI? 15:21:45 <mvollmer> Can we advertise it for F21? 15:22:24 <andreasn> I have to admit I have no idea what gssapi is 15:22:29 <mvollmer> You can't log in as wheel with Kerberos, and you can't use really use Cockpit as non-hweel. 15:22:46 <andreasn> oh, is it the single-sign-on? 15:22:54 <stefw> i don't think we can advertise it 15:23:13 <stefw> but i like it being there since it lights a fire under people to actually fix the underlying platform bugs we have have listed. 15:23:20 <mvollmer> yes. 15:23:43 <stefw> sgallagh, does this affect our Fedora Server documentation or release notes? 15:23:55 <mvollmer> remoting via ssh doesn't work with SSO either, right? 15:23:56 <stefw> the fact that GSSAPI works, but actually doesn't (because no wheel)? 15:24:00 <andreasn> #info You can't log in as wheel with Kerberos, and you can't use really use Cockpit as non-hweel. 15:24:05 <stefw> mvollmer, yes, that branch has not been merged yet 15:24:13 <stefw> and needs rebasing, testing and fixing 15:24:21 <mvollmer> stefw, should we try to do that? 15:24:28 <stefw> yes, i think so 15:24:29 <mvollmer> to make the fire hotter? 15:24:33 <mvollmer> ok. 15:24:40 <stefw> it would be nice to get that into f21, again for the reason above 15:24:47 <stefw> so people can try out kerberos support 15:24:51 <sgallagh> stefw: We haven't made any statements about SSO at all 15:24:53 <andreasn> #info even though not everything works, it's a good motivator for others to fix underlying platform bugs 15:24:54 <stefw> and very easily reproduce the issues that don't work 15:24:57 <stefw> sgallagh, ok, good to know 15:25:02 <sgallagh> So I think continuing to not say anything is probably fine 15:25:05 <mvollmer> #info finish #1226 for Fedora 21. 15:25:58 <andreasn> is that a action for stefw or mvollmer, or both? 15:26:08 <mvollmer> both, I guess. 15:26:24 <mvollmer> tbh, I have ignored that pr mostly... :-/ 15:26:31 <andreasn> #action mvollmer and stefw to finish #1226 for Fedora 21 15:26:50 <andreasn> great, next topic? 15:27:25 * mvollmer nods 15:27:40 <andreasn> #topic Modular architecture 15:28:13 <mvollmer> so, as discussed, I'd say it has low priority for this week 15:28:18 <andreasn> sure 15:28:27 <mvollmer> stefw, do you agree? 15:28:41 <stefw> low priority for a merge 15:28:48 <mvollmer> yes 15:28:52 <stefw> but a high priority to get it in a reviewable and discussable state 15:29:00 <stefw> too many people are waiting for this 15:29:13 <stefw> and soon it will become a road block in our path forward 15:29:18 <mvollmer> yes, but this week we really want to get F21 done. :-) 15:29:36 <stefw> indeed 15:29:37 <andreasn> #info won't merge this week, since we're focusing on F21 15:29:53 <andreasn> #info really important path forward though 15:29:54 <stefw> so for me that means working on the gssapi delegation pullreq 15:29:57 <stefw> and doing reviews 15:29:59 <mvollmer> next week, full speed on modular stuff 15:30:24 <mvollmer> yes 15:30:26 <andreasn> #info next week will be focus on modular arch 15:30:30 <andreasn> great! 15:30:53 <andreasn> next topic? (hope I'm not rushing things) 15:31:07 <stefw> no this is good 15:31:15 <mvollmer> agree 15:31:18 <andreasn> #topic Roadmap 15:31:34 <mvollmer> (this should be a regular topic every week) 15:31:47 <mvollmer> (and means: update the roadmap wiki page) 15:31:50 <andreasn> #info first roadmap now up at https://github.com/cockpit-project/cockpit/wiki/Roadmap 15:32:41 <mvollmer> ok, next four weeks: 15:32:49 <mvollmer> let's remove the second bullet. 15:33:00 <andreasn> is mount points merged? 15:33:20 <mvollmer> no, but it's in good shape and doesn't need to be pointed out, I'd say. 15:33:39 <andreasn> so likley to be merged this week? 15:33:53 <jscotka> stefw, mvollmer andreasn Hi, sorry for late coming at the meeting 15:33:57 <mvollmer> let's add "SSO enablers" and move "Modular architecture" to 4w. 15:33:58 <stefw> no worries 15:34:35 <andreasn> #info mount points fix in good shape to be merged soon. Can be removed from roadmap 15:34:44 <jscotka> something important was introduced today from QA side? 15:34:48 <mvollmer> I'll just do it live 15:35:04 <andreasn> #info SSO enablers and Modular architecture can be moved from the section 4 months to 4 weeks 15:35:20 <mvollmer> jscotka, let's add a topic to todays agenda. 15:35:46 <mvollmer> we'll quickly finish the roadmap polish, and then we can talk about QA. 15:35:53 <andreasn> sounds good 15:36:03 <jscotka> thanks 15:36:20 <mvollmer> so, atomic. 15:36:32 <mvollmer> i think we should have a go at it soonish. 15:36:37 <andreasn> is Embedding still in the right place on the roadmap? 15:36:53 <mvollmer> not sure... 15:37:09 <mvollmer> is it tied to the navigation redesign? 15:37:19 <andreasn> I think so. stefw? 15:37:24 <stefw> it is indeed 15:37:40 <stefw> embedded will arrive along with modular work 15:37:44 <stefw> and the two need to be considered together 15:37:51 <andreasn> so also 4 weeks actually? 15:38:00 <mvollmer> right 15:38:05 <mvollmer> navigation also? 15:38:07 <stefw> yes, for a initial work 15:38:08 <andreasn> yeah 15:38:15 <stefw> it would be nice if navigation did 15:38:23 <stefw> it would make life much simpler 15:38:26 <stefw> but it could be decopled 15:38:32 <mvollmer> then "4 months" is empty. :-) 15:38:42 <andreasn> #info move navigation and embedability to 4 weeks 15:39:02 <andreasn> lets fill that from 2 years then :) 15:39:08 <mvollmer> sounds like we need a more detailed plan of attack. 15:39:35 <andreasn> I have at least two things we need at "some point in the future" 15:39:48 <andreasn> better timers UI: https://github.com/cockpit-project/cockpit/issues/1297 15:40:10 <andreasn> Firewall support: https://github.com/cockpit-project/cockpit/issues/1297 15:40:19 <mvollmer> yeah, 15:40:22 <mvollmer> I make links.. 15:40:45 <andreasn> we have a whole bunch of things under https://github.com/cockpit-project/cockpit/issues/1297 15:41:15 <andreasn> might be a good time to update patternfly for example after f21 15:41:41 <andreasn> ups, wrong link above https://github.com/cockpit-project/cockpit/labels/enhancement 15:41:45 <mvollmer> andreasn, is the link correct? it only talks about cron. 15:41:49 <mvollmer> i see 15:41:59 <andreasn> patternfly update: https://github.com/cockpit-project/cockpit/issues/788 15:42:59 <andreasn> firewall support: https://github.com/cockpit-project/cockpit/issues/1094 15:43:16 <mvollmer> andreasn, could you sort those into the roadmap offline? 15:43:22 <andreasn> yep! 15:43:34 <mvollmer> and then we discuss that next week, or ad-hoc. 15:43:36 <andreasn> I'll put them under 2 years for now 15:43:53 <mvollmer> i have put patternfly under 4m now. 15:44:01 <andreasn> #action andreas to sort issues from https://github.com/cockpit-project/cockpit/labels/enhancement on to the roadmap page 15:44:12 <andreasn> sounds correct 15:44:42 <andreasn> so server roles under 4 months? 15:44:48 <mvollmer> sounds like it, no? 15:44:59 <jscotka> andreasn, Is there any future feature page, what is planned (like adding support for firewall as mentioned, or cron, or new one, like printers, etc) ? 15:45:02 <mvollmer> ahh, actually, maybe not. 15:45:16 <andreasn> jscotka: yes, here you go https://github.com/cockpit-project/cockpit/wiki/Roadmap 15:45:49 <jscotka> andreasn, thanks, it is exactly what I was searching for :-) 15:45:49 <mvollmer> ok, I think the 4w section is OK, so let's move on to QA, ok? 15:46:29 <andreasn> I would move Atomic stuff from 4 weeks to 4 months 15:46:40 <andreasn> but anyway, yeah, next topic 15:46:47 <andreasn> #topic Q&A 15:47:02 <mvollmer> andreasn, yeah, agreed. 15:47:33 <mvollmer> cool navigation redesign is coming! 15:47:53 <andreasn> yesss 15:48:07 <jscotka> I'm looking forward 15:48:13 * mvollmer makes complicated secret hadn shake with andreasn 15:48:19 <andreasn> hehe 15:48:51 <andreasn> so, what do we have for Q&A? 15:49:07 <mvollmer> no, "QA" 15:49:13 <mvollmer> Quality Assurance 15:49:21 <mvollmer> :-) 15:49:45 <jscotka> sounds better :-) Q&A sound like questions & answers 15:49:48 <andreasn> oh, right 15:49:50 <mvollmer> jscotka, so you are getting up to speed? 15:49:53 <andreasn> sorry 15:49:55 <andreasn> #topic QA 15:50:35 <jscotka> but I need some Questions :-) where are stored your #xy logs? 15:50:54 <stefw> our IRC logs? 15:51:01 <stefw> andreasn will send them around after the meeting 15:51:12 <jscotka> yes, ah okay 15:51:20 <sgallagh> #undo 15:51:20 <zodbot> Removing item from minutes: <MeetBot.items.Topic object at 0x2bbc8d90> 15:51:22 <sgallagh> #undo 15:51:22 <zodbot> Removing item from minutes: <MeetBot.items.Topic object at 0x29ecac10> 15:51:28 <sgallagh> #topic QA 15:51:44 <jscotka> so it is not somewhere sent automatically 15:51:48 <andreasn> thanks sgallagh! 15:52:36 <andreasn> the meeting logs get stored here http://meetbot.fedoraproject.org/cockpit/ 15:53:16 <jscotka> I'll plan to integrate your current testsuite to internal testing infrastructure as first step. 15:54:15 <mvollmer> jscotka, we currently run the tests only manually. 15:54:16 <jscotka> thanks for link to meetbot 15:54:45 <mvollmer> and we would like to run them automatically for each github pullrequest 15:54:48 <stefw> indeed 15:54:53 <mvollmer> and publish the results 15:55:04 <mvollmer> can the internal infrstructure do that? 15:55:05 <jscotka> I'll create it beaker compliant 15:55:10 <andreasn> #info jscotka will intergrate the current cockpit test suite to the interal RH testing infrastructure as a first step 15:55:46 <mvollmer> beaker means testing on bare metal, right? 15:56:06 <jscotka> mvollmer, on various HW bare metal and virt as well 15:56:14 <mvollmer> ok 15:56:47 <mvollmer> so, can we publish the results? 15:56:51 <jscotka> It should also help me to find how it works now 15:57:23 <stefw> right 15:57:23 <jscotka> mvollmer, it would be possible without problems 15:57:28 <mvollmer> ok, nice! 15:58:41 <mvollmer> so we hear from you how it is going, right? 15:58:41 <jscotka> how is testsute scheduled now? manually via make test? or is it internal part of build process? 15:58:58 <stefw> we run it manually 15:59:03 <stefw> during our work process 15:59:03 <mvollmer> there is also travis 15:59:05 <stefw> on our own machines 15:59:08 <stefw> there is 'make check' 15:59:11 <stefw> which is unit tests 15:59:19 <stefw> that is scheduled automaticalcly via a github.com pull request 15:59:21 <jscotka> okay, 15:59:24 <stefw> the integration test suite 15:59:26 <stefw> does not run automatically 15:59:42 <stefw> but the end goal would be to have it run automatically for each github.com pull request 15:59:48 <stefw> and or update to a pull request 16:00:00 <stefw> similar to how travis works for us right now, with the unit test suite 16:00:00 <mvollmer> just like travis, basically. 16:00:10 <andreasn> #info end goal would be for the test suite to run for each github.com pull request 16:00:47 <jscotka> yes, I'll plan also to do it as part of our internal contionous integration project and results could be exported somewhere. 16:01:15 <mvollmer> jscotka, puiterwijk has been working on this in the past. you could contact him. 16:01:42 <jscotka> mvollmer, thanks I'll do. 16:02:00 <andreasn> #action jscotka to contact puiterwijk, who's been working on this in the past 16:02:07 <jscotka> how long running testsuite takes now? 16:02:24 <stefw> several minutes, if done in parallel 16:02:32 <stefw> using the TEST_JOBS=4 environment variable 16:02:39 <mvollmer> on one core, maybe 15 minutes? 16:02:42 <stefw> but it can take about 15 minutes if TEST_JOBS is not defined 16:02:49 <jscotka> andreasn, It is so fast, so it is perfect 16:03:13 <mvollmer> jscotka, it is still small, but it should grow! :-) 16:03:22 <stefw> it doesn't feel fast to me when waiting on it :) 16:03:33 <jscotka> andreasn, I have to learn how to work with your hashtags syntax :-) 16:03:58 <andreasn> it's for the meeting minutes 16:05:23 <mvollmer> jscotka, ok, so you are happy for now? Anything else to discuss now? 16:05:35 <jscotka> mvollmer, is there any possibility to automate also webUI tasks? I know that there is possible to use frameworks like seleniun-ish, but it is possible to do it via curl (with some stored secret cookie or another session storing abilities)? 16:05:45 <jscotka> mvollmer, yes I'm happy :-) 16:05:49 <stefw> we use phantomjs 16:05:55 <stefw> and sometimes ucrl 16:05:56 <stefw> curl 16:06:04 <jscotka> stefw, perfect 16:06:16 <stefw> most of our test suite revolves around automating web ui tasks 16:06:27 <stefw> the integration test suite that is ^^ 16:06:30 <mvollmer> very racy stuff... 16:06:47 <stefw> most of the *unit* test suite revolves around testing backend logic, javascript internals, etc.. 16:06:59 <jscotka> mvollmer, stefw mvollmer for now it is everything. Too much info for now :-) 16:07:03 <stefw> heh 16:07:06 <mvollmer> ok! 16:07:10 <mvollmer> one more thing, though! 16:07:41 <mvollmer> jscotka, we have some abstractions in place that might suit you: 16:08:04 <mvollmer> (hmm, how do I put this briefly.) 16:08:24 <mvollmer> we had planned to run the tests on either local vms or remote openstack instances 16:08:38 <mvollmer> so maybe you can hook into that when running things on beaker. 16:08:56 <mvollmer> just to keep in mind that there is already something in the code that might do this. 16:09:00 <jscotka> mvollmer, perfect, I hope that in next 2 weeks I'll understand that 16:09:07 <mvollmer> when you get to it, just ask again for the details. 16:09:29 <jscotka> mvollmer, thanks a lot 16:09:59 <jscotka> nowadays I undestand Cockpit only from user side :-) 16:10:25 <andreasn> all right, we're 10 past the hour. Anything else for the meeting? 16:10:50 * mvollmer puts down pencil. 16:11:04 <andreasn> all right 16:11:11 * jscotka leaving&eating :-) 16:11:20 <mvollmer> enjoy! 16:11:22 <andreasn> thanks everyone for your time! 16:11:27 <andreasn> #endmeeting