15:01:08 <pknirsch> #startmeeting Fedora Base Design Working Group (2014-12-05)
15:01:08 <zodbot> Meeting started Fri Dec  5 15:01:08 2014 UTC.  The chair is pknirsch. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:08 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:01:16 <pknirsch> #meetingname  Fedora Base Design Working Group
15:01:16 <zodbot> The meeting name has been set to 'fedora_base_design_working_group'
15:01:27 <pknirsch> Hey and good morning/afternoon everyone!
15:01:56 <pknirsch> #chair haraldh msekleta dgilmore masta
15:01:56 <zodbot> Current chairs: dgilmore haraldh masta msekleta pknirsch
15:02:04 <haraldh> <-
15:02:21 <pknirsch> ping dazo nphilipp for first topic ;)
15:02:27 <masta> Howdy folks, I'm half here today
15:02:31 <nphilipp> Hi
15:02:34 <dazo> hey!
15:02:35 <msekleta> hello
15:02:52 <pknirsch> alright, lots of topics today, so lets quickly jump in
15:03:20 <pknirsch> #topic Status buildrequires cleanup work (davids & nils!)
15:03:31 <pknirsch> So dazo, nphilipp, whats new? :)
15:03:32 <dazo> I've not had time to look much at buildreq-check this week, but XML-RPC prototype server is running as openshift instance and seems to work ... awaiting some real test runs
15:04:06 <dazo> code needs some clean-up and feature improvements ... but it's ready for real PoC testing
15:04:08 <nphilipp> to add to that, I've worked on configuration (dazo: just pushed the code in my branch)
15:04:22 <nphilipp> and I've continued on cleanup :)
15:04:26 <pknirsch> nice :)
15:04:32 <pknirsch> so lets do some real PoC then!
15:04:46 <nphilipp> right now everything is in one module and I'm in the process of pulling it apart
15:05:01 <nphilipp> e.g. DB model in its own module, config in its own (that one is already so)
15:05:03 <nphilipp> and so forth
15:05:08 * pknirsch nods
15:05:14 <nphilipp> XMLRPC app into its own module
15:05:17 <nphilipp> stuff like that
15:05:50 <nphilipp> we also got the way we worked together ironed out, that helps as well :)
15:05:59 <pknirsch> Who want's to kick off the real test run then?
15:06:53 <nphilipp> pknirsch: jog my memory, where's the authoritative list with "base packages"?
15:07:14 <pknirsch> http://www.harald-hoyer.de/2014/01/14/self-hosting-fedora-base/
15:07:22 <nphilipp> I have murky recollections about some of haraldh's graphs, yes
15:07:27 <pknirsch> aye
15:07:54 <nphilipp> so it's basically this: https://harald.fedorapeople.org/kernel-br-all.txt right?
15:07:58 <haraldh> well, that's only for the kernel build
15:08:01 <pknirsch> yep
15:08:09 <haraldh> not the authoritative list with "base packages"
15:08:11 <nphilipp> right
15:08:33 <pknirsch> but it's a good starting point for a test
15:08:42 <nphilipp> in an ideal world we wouldn't need the java cruft in there, do we know if there's progress been made there?
15:09:07 <nphilipp> That's already 1806 packages, so I guess that's a starting point
15:09:58 <nphilipp> anyway, there's some schema-related clean up I want to do before we start filling the DB
15:10:05 <pknirsch> java was among the 65 choke point packages identified by haraldh later in the documentation.
15:10:21 <nphilipp> I plan to add schema migration but it'd be swell if we got by without it during the test run :)
15:10:47 <pknirsch> From my review i did earlier this year the main blocker was the dbX-java subpackage.
15:11:00 <nphilipp> ahh yes
15:11:26 <pknirsch> if either you nphilipp or dazo could pick up that review that would be fantastic, too.
15:11:38 <pknirsch> and should help a long way to reduce our initial number, too.
15:11:53 <pknirsch> see http://harald.fedorapeople.org/bootstrap-not-needed-deps.txt
15:11:56 <haraldh> https://fedoraproject.org/wiki/Base#Details
15:12:00 <pknirsch> or that :)
15:12:01 <haraldh> also as a start
15:12:33 <nphilipp> that'd be even more minimal
15:12:38 <nphilipp> cool
15:13:00 <pknirsch> feel free to kick off a test with any of those lists.
15:13:07 <nphilipp> ok
15:13:08 <pknirsch> maybe starting small might be a better idea actually
15:13:16 <dazo> yeah, starting with 125 packages (kernel rpm yum) seems to be a good place
15:13:20 <pknirsch> aye
15:13:22 <nphilipp> yeah, we'll probably run into issues soon enough :)
15:13:29 <pknirsch> to work out any kinks or issues in the system
15:13:37 <dazo> nphilipp, you don't think my code is perfect!? ;-)
15:13:45 <pknirsch> ha!
15:14:01 <nphilipp> dazo: well, if it's perfect I can go into the weekend, now can't I :P?
15:14:09 <dazo> of course!
15:14:11 <dazo> ;-)
15:14:12 <pknirsch> i think i only once wrote perfect code on the first try ;P
15:14:14 <nphilipp> world peace could be so easy :)
15:14:19 <pknirsch> in my 30 years of coding ;)
15:14:30 <pknirsch> (non trivial code that is)
15:15:41 <pknirsch> anyway, lets move on, lots of other stuff to cover today and don't want to keep EMEA folks away from their well deserved weekend too long :)
15:15:53 <pknirsch> any other important things about the buildreq checks?
15:16:05 <dazo> not from me
15:16:10 <nphilipp> neither from me
15:16:18 <pknirsch> okidokie, thanks guys!
15:16:33 <pknirsch> next up then
15:16:37 <pknirsch> #topic Docker update / Rocket container check
15:16:54 <pknirsch> Not sure if vpavlin is with us today, haven't seen him on irc at least
15:17:04 <pknirsch> anyone else who could give a Docker update? msekleta maybe?
15:18:23 <msekleta> I don't think anything (out PRs on GitHub) moved on
15:18:32 <msekleta> but let me check on that real quick
15:18:38 <pknirsch> ok, thanks msekleta
15:19:05 <msekleta> https://github.com/docker/docker/pull/8478
15:19:33 <msekleta> this is the show stopper, waiting on approval from Solomon
15:19:47 <pknirsch> In the meantime, i just wanted to mention the new container tech from CoreOS, Rocket. Not sure if we're interested in that yet, but imho it's a good idea to keep that on our radar and see if that would be something we'd want in Fedora, too.
15:20:18 <pknirsch> is anyone in contact with Solomon to drive this?
15:20:25 <pknirsch> i suspect vpavlin :)
15:21:00 <msekleta> they are both on DockerCon in Amsterdam
15:21:20 <pknirsch> alright. lets hope they get it resolved there maybe in person ;)
15:21:51 <pknirsch> And iirc vpavlin already had a look at Rocket, too, so we'll have to wait till next week for some more info on that
15:22:08 <msekleta> not sure, seems like DockerCon is more about marketing rather then hacking and talking code
15:22:38 <pknirsch> ah, there's vpavlin, right on time :)
15:23:32 <pknirsch> vpavlin: any other news on Docket except the pending pull request for https://github.com/docker/docker/pull/8478 ?
15:23:46 <pknirsch> vpavlin: and any quick thoughts on Rocket?
15:25:22 <msekleta> pknirsch, until vpavlin chimes in, I'd like to mention docker people announced couple new project docker-swarm and docker-machine
15:25:36 <pknirsch> oh, whats those about?
15:26:00 <msekleta> https://github.com/docker/swarm
15:26:01 <vpavlin> pknirsch: uh..I just woke my laptop to write some notes from discussion with Docker guys:)
15:26:08 * vpavlin is on DockerCon
15:26:12 <vpavlin> *at
15:26:19 <pknirsch> we know vpavlin :)
15:26:19 <msekleta> https://github.com/docker/machine
15:26:37 <vpavlin> pknirsch: So turning off again - last talk from Hykes and Golub ahead..
15:26:41 <vpavlin> Sorry:)
15:26:51 <pknirsch> alright, cya vpavlin !
15:27:08 <pknirsch> msekleta: interesting
15:27:19 <pknirsch> so docker clusters and docker mgmt
15:27:35 * msekleta admits he didn't look into much details yet
15:28:09 <vpavlin> I'll post some status on ML:) But...talked to @unclejack about systemd and I maybe convinced him it's benefitial for Docker to work well with systemd..we will se..
15:28:47 <pknirsch> thanks vpavlin
15:28:52 <msekleta> imho the take away is that CoreOS and Docker are now direct competitors on all the levels of the container stack
15:29:04 <pknirsch> aye
15:29:12 <pknirsch> and with CoreOS launching Rocket
15:29:27 <pknirsch> they have on their side now their own container tech
15:29:56 <msekleta> and they already have cluster management, etcd and fleet
15:30:00 <pknirsch> mhm
15:30:31 <vpavlin> Docker announced cooperation with Mesos on cluster mgmt..
15:30:48 <vpavlin> i.e. Mesos will be backend for swarm...
15:30:56 <pknirsch> no surprise there ;)
15:31:45 <msekleta> vpavlin, are kubernetess still in the game then?
15:32:16 <vpavlin> yup - I mean that it's basically that both has it all now...though Docker has waaaaaaay bigger community and real usecases presented by big companies (f.e. BBC)
15:32:37 * pknirsch nods
15:33:33 <vpavlin> msekleta: I talked to Solomon about this yesterday and he says the only thing they want is standard interface - swarm is really stupid and you can write your own orchestrator as backend for it - so yes, they are
15:34:56 * msekleta nods, consolidation and common APIs everybody agrees on are way to go
15:35:04 <pknirsch> jup
15:35:38 <vpavlin> the biggest problems people mention are secrets for build and runtime and that Dockerfile builds are slow - lot of them write custom solution very similar to Openshift's STI
15:36:26 <vpavlin> I wanted to talk to tianon about Fedora base image but was not able to find him (not sure if he is here...) but I am going to find somebody else..
15:36:35 <pknirsch> cool
15:37:11 <vpavlin> Ok...bye guys...:) Have a nice weekend!
15:37:19 <pknirsch> cya vpavlin ! and thanks for the infos
15:37:30 <msekleta> vpavlin, bye, enjoy rest of the conference
15:38:01 <pknirsch> alright, lets move on then to the next topic
15:38:15 <pknirsch> #topic Status rpm mechanisms for multiple config subpackages
15:38:27 <pknirsch> Not sure if we got any news on that. Maybe haraldh or ffesti?
15:38:53 <haraldh> heard nothing from the rpm front :-/
15:38:58 <pknirsch> hm, alright
15:39:20 <msekleta> not sure you guys know or not, but Panu left rpm crew
15:39:32 <pknirsch> yea, i read about that last week
15:39:41 <pknirsch> already talked with him privately, too :)
15:40:04 <haraldh> oh
15:40:05 <haraldh> ok
15:40:19 <pknirsch> ffesti will be taking over full ownership last i heard
15:40:51 <msekleta> yep, afaict it is ffesti only now
15:41:50 <pknirsch> though jzeleny mentioned getting one more guy on board
15:42:34 <pknirsch> #info no news on the rpm front
15:42:40 <pknirsch> next topic then
15:42:44 <pknirsch> #topic Status rpm mechanisms for factory reset files
15:42:51 <pknirsch> i guess same here then haraldh ?
15:42:56 <haraldh> yes
15:43:05 <pknirsch> #info same here, need input from rpm team
15:43:08 <pknirsch> alright
15:43:18 <pknirsch> last topic then
15:43:25 <pknirsch> #topic Base WG ownership of generic network install images
15:43:49 <pknirsch> As we're nearly all back now lets see what this is about. anyone can give me an update on this?
15:44:44 <haraldh> +1
15:44:49 <msekleta> +1
15:47:31 <pknirsch> so what would the Base WGs responsibilities and work then be? and who'd be doing it, haraldh and msekleta ?
15:47:49 * pknirsch wants to understand a bit more in detail what this is exactly about
15:48:04 <msekleta> pknirsch, just to explain quick +1, we already discussed this, and me and haraldh were in favor
15:48:48 <pknirsch> sure, but before i can give a + or - i'd like to understand the details a bit more
15:49:01 <pknirsch> apologies for not going through the last weeks logs in more detail
15:50:07 <haraldh> because it's not a product on its own, I don't think there is much to discuss about what should be in there
15:50:21 <haraldh> it's more like "owning" the packages
15:50:39 <haraldh> or being the "authority" of it
15:51:04 <haraldh> should there be any concern about it, we would be the deciding working group
15:51:13 <haraldh> and not server and not desktop
15:51:46 <haraldh> because they basically share this block
15:52:41 <msekleta> as I understand, we would monitor the situation, and I case there is a problem with the packages (base packages needed for partitioning, anaconda) we would pull respective maintainers into meeting to decide on next steps
15:53:27 <msekleta> s/I case/in case/
15:53:31 <pknirsch> ok, that makes sense
15:53:38 <pknirsch> as it's happening anyway
15:53:45 <haraldh> exactly
15:54:04 <pknirsch> sounds reasonable then
15:54:06 <pknirsch> +1
15:54:08 <pknirsch> then as well
15:54:14 <haraldh> We discussed about anaconda being part of base already..
15:54:21 <pknirsch> aye
15:55:37 <pknirsch> lets double check with the remaining guys next week, but i think this item shouldn't be a big point of discussion
15:56:52 <pknirsch> #info haraldh, msekleta and pknirsch all in favor of Base owning the netinstall images. Getting remaining input/votes from other members next week.
15:57:00 <pknirsch> and last thing for today:
15:57:06 <pknirsch> #topic Open Floor
15:57:14 <pknirsch> anything else from anyone for today?
15:57:56 <haraldh> -1
15:58:10 <pknirsch> msekleta, anything else from you?
15:58:40 <msekleta> I don't have anything else to add
15:58:50 <pknirsch> oki
15:59:10 <pknirsch> i have one last thing organizationally: Meetings over the next weeks.
15:59:25 <pknirsch> As we're heading into holiday season
16:00:08 <pknirsch> Next week i won't be able to attend the meeting at least due to our office holiday party, so if anyone else would take over that that would be great
16:00:19 <pknirsch> the week after, the 19th, we could still hold one
16:00:38 <pknirsch> but then the 2 weeks after i'll at least be out again, and i expect most of us will be on Dec 26th
16:00:46 <haraldh> I am on the same office holiday party also :)
16:00:59 <pknirsch> right!
16:02:11 <pknirsch> msekleta, would you be able to hold the meeting next week?
16:02:30 <pknirsch> otherwise we'd only have 19th this month
16:02:43 <msekleta> sure I can do that, not sure how much sense does it make
16:02:53 <haraldh> I would just do the 19th
16:02:57 <msekleta> without you gone
16:02:59 <pknirsch> ok, lets do that then
16:03:00 <haraldh> nothing critical anyway
16:03:03 <pknirsch> yea
16:03:21 <msekleta> s/without/with
16:04:01 <pknirsch> #info proposal: Due to holiday season proposal to do last meeting this year on 19th. haraldh, msekleta and pknirsch all in favor.
16:04:21 <pknirsch> ok, thats all
16:04:26 <pknirsch> thanks everyone then!
16:04:38 <pknirsch> and thanks haraldh for covering the meetings while i was gone :)
16:04:39 <haraldh> have a nice weekend!
16:04:44 <pknirsch> same to you guys
16:04:45 <haraldh> welcome
16:04:47 <pknirsch> o/
16:04:50 <pknirsch> EoM
16:04:52 <pknirsch> #endmeeting