14:59:44 #startmeeting Server SIG Weekly Meeting (2015-07-14) 14:59:44 Meeting started Tue Jul 14 14:59:44 2015 UTC. The chair is sgallagh. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:59:44 Useful Commands: #action #agreed #halp #info #idea #link #topic. 14:59:44 #chair sgallagh mizmo nirik stefw adamw simo tuanta mitr danofsatx 14:59:44 #topic roll call 14:59:44 Current chairs: adamw danofsatx mitr mizmo nirik sgallagh simo stefw tuanta 14:59:48 hi 14:59:49 morning 14:59:55 yo 14:59:55 Hello 15:00:03 hola, peeps 15:01:05 I'll leave a couple minutes for folks to filter in 15:03:29 ok, I guess we're it today. 15:03:44 #topic Agenda 15:04:01 I had one additional small item that I forgot to put in the email 15:04:04 #topic Agenda Item 15:04:08 #undo 15:04:08 Removing item from minutes: 15:04:19 #info Agenda Item: Getty IP Address 15:04:29 #info Agenda Item: Fedora 23 Progress 15:04:39 #info Agenda Item: Atomic Server 15:04:42 Other items? 15:05:22 not sure - I need to look at the minutes for the meetings I missed 15:05:36 OK, if you think of anything, we'll get it in Open Floor 15:06:32 Hello 15:06:46 Hello 15:07:25 OK, let's proceed. 15:07:31 #topic Getty IP Address 15:07:53 So, stefw wrote some patches to support printing the IP on getty 15:08:12 I sent out a proposal for how to phrase this on the terminal. 15:08:39 Just to put it to rest: 15:09:14 Proposal: Use the approach described in https://lists.fedoraproject.org/pipermail/server/2015-July/001936.html with the \S amendment 15:09:41 * nirik nods. seems fine to me. 15:09:51 nice. 15:10:01 nice++ 15:11:00 #link https://lists.fedoraproject.org/pipermail/server/2015-July/001936.html 15:11:23 #idea Use the approach described in https://lists.fedoraproject.org/pipermail/server/2015-July/001936.html with the \S amendment 15:13:20 Votes? 15:14:16 oh, i didn't realize we were voting on ideas here 15:14:37 sure, whatever, +1. 15:14:47 Well, I raised it as a Proposal: above 15:14:54 Then I figured I'd add it to the notes 15:15:04 +1 15:15:10 +1 15:15:15 +1 sure 15:15:17 oh right, missed that. 15:15:36 * nirik steps away for a min 15:16:20 OK, with my own +1, I think that's accepted. 15:16:26 #agreed Use the approach described in https://lists.fedoraproject.org/pipermail/server/2015-July/001936.html with the \S amendment 15:16:31 #topic Fedora 23 Progress 15:16:37 ... or lack thereof 15:17:08 So, we're at branch, with Alpha Freeze only two weeks away. 15:17:19 Things are not going particularly well this cycle (by my estimation) 15:17:38 yeah, from here it's been a bit of a mess too. 15:18:06 I've been out of the loop....too darned busy with RL 15:18:28 I myself am behind on the role containerization work, mostly due to a number of unexpected fires 15:19:04 I'm going to see what I can manage by the 28th, but I have a few days of PTO coming up as well 15:19:23 sgallagh: Unfortunately, sometimes PTO days are the ones with the fewest distractions :-( 15:19:28 The Cockpit domain controller setup had to be deferred due to GSoC chaos. 15:19:42 jsmith: I'm taking my children to a theme park. No hacking allowed. 15:20:02 sgallagh: Yeah, laptops and roller coasters don't mix well... 15:20:17 sgallagh: so just to be clear, https://fedoraproject.org/wiki/Changes/Cockpit_Domain_Controller_GUI is not happening for 23, right? 15:20:23 adamw: Correct. 15:20:38 Until we can find some resources to pick it up, it's on hold. 15:21:13 I think those are the only things we planned on delivering this cycle 15:21:31 So unless I miss my guess, there's not going to be a lot of marketable change in this release. 15:21:44 Maybe the FreeIPA 4.2 release will help with that, though 15:22:08 Anyone else working on anything exciting for this cycle? 15:23:01 not in regards to Server ;) 15:23:15 * nirik is still working on passphrase policy, but thats not really exciting. ;) 15:23:26 nirik: But extremely noteworthy 15:24:07 So, the biggest problem I think we're facing is a lack of resources (primarily available cycles). 15:24:35 I'd really appreciate it if everyone on the WG (and in the SIG) would make an effort to publicize our work and goals a bit more. 15:24:49 I'd like for us to be semi-actively recruiting folks. 15:26:16 I'd also like if you could help with the occasional blog post or lightning talk out in the wild. 15:26:21 i didn't get anywhere with openQA stuff, unfortunately, as i wound up spending the whole week working out why all our existing openQA tests didn't work and getting those problems fixed. 15:26:36 (If I have to hear "Why would I use Fedora Server instead of CentOS?" *one more time*...) 15:26:38 i was looking at the possibility of doing some test days for server stuff this cycle, though; anyone interested in that? 15:26:56 adamw: What in particular? 15:27:05 * nirik might be able to help with test days. 15:27:31 er, what have we got? :) how are Containerized Server Roles coming? 15:27:47 "(11:18:28 AM) sgallagh: I myself am behind on the role containerization work, mostly due to a number of unexpected fires" 15:27:52 if we don't have any new stuff we could at least try and run one as an opportunity to get a wider audience looking at existing bits 15:27:58 ah fun, /me will learn to read one of these days 15:28:19 adamw: reading is for loosers. 15:28:33 right now it seems like validation testing is basically done by two or three of us, who might miss stuff a wider audience would see... 15:28:36 adamw: I would like to run a Domain Controller/DB server test day though 15:28:37 danofsatx: so's spelling 15:28:45 sgallagh: right, something like that 15:28:50 If nothing else, it might catch issues introduced by switching rolekit and firewalld over to python 3 15:28:55 let's get a ticket filed this week 15:29:24 /me nods 15:29:37 #action sgallagh to file a ticket for a Server Role Test Day 15:29:51 did we ever test DB role? 15:30:17 danofsatx: No Test Day, but I ran it through a lot of release validation 15:30:22 like, test-day type 15:30:43 roger 15:31:34 OK, so the short version is that we should probably let Marketing know that Fedora Server will likely have few visible changes in F23 and they should focus on the other editions this time around 15:31:45 agreed 15:31:55 refocus efforts for F24 15:32:13 By "refocus" I hope we mean "go on a recruiting spree" 15:32:19 that too ;) 15:32:35 hopefully, I'll have new employment that will allow me to devote more resources to open source 15:33:27 #info Fedora Server 23 will have few user-visible changes. Encourage marketing to focus efforts on the other Editions. 15:33:50 OK, last agenda item: 15:33:59 #topic Atomic Server 15:34:39 As many of you have probably seen, the Workstation team is looking into eventually providing an Atomic Workstation system 15:35:15 The basic idea around it is that the core platform will be a read-only, atomically-updated system and that all of the user and system applications will be containerized. 15:35:30 Is this something we would also like to be exploring in the Fedora Server? 15:36:21 it seems like a direction lots of people are thinking/moving in, but it feels slightly odd to do it at the flavor level, to me... 15:37:21 adamw: Well, the plan is to not abandon RPMs as a deployment mechanism in the near future 15:37:45 (And with rpm-ostree to use those RPMs to create the read-only platforms) 15:37:49 adamw: yeah, seems also to me more like a format 15:38:01 So I think technologically it makes sense, even if it's a little awkward conceptually 15:39:38 OK, I'm not hearing an overwhelming response either way. Is this something we just don't care about for the immediate future? 15:39:40 An atomic OS (meaning single-package, take-it-or-leavit deal, as opposed to “unwritable /bin /lib /usr” we’ve always had) makes sense to me; moving libraries into applicatinos not really. 15:40:43 * adamw tends to be bad at thinking about things that don't exist, so isn't very keen to give thoughts 15:40:54 * danofsatx has to bow out. Duty calls. 15:41:07 adamw: Lies: you talk about "quality" all the time :-P 15:41:11 :P 15:41:14 mitr: yeah. +1 15:41:27 sgallagh: quality is like luck, it exists in 'bad' and 'good' varieties ;P 15:41:31 It would be nice to explore this, sure; can anyone spare the time? 15:42:22 (betting that the atomic OS is a prerequisite to applications actually relying on an OS, and thus not needing libraries so much)\ 15:43:02 would we be using the same core as Workstation or not? 15:43:20 adamw: Great question. 15:43:39 I suppose that's going to depend on how the layering works 15:43:52 I expect that we won't want the full GNOME Shell environment 15:44:09 Which I would think would be part of the atomic image. The applications, probably not. 15:44:13 I’d expect there to be push to keep GUI out of the OS, but having the same set of API/ABI libraries in both would make sense to me. 15:44:24 mitr++ 15:44:48 sgallagh: (This means also having libjpeg and libgtk3 in the OS rather than Workstation pushing copies of these into apps) 15:45:11 mitr: I don't think it necessary mandates that 15:45:26 But that's certainly important to figure out 15:46:06 sgallagh: We could perhaps split into base->{server,workstation} but that’s making the job of attracting developers harder 15:46:28 Anyway… spare cycles? 15:46:37 I absolutely don't have any :( 15:46:52 I started with zero and it's gone downhill from there. 15:47:24 My time allocation is completely unknown at this moment 15:49:11 i don't think i'd be the right person for this anyway. 15:49:36 #info Help wanted to investigate an Atomic Server platform. 15:49:50 * nirik is also very busy. 15:51:03 OK, let's move on 15:51:06 #topic Open Floor 15:52:22 * nirik has nothing 15:53:24 OK, setting the fuse 15:57:32 #endmeeting