18:59:59 <roshi> #startmeeting Cloud WG
18:59:59 <zodbot> Meeting started Wed Apr  8 18:59:59 2015 UTC.  The chair is roshi. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:59:59 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
19:00:11 <roshi> #meetingname Cloud WG
19:00:11 <zodbot> The meeting name has been set to 'cloud_wg'
19:00:21 <oddshocks> .hello oddshocks
19:00:22 <zodbot> oddshocks: oddshocks 'David Gay' <dgay@redhat.com>
19:00:24 <roshi> #topic Roll Call
19:00:28 <roshi> .hello roshi
19:00:29 <Corey84> .fas Corey84
19:00:30 <zodbot> roshi: roshi 'Mike Ruckman' <mruckman@redhat.com>
19:00:33 <zodbot> Corey84: corey84 'Corey84' <sheldon.corey@gmail.com>
19:01:38 <roshi> #topic Previous Meeting Followup
19:01:41 <roshi> * jzb to take the Atomic Spin idea to the Spin SIG
19:01:42 <roshi> * dustymabe to update the ticket saying this will land in F22
19:01:47 <roshi> jzb: ?
19:02:36 <roshi> dusty's task is done
19:02:48 <jzb> roshi: yo
19:02:54 <jzb> .hellomynameis jzb
19:02:55 <zodbot> jzb: jzb 'Joe Brockmeier' <jzb@redhat.com>
19:03:00 <roshi> o/
19:03:16 <jzb> roshi: have not done that yet.
19:03:44 <roshi> kk, no worries
19:03:49 <roshi> any ETA in mind?
19:04:38 <jzb> roshi: walters just reopened (temporarily) the discussion
19:04:44 <jzb> roshi: so after that is closed again.
19:04:55 <roshi> ok
19:04:57 <roshi> thanks
19:05:01 <jzb> roshi: note that we may not qualify for Spin as it includes copr builds
19:05:11 <roshi> ah
19:05:23 <oddshocks> Did not know that was a thing
19:05:26 <jzb> so it may be we need to 1) either get an exception (possible but difficult) or 2) go with Remix instead of Spin
19:05:28 <roshi> me either
19:05:47 <jzb> basically anything for a Spin still needs to be built in Koji
19:05:49 <jzb> per spot
19:05:57 <roshi> good to know
19:06:08 <roshi> any reason why stuff can't be built in koji?
19:06:49 <jzb> roshi: we're building newer versions of the packages so ...
19:06:50 * roshi isn't that familiar
19:07:02 <dgilmore> you can not use copr builds in a spin
19:07:18 <jzb> dgilmore: yeah, we know :-)
19:07:20 <dgilmore> it would have to be a remix, which is entirely on you to put together
19:07:39 <roshi> if it's a remix, could it ever be blocking?
19:08:01 <roshi> oh wait, this is in leiu of it being a blocker for F23, right?
19:08:08 <roshi> an in the mean time fix
19:08:09 <jzb> roshi: blocking what? we would be (and want to be) outside the regular release cycle
19:08:37 <roshi> that's right, nvm - I'm just blabbering now
19:08:59 <roshi> #info ignore the crazy guy in the corner talking to himself about remixes
19:09:24 <roshi> alright, onto topics?
19:09:30 <jzb> roshi: go for it!
19:09:48 <roshi> #topic Maintaining docker images for F22
19:09:51 <roshi> #link https://fedorahosted.org/cloud/ticket/97
19:10:44 <roshi> you again jzb :)
19:10:55 * jzb grumbles about having to log into each fedora hosted trac instance separately
19:11:09 <jzb> yeah, I haven't had time to update the ticket, but
19:11:23 <jzb> 1. The current tc8 docker image seems fine.
19:11:35 <jzb> I think the problems we had early on are solved.
19:11:48 <roshi> and we're going to be doing the testing of the docker image, not the base WG, right?
19:11:50 <jzb> 2. I've sent a note to -devel about the current location of the Docker images
19:11:51 <roshi> good good
19:11:57 <jzb> roshi: hang on, still typing ;-)
19:12:21 <roshi> kk :p
19:12:27 <jzb> hoping to move promotion and listing of the Docker images to Cloud
19:12:48 <jzb> 3. I haven't gotten as far as testing yet :-(
19:13:03 <roshi> got a link to the mail sent to devel?
19:13:09 <jzb> but I will bring that up in the location/promotion of the images discussion.
19:13:13 <jzb> roshi: just a sec
19:13:22 <roshi> thanks, it'll be useful in the meeting logs
19:13:55 <jzb> #info https://lists.fedoraproject.org/pipermail/devel/2015-April/209742.html
19:14:28 <roshi> thanks
19:14:38 <jzb> and will update the ticket momentarily
19:14:45 <roshi> sweet, thanks
19:14:51 <roshi> anybody have anything else?
19:16:21 <roshi> moving on then...
19:17:38 <roshi> #Topic Dockerfiles care and feeding
19:17:41 <roshi> #link https://fedorahosted.org/cloud/ticket/84
19:18:34 <jzb> roshi: no update this week
19:18:34 * roshi is supposed to be helping with this, but have been doing release validation and making sure things are good to go before I head out on some PTO...
19:18:56 <roshi> wfm jzb
19:20:14 <roshi> anybody have anything for this ticket?
19:20:43 * roshi moves on
19:20:53 <roshi> #topic sha256sums
19:20:58 <roshi> #link https://fedorahosted.org/cloud/ticket/93
19:21:06 <roshi> I think we can close this one and reopen it for F23
19:21:45 <jzb> roshi: yeah
19:21:48 <roshi> proposal: remove the meeting tag from this ticket until F23
19:21:51 <roshi> +1
19:22:58 <oddshocks> +1
19:23:18 * roshi removes the tag - counting jzb in favor due to his 'yeah'
19:23:49 <roshi> #topic Updated Images
19:23:51 <roshi> #link https://fedorahosted.org/cloud/ticket/94
19:24:18 <roshi> I got nothing for this ticket
19:24:34 <jzb> oddshocks: ^^?
19:25:13 <oddshocks> Nothing more from me on that.
19:25:53 <oddshocks> Is there really anything we have to worry about with that for F21?
19:26:02 <roshi> and you're going to be shephearding that effort, right?
19:26:19 <roshi> don't think so
19:26:24 <roshi> per the last comment in the ticket
19:26:31 <jzb> if it's done let's close the ticket.
19:26:35 <roshi> we decided to leave it alone for F21 and start it for F22
19:26:43 <roshi> +1 jzb
19:26:53 <oddshocks> I don't know what else there is to do on that. I say +1, let's shove it off until F22
19:27:00 <oddshocks> er, until after F21 release
19:27:20 <roshi> F22 release you mean
19:27:21 <roshi> ?
19:28:06 <oddshocks> yes.
19:28:07 <oddshocks> yes, that.
19:28:08 <oddshocks> :P
19:28:24 * roshi closes the ticket
19:28:50 <roshi> #topic Open Floor
19:28:57 <roshi> that's all the tickets
19:29:05 <roshi> so, I wanted to take a quick poll
19:29:06 <kushal> ah.
19:29:15 <roshi> out of the following:
19:29:21 <roshi> 1 - Cloud Base Image
19:29:29 <roshi> 2 - Atomic Host Image
19:29:35 <roshi> 3 - Vagrant Images
19:29:50 <roshi> 4 - Atomic Images (not the bare metal host)
19:29:55 <roshi> 5 - Docker Images
19:30:13 <roshi> which of these are people under the impression they're release blocking for F22?
19:30:34 * roshi 1
19:30:43 <kushal> 1,2,4,5
19:30:50 <kushal> ^^ should be
19:31:02 <jzb> roshi: what's the diff between 2 and 4?
19:31:18 <jzb> roshi: 1,3,5
19:31:20 <roshi> one runs on bare metal and the other in the cloud
19:31:47 <roshi> was my impression, anyways
19:31:56 <roshi> oddshocks: ?
19:32:03 <roshi> jsmith: ^^
19:32:06 <roshi> number80: ^^
19:32:16 <oddshocks> hm
19:33:07 <oddshocks> I think I'm gonna agree with kushal. Definitely not 3. But we seem to be pretty focused on docker/atomic, and of course the base image should be blocking
19:33:33 <oddshocks> jzb: What are your thoughts on why atomic shouldn't be a blocker?
19:33:49 * oddshocks probably doesn't fully grok the entire situation
19:34:05 <jzb> oddshocks: because we're already proposing the 2-week spin/remix and not planning to have an Atomic Host f23 edition.
19:34:05 <roshi> atomic will be a spin/remix
19:34:09 <roshi> so outside the cycle
19:34:17 <jzb> so I don't see any reason to block F22 if there's some bug in the process that afflicts it.
19:34:35 <oddshocks> Ahhh
19:34:50 <roshi> out of those, does anybody have any notes that fesco and QA are aware of more than 1 being a blocker?
19:34:55 <roshi> because I don't have any
19:35:03 <oddshocks> Alright, I'll revise my answer, then :P
19:35:11 <roshi> the only one I know as blocker F22 is the base cloud image
19:35:55 <roshi> and that's also the only on QA has been tracking, has criteria associated with it or any amount of testcases for it
19:36:02 <oddshocks> That's the only one I'm 100% sure we should be blocking on
19:36:26 <roshi> I think how the process should go is this:
19:36:40 <jzb> oddshocks: we should absolutely be blocking on the Docker image if we're not.
19:36:55 <oddshocks> jzb: Alright
19:37:15 <roshi> we decide what we want to block on (too late for F22, IMO), we propose to fesco/qa/whoever, docs get written for it once decided (release criteria, test plans, etc)
19:39:47 <roshi> thoughts?
19:40:18 <jzb> roshi: I don't.
19:40:39 <jzb> roshi: is there a checklist somewhere that we should follow for these new things?
19:40:55 <roshi> not that I know of, but I can write one up
19:41:37 <jzb> roshi: I kind of feel like much of my life these days is about finding out new and interesting things that aren't obvious but should have been done and now it's a bit too late to do them, but check back in six months. ;-)
19:41:51 <roshi> I know the feeling :)
19:42:01 <roshi> I just want to make sure we're on the same page
19:42:19 <jzb> roshi: heck, at this point I'd settle for same book and chapter.
19:42:31 <jzb> we're probably all in the same library.
19:42:34 <roshi> I don't have strong feelings towards any of the particular images blocking or not blocking - I just want to make sure we're set up to succeed when we try to do them
19:42:41 <jzb> maybe even the same Dewey Decimal section.
19:42:47 <roshi> haha
19:43:13 <roshi> so, unless someone has notes somewhere - the only blocker we have for F22 is the cloud base image
19:43:36 <roshi> and the need for an impending vote when we have more people present
19:43:44 <roshi> then a fesco ticket and fesco meeting
19:43:49 <jzb> roshi: where would these notes *be*?
19:43:50 <roshi> that sound about right?
19:43:59 <roshi> devel list, meeting logs
19:44:06 <roshi> any mailing list, actually
19:44:12 <jzb> roshi: let's take a step back
19:44:25 <jzb> roshi: what's the Official Sanctioned Process for declaring something a blocker?
19:44:37 <roshi> afaik, there isn't one
19:44:37 <jzb> e.g. who decides that? Us? FeSCO?
19:44:53 <roshi> since in the past, it was fesco and the board maybe - and it was just desktops and arches
19:45:10 <jzb> lemme see if I can summon mattdm
19:45:25 <roshi> already invoked his name in cloud
19:46:07 <jzb> adamw: maybe you would know?
19:46:23 <jzb> adamw: what's the official decider for making something a release blocker or not?
19:46:42 <roshi> haha, I just pinged him in #fedora-qa about it
19:46:51 <jzb> roshi: he's here
19:46:52 * roshi mumbles something about great minds...
19:47:04 <jzb> well, he's lurking at any rate.
19:47:17 <roshi> yeah, I just know he monitors #fedora-qa pretty close
19:47:28 <kushal> adamw, adamw adamw adamw
19:47:44 <kushal> roshi, ^^^ adamw will notice this one.
19:47:52 <roshi> sgallagh: just answered in qa channel
19:48:10 <roshi> 19:47 < sgallagh> roshi: For arches, I think that was basically a joint Board/FESCo decision
19:48:39 <roshi> I think this is a new process that needs to come along with the fedora.next stuff
19:48:54 <sgallagh> Sorry, what's the context here?
19:49:07 <roshi> cloud has several differing images
19:49:18 <roshi> and several differing ideas of what's blocking release
19:49:22 <sgallagh> OK
19:49:27 <roshi> I only know of one, which is the cloud base image
19:49:38 <roshi> if you scroll up, you'll see a list of the images
19:49:49 <sgallagh> So in terms of the release media, each of the Product WGs is supposed to decide that, include it on the PRD and communicate it to FESCo and QA
19:50:04 <sgallagh> That was part of the initial planning stuff we did during the gap between F20 and F21
19:50:09 <sgallagh> It was meant to be ongoing.
19:50:13 <jzb> sgallagh: I think we may be in deep need of updating the PRD for Cloud
19:50:14 <roshi> that was my impression
19:50:28 <roshi> and we need to update the PRD and take it to fesco
19:50:30 <sgallagh> jzb: Yes, without a doubt.
19:50:37 <jzb> sgallagh: as I recall working on the PRD quite some time ago, finishing it, and I don't think we've referred to it since.
19:50:46 <sgallagh> roshi: It's too late for F22, obviously.
19:50:50 <roshi> we haven't really
19:50:52 <jzb> sgallagh: a PRD review should be part of the prep process?
19:50:54 <sgallagh> But yes, please update it early for F23
19:50:58 <roshi> that's what I thought too sgallagh
19:51:00 <kushal> jzb, we have to update it.
19:51:06 <jzb> kushal: yes
19:51:17 <kushal> jzb, that is actually in my TODO list.
19:51:38 <roshi> I'll add a meeting item to track progess on updating that for F23
19:51:40 <sgallagh> I'd say that decisions about blocking install media need to be approved before Alpha Freeze, but that's not a codified requirement.
19:51:44 <sgallagh> Just sensible (IMHO)
19:51:52 <roshi> +1 sgallagh
19:52:40 <jzb> kushal: yay! does that mean we're off the hook?
19:53:02 <sgallagh> jzb: The opposite. It means you can't slack off and then slip the release to finish :)
19:53:15 <jzb> sgallagh: I <3 sensible, but maybe codifying that would prevent unhappiness later?
19:53:29 <roshi> probably would be a good idea
19:53:30 <sgallagh> jzb: Oh absolutely.
19:53:40 <sgallagh> I just meant that I'm not authoritative, so take that for what it's worth
19:53:53 <jzb> sgallagh: you're not? oh.
19:53:53 <roshi> I think we just got stung by a new process we weren't really paying attention to and caught it late
19:53:55 <sgallagh> (Occasionally authoritarian, but rarely authoritative)
19:54:19 <sgallagh> jzb: I may be confusing you.
19:54:38 <sgallagh> New blocking media MUST be approved by FESCo (who will always be expected to check with QA before doing so)
19:54:46 <sgallagh> Right now, there's no official time this has to happen.
19:55:04 <sgallagh> On my personal opinion, the policy should be "before Alpha Freeze", but it's not written down anywhere at present.
19:55:19 <sgallagh> So in theory, FESCo could buck good sense and do so, but I'd vote against it :)
19:55:40 <kushal> jzb, nope, my TODO list says get jzb to write the updated PRD
19:55:45 <jzb> heh
19:56:04 <jzb> sgallagh: wait, isn't "is/isn't blocker" part of feature proposals?
19:56:30 <sgallagh> jzb: That would qualify as a "FESCo approval request"
19:56:51 <roshi> are features meant to be for new blocking products though?
19:56:58 <jzb> OK, so it's less wild west than I was thinking.
19:57:02 <sgallagh> roshi: Excellent question
19:57:03 <roshi> I see that as new included libs and whatnot
19:57:22 <roshi> new apps/functionality - ie, features
19:57:37 <sgallagh> I think I'd argue that if the Change was filed by a WG, then asking it to include blocking media is reasonable
19:57:38 <roshi> a feature of a new car is the sun roof, not another model
19:57:46 <sgallagh> I wouldn't necessarily say the same from J Random Contributor
19:58:15 <sgallagh> roshi: I have no problem with that mechanism starting the process.
19:58:21 <sgallagh> How we proceed from there is a matter of some discussion
19:58:30 <roshi> for sure
19:58:52 <sgallagh> (Also, we changed the term from Features to Changes for a reason)
19:59:35 <sgallagh> Since these are meant to be high-value alterations to The Fedora Project, not necessarily technical or user-facing features of one or more editions
20:00:03 <roshi> ah
20:00:15 <roshi> sorry for the equivocation :)
20:00:49 <roshi> ok, so what does this all mean for us *right now*?
20:01:31 <jzb> sgallagh: I should know this, but
20:01:43 <jzb> sgallagh: do we do any kind of release post-mortem?
20:01:51 <sgallagh> IMHO, *right now* it means that https://fedoraproject.org/wiki/Cloud/Cloud_PRD?rd=Cloud_PRD#Delivery_Mechanisms provides the list of blocking media
20:02:05 <jzb> I know marketing is *supposed* to, but I don't know if we do a project-wide one.
20:02:27 <sgallagh> The Release Readiness Meeting usually doubles as such
20:02:31 <sgallagh> But it's not really project-wide
20:02:59 <jzb> roshi: ok, so now I've figured out why I believe Atomic is blocking
20:03:02 <sgallagh> It's hard to do a project-wide meeting because it usually just ends up a flamewar on devel@
20:03:05 <jzb> roshi: http://fedoraproject.org/wiki/Changes/Atomic_Cloud_Image
20:03:16 <jzb> roshi: under Contingency Plan
20:03:28 <jzb> roshi: "Blocks product? Yes, Atomic Cloud Image (duh!)"
20:03:39 <jzb> roshi: so, yes, there are notes! :-D
20:03:45 <jzb> roshi: and it was accepted by FeSCO
20:04:12 <roshi> so it would block the atomic cloud image, but does that mean full release?
20:04:25 <roshi> I mean, I thought that line was kinda tautological
20:04:25 <jzb> roshi: what's missing here, is that using the feature mechanism, there's no automatic approval / disapproval
20:04:46 <jzb> roshi: you're right, but that's why I *thought* it ;-)
20:04:51 <roshi> ah
20:04:57 <roshi> that clears that up then
20:05:05 <roshi> and why I thought it didn't
20:05:12 <roshi> "Of course it blocks itself"
20:05:13 <roshi> lol
20:05:51 <jzb> roshi: yes, but if it's an accepted change we should make all necessary efforts to make sure it succeeds.
20:05:51 <sgallagh> jzb: My advice at this point is to not expect that issues exclusive to Atomic Cloud Image will block the release
20:06:00 <jzb> sgallagh: Agreed.
20:06:03 <sgallagh> But *do* apply for Freeze Exception where needed.
20:06:06 <oddshocks> +1
20:06:15 <adamw> jzb: sorry, i'm multitasking. we don't have a super solid process for it, it's been kind of ad hoc and on the fly so far.
20:06:33 <jzb> adamw: no worries, thanks. Maybe we should firm that up for 23.
20:06:43 <adamw> in ye olde days we only *had* a few images so it was fairly obvious.
20:06:46 <jzb> sgallagh: and I wonder if there's a way to do an informal postmortem.
20:06:52 <jzb> without it devolving.
20:07:01 <jzb> sgallagh: I think it'd be valuable.
20:07:09 <sgallagh> jzb: Like I said, we usually end up doing some of that durng the Final Release Readiness meeting
20:07:16 <roshi> I'm all for FE's and whatnot for atomic
20:07:27 <sgallagh> But maybe ask mattdm to set up a formal townhall meeting or something
20:07:28 <roshi> just wanted to nail down what would make us slip so we know ahead of time
20:07:42 <jzb> sgallagh: +1
20:08:15 <jzb> adamw: growth is painful. We seem to be experiencing pain, therefore we must be growing successfully, right? ;-)
20:08:39 <adamw> i'll drink to that...
20:08:47 <sgallagh> /me looks at all the eggshells around
20:08:52 <roshi> same here
20:08:59 <roshi> to the drinking
20:09:04 <roshi> not sure about egg shells
20:09:28 * jzb notices we may have gone off the rails slightly
20:09:35 <jzb> roshi: should we wrap it up?
20:10:15 <roshi> yep - I think so
20:10:27 <roshi> just wanted to try to reconvene in the same section of the library :p
20:10:32 * roshi sets the fuse
20:10:35 <roshi> 3...
20:10:46 <roshi> 2...
20:10:52 <roshi> 1...
20:10:57 <roshi> thanks for coming folks!
20:11:01 <roshi> #endmeeting