16:02:05 <Rathann> #startmeeting FESCO (2017-03-10)
16:02:05 <zodbot> Meeting started Fri Mar 10 16:02:05 2017 UTC.  The chair is Rathann. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:02:05 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:02:05 <zodbot> The meeting name has been set to 'fesco_(2017-03-10)'
16:02:05 <Rathann> #meetingname fesco
16:02:05 <zodbot> The meeting name has been set to 'fesco'
16:02:35 <jforbes> .hello jforbes
16:02:36 <zodbot> jforbes: jforbes 'Justin M. Forbes' <jforbes@redhat.com>
16:02:43 <nirik> morning everyone.
16:02:53 <jwb> hi
16:03:31 <Rathann> #chair rathann jforbes nirik jwb
16:03:31 <zodbot> Current chairs: Rathann jforbes jwb nirik rathann
16:04:27 <Rathann> jsmith, paragan, kalev-afk, sgallagh: ping
16:05:21 <sgallagh_> .hello sgallagh
16:05:22 <zodbot> sgallagh_: sgallagh 'Stephen Gallagher' <sgallagh@redhat.com>
16:06:16 <Rathann> #chair sgallagh
16:06:16 <zodbot> Current chairs: Rathann jforbes jwb nirik rathann sgallagh
16:06:31 <Rathann> ok, that's 5
16:06:51 * Rathann wonders why zodbot is case sensitive
16:07:09 <maxamillion> .hello maxamillion
16:07:10 <zodbot> maxamillion: maxamillion 'Adam Miller' <maxamillion@gmail.com>
16:07:16 <Rathann> welcome
16:07:22 <Rathann> #chair maxamillion
16:07:22 <zodbot> Current chairs: Rathann jforbes jwb maxamillion nirik rathann sgallagh
16:07:31 <maxamillion> sorry, in the Atomic Host WG VFAD and wasn't paying attention to time
16:07:35 <Rathann> #topic init process
16:08:08 <Rathann> not much on the agenda today
16:08:24 <Rathann> #topic #1686 F27 System Wide Change: No More Alphas
16:08:24 <Rathann> .fesco 1686
16:08:25 <zodbot> Rathann: Issue #1686: F27 System Wide Change: No More Alphas - fesco - Pagure - https://pagure.io/fesco/issue/1686
16:09:10 <Rathann> adamw: around?
16:09:32 <nirik> I don't think he's around today...
16:09:36 <nirik> and dgilmore is in meetings.
16:09:47 * nirik has no questions tho, I'm +1 to the change.
16:09:50 <jforbes> While I think the underlying changes here will gate the kernel process since we do daily rawhide builds, I do see the overall benefit to fedora
16:10:04 <jforbes> I'm +1 as well
16:10:30 <jwb> 0
16:10:42 <maxamillion> +1 here
16:11:03 <sgallagh> Alpha does provide an important check-in date though
16:12:08 <jforbes> sgallagh: My impression was that the feature status checks would still be included in the schedule
16:12:12 <sgallagh> We have a fair amount of process around Alpha, including the contingency plan stuff
16:12:22 <nirik> "We will likely want to add a new checkpoint for change implementation that currently needs to be checked at Alpha "
16:12:29 <sgallagh> /me nods
16:12:55 <Rathann> yes, I think we should keep the other checkpoints, too
16:13:16 <jwb> i guess i've either missed it, or it wasn't spelled out clearly, but what exactly is doing the gating
16:13:39 <sgallagh> I'm +1, but I think it's going to necessitate some non-technical process changes as well as technical ones.
16:13:59 <Rathann> jwb: the compose process, IIUC
16:14:06 <jwb> that doesn't help
16:14:07 <jforbes> jwb: There was talk in the thread about either something automatic with bodhi or using a different system, but basically running automatic tests
16:14:21 <jwb> right, all of this is handwavy
16:14:56 <nirik> not bodhi
16:15:03 <Rathann> I am +1 in general
16:15:12 <sgallagh> jwb: What I was hearing this week was basically that Bodhi would be enabled for Rawhide, but that by default it would go automatically to the repo unless tests failed
16:15:13 <nirik> taskotron and/or openqa
16:15:22 <Rathann> so that's +1:5 and 0:1
16:15:25 <sgallagh> Which apparently disagrees with what others are hearing...
16:15:38 <jwb> you guys are just illustrating my point for me.  keep going.
16:16:07 <nirik> "But that's not a part of the proposal. Nothing in the proposal involves using Bodhi."
16:16:13 <nirik> (from adamw's post to devel list)
16:16:21 <jwb> i like the idea behind this.  i'd like to understand how it will actually impact contributors
16:16:43 <jwb> Rathann: yes, it's passed.  we can likely move on since nobody actually has details
16:17:03 <jforbes> Taskotron runs the tests.  The real question is where results can be seen. Bodhi was mentioned by adamw as a possible integration point there, but entirely automated
16:17:56 <Rathann> jwb: there's no rush, we can defer and ask for more implementation details
16:18:02 <nirik> well, the idea is pagure in front of dist git
16:18:08 <nirik> but I am not sure the status of that.
16:18:13 <jforbes> "The major focus at present is to deploy a Pagure instance on top of dist-git, as that provides another point at which we can provide a proper test feedback loop. "
16:18:24 <jwb> Rathann: i kind of disagree.  this stuff should be landing now
16:18:32 <jwb> rawhide is off churning away for f27
16:18:37 <jwb> if this is an f27 change...
16:18:53 <nirik> right and it says it should be landing shortly after f26 branches... so... now.
16:19:08 <jwb> yet nobody has details
16:19:15 <jwb> all was have are ideas
16:19:28 <jwb> which means i question if this can really land in time for f27
16:19:37 <jwb> but i'm not doubtful enough to block it
16:20:36 <jforbes> jwb: I would agree, but it seems the handwavy ideas are really all in the reporting portion of it, the actual interface.  The tests themselves are there.
16:20:38 <nirik> it's hard without having the change owners here.
16:20:49 <jwb> it's already passed guys.  move on
16:21:25 <Rathann> ok, I'll ping the change owners for more details
16:21:50 <jforbes> And the big thing here, is the gating of rawhide doesn't seem to be dependent on this change
16:22:05 <Rathann> #action Rathann will ping adamw and ausil for more details on  F27 System Wide Change: No More Alphas
16:22:41 <nirik> taskotron is already running on rawhide builds... but yeah.
16:23:35 <Rathann> #agreed Approve the  F27 System Wide Change: No More Alphas (+1:5, 0:1, -1:0)
16:23:49 <Rathann> #topic #1687 Consider adding SSSD to one of the default comps groups
16:23:55 <Rathann> .fesco 1687
16:23:56 <zodbot> Rathann: Issue #1687: Consider adding SSSD to one of the default comps groups - fesco - Pagure - https://pagure.io/fesco/issue/1687
16:24:48 <sgallagh> /me can answer questions on this if required
16:24:57 <nirik> sure. +1
16:26:03 <Rathann> I'm +1
16:26:10 <jforbes> sgallagh: Is this something that is going to annoy atomic?
16:26:13 <Rathann> sgallagh: does this replace nscd?
16:26:35 <sgallagh> jforbes: I doubt it; they don't use @core as far as I am aware
16:27:03 <sgallagh> Rathann: it's not a 100% replacement
16:27:16 <jforbes> sgallagh: okay, I thought they were still using core.  If not, I am all for it.
16:27:20 <sgallagh> Because SSSD doesn't handle the hosts or services maps
16:27:27 <jforbes> +1
16:27:59 <sgallagh> jforbes: That said, I think atomic ships sssd-common in the host already anyway
16:28:18 <jwb> this puts SSSD in the path of all logins?
16:28:40 <Rathann> sgallagh: how does it interact with nscd, then?
16:28:57 <jforbes> sgallagh: Yeah, I could see a lot of people would want it, but a lot of the atomic folks are really squidgy about any size changes too.
16:29:05 <sgallagh> jwb: No, it's not being added to PAM by default
16:29:13 <sgallagh> It's being added only for name-service
16:29:26 <jwb> like dns?
16:29:42 <sgallagh> jwb: SSSD doesn't handle the hosts map
16:29:43 <jwb> i know absolutely nothing about SSSD
16:30:02 <sgallagh> It handles users, groups, initgroups, sudo, autofs and netgroups
16:30:16 <sgallagh> Might be one more I'm forgetting
16:30:39 <jwb> so i'm trying to figure out how it's going to impact users on a daily basis
16:30:50 <jwb> how does it handle all those things?
16:31:05 <sgallagh> jwb: It provides libnss_sss.so.2 for glibc
16:31:34 <sgallagh> SSSD also has a fast in-memory cache these days which is comparable in performance to nscd
16:31:40 <jwb> sgallagh: ok, that means nothing to me.  let's take this offline
16:31:48 <maxamillion> I think SSSD is already in the atomic host image actually
16:31:57 <sgallagh> jwb: name-service uses plugins configured in nsswitch.conf
16:32:06 <sgallagh> SSSD provides a thin shim plugin that talks to the daemon
16:32:26 <sgallagh> maxamillion: Yes, it is. They're looking at moving it to a container, but for now it's tightly part of the host
16:32:30 <Rathann> I'm counting 3 +1
16:32:36 <Rathann> anyone else want to vote?
16:32:49 <Rathann> or defer?
16:32:57 <sgallagh> I have a couple more things to say, if I may
16:33:01 <Rathann> sure
16:33:09 <jwb> sgallagh: i know you're trying to be helpful, but you're talking at levels that all sound like gibberish to me.  it's OK, i can go read
16:33:29 <sgallagh> Today, if you use SSSD, there's a performance penalty because SSSD doesn't play well with nscd (you have to disable nscd for the maps handled by SSSD)
16:33:50 <sgallagh> This means that looking up local users actually has to hit the disk, whereas if nscd was in play it would just go to *its* cache.
16:33:58 <maxamillion> sgallagh: rgr
16:34:08 <sgallagh> We fixed that in SSSD recently so that now the local users can also be cached by SSSD for significant performance improvement
16:34:26 <sgallagh> So now that we have that, we want to offer it by default
16:34:37 <maxamillion> yeah, arlight
16:34:38 <maxamillion> +1
16:34:52 <sgallagh> This request is to put just the pieces of SSSD that enable the "files provider" onto the system by default.
16:35:06 <sgallagh> It does not include the LDAP, Kerberos, FreeIPA, AD, etc. functionality.
16:35:14 <Rathann> right
16:35:15 <sgallagh> That's available in @standard today.
16:35:43 <Rathann> I'm slightly worried about having both nscd and sssd running by default since they overlap in this case
16:35:55 <sgallagh> Rathann: nscd doesn't run by default on Fedora
16:36:13 <Rathann> oh, right
16:36:19 <Rathann> I've just checked it too
16:36:26 * Rathann somehow missed this change
16:36:38 <sgallagh> It's been that way for at least five years
16:37:34 <sgallagh> Any further questions?
16:37:43 <Rathann> right, no questions from me then
16:38:05 <Rathann> and we're missing votes from sgallagh and jwb
16:38:17 <sgallagh> jwb: https://fedorahosted.org/sssd/wiki/DesignDocs/FilesProvider is a good resource, FWIW
16:38:23 <sgallagh> I voted +1
16:38:37 <sgallagh> Or, I thought I did
16:38:54 <Rathann> not that I can see
16:39:02 <Rathann> but I was expecting a +1 from you anyway
16:39:05 <sgallagh> Well, I just did, then :)
16:39:07 <Rathann> right
16:39:15 <Rathann> jwb?
16:39:25 <jwb> +1 i guess
16:39:32 <Rathann> #agreed Approve the Consider adding SSSD to one of the default comps groups (+1:6, 0:0, -1:0)
16:39:34 <Rathann> great
16:39:41 <sgallagh> #action sgallagh to send a PR for comps
16:39:53 <Rathann> #topic #1688 Incomplete (non testable) Changes of F26
16:39:53 <Rathann> .fesco 1688
16:39:54 <zodbot> Rathann: Issue #1688: Incomplete (non testable) Changes of F26 - fesco - Pagure - https://pagure.io/fesco/issue/1688
16:40:52 <Rathann> wow... that's a lot
16:41:06 <Rathann> probably people forgot to update at least some
16:41:41 <nirik> the static info there is a week old, FWIW
16:41:51 <jforbes> Some are notated that way.
16:42:00 <nirik> but wait, edited 13 hours ago. nevermind
16:42:25 <jwb> i'd suggest just approving the deferal of the ones that wish to move to f27
16:42:36 <jwb> and revisiting the rest of the list next week
16:42:41 <sgallagh_> jwb: +1
16:42:45 <nirik> jwb: +1
16:42:54 <nirik> some of these are clearly done, just not updated... like gcc7
16:43:02 <Rathann> ok, +1
16:44:19 <jforbes> jwb: +1
16:44:23 <dgilmore> hi all sorry was in meetings
16:45:00 <maxamillion> jwb: +1
16:45:28 <jwb> i'll comment the ticket
16:46:03 <Rathann> #agreed Approve deferring the incomplete F26 changes that were asked to be moved to F27, revisit the rest next week (+1:5, 0:0, -1:0)
16:46:06 <Rathann> thanks jwb
16:48:18 <Rathann> #topic #1635 F26 Self Contained Changes
16:48:18 <Rathann> .fesco 1635
16:48:18 <Rathann> there are two changes there
16:48:19 <Rathann> Arm Support In FMW https://fedoraproject.org/wiki/Changes/ArmSupportInFmw and Python 3 C.UTF-8 locale https://fedoraproject.org/wiki/Changes/python3_c.utf-8_locale
16:48:20 <zodbot> Rathann: Issue #1635: F26 Self Contained Changes - fesco - Pagure - https://pagure.io/fesco/issue/1635
16:48:32 <Rathann> jkurik suggested deferring them to F27
16:48:44 <dgilmore> +1 to both
16:49:13 <dgilmore> assuming they are done/testable now
16:49:59 <jforbes> Well, kind of crappy to defer when the reason is we didn't hold a meeting last week.
16:50:17 <Rathann> agreed
16:50:38 <cstratak> Hello
16:50:40 <Rathann> cstratak: what's the status of Python 3 C.UTF-8 locale https://fedoraproject.org/wiki/Changes/python3_c.utf-8_locale ?
16:50:49 <Rathann> is it in testable state?
16:51:07 <sgallagh> jforbes: Well, at the same time if they didn't get the changes in before Alpha freeze...
16:51:12 <cstratak> I haven't pushed the changes in rawhide as the change was not accepted so far, however I have testable builds in copr.
16:51:54 <Rathann> I'm +1 to both for the record, even if they missed the freeze
16:52:16 <cstratak> The change was sent the last day of the deadline but it has to stay for one week at the devel mailing list before it's considered to be discussed
16:52:18 <sgallagh> I'm a little concerned about the Python one, since I'm not sure how much fallout this will have
16:52:33 <sgallagh> I know the *existing* default locale has caused lots of problems in the past
16:52:39 <cstratak> during that time a fesco meeting was cancelled and the alpha freeze came around
16:53:41 <cstratak> sgallagh, I believe everything is mentioned at the change, is it something not clean about it?
16:53:47 <cstratak> clear
16:53:53 <nirik> Looks like FMW has arm changes upstream, but also note it wasn't going to replace the command line arm image as default, just start offering these.
16:54:02 <nirik> so I am +1 to both
16:54:07 <Rathann> #chair dgilmore
16:54:07 <zodbot> Current chairs: Rathann dgilmore jforbes jwb maxamillion nirik rathann sgallagh
16:54:19 <sgallagh> "upgrade/compatibility impact" has nothing, which I doubt is true
16:54:23 <maxamillion> I'm +1 to both
16:54:48 <sgallagh> To be clear, this is something I want to see land, but I'm really concerned about landing it this late in F26
16:54:48 <jforbes> sgallagh: They didn't get the changes in before alpha freeze because they weren't approved last week, because we didn't have a meeting.
16:54:48 <cstratak> sgallagh, well it is true actually
16:56:00 <cstratak> sgallagh, it only affects isolated environments where the locale is not set and only python application during their run time. So essentially it kinda makes things "more" compatible in most aspects
16:56:08 <sgallagh> So things like... Anaconda.
16:56:18 <cstratak> anaconda shouldnt have any issues by it
16:56:34 <sgallagh> Locale issues have plagued anaconda for years.
16:56:36 <Rathann> cstratak: have you tested anaconda with this change?
16:56:38 <sgallagh> It's why I'm so hesitant
16:56:55 <cstratak> and if they do actually I'm in contact as well with anaconda developers for any issues that might arise by changes in the interpreter
16:57:22 <cstratak> so they would be of the first people to actually be aware of this change and test it as soon as it's accepted.
16:57:29 <Rathann> ok...
16:57:29 <sgallagh> cstratak: Right... but since this missed Alpha, we really only have Beta to test and fix issues in anaconda that arise from it
16:57:52 <sgallagh> (for wide testing, I mean)
16:58:41 <Rathann> I'm counting 4 +1 to both changes so far
16:58:46 <nirik> well, we have not yet made a alpha rc...
16:58:54 <jwb> i'm +1 for the arm thing
16:58:56 <nirik> but that would require passing this in...
16:58:56 <cstratak> sgallagh, well while that might hold true in the sense that the change might have a wide effect (which it does not), the change not landing in the alpha was due to scheduling
16:59:20 <cstratak> and I can push the patch even today as it is ready
16:59:27 <sgallagh> cstratak: I know, and I am not happy about that.
16:59:43 <sgallagh> (FWIW, I showed up last Friday hoping we could address these questions then)
17:00:01 <Rathann> #agreed Approve Arm Support In FMW (+1:5, 0:0, -1:0)
17:00:11 <jforbes> +1 to both
17:00:15 <Rathann> #undo
17:00:15 <zodbot> Removing item from minutes: AGREED by Rathann at 17:00:01 : Approve Arm Support In FMW (+1:5, 0:0, -1:0)
17:00:29 <sgallagh> If the rest of FESCo is willing to give it the go-ahead to be a Freeze Exception for Alpha and land there, I guess I could go along with that.
17:00:35 <Rathann> #agreed Approve Arm Support In FMW (+1:6, 0:0, -1:0)
17:00:40 <jforbes> sgallagh: since we have no alpha rc yet, it might be worth asking for a freeze exception
17:00:46 <Rathann> sgallagh: yes, I'd be +1 to that
17:01:50 <cstratak> so would a freeze exception work better in that case?
17:02:07 <sgallagh> It would give us two milestones to get real testing
17:02:12 <dgilmore> we would need a FE to get it in
17:02:17 <sgallagh> Which would make me more comfortable
17:02:18 <dgilmore> which FESCo can do
17:02:26 <nirik> and we should land it asap... so we could see the nightly composes
17:02:28 <sgallagh> Right
17:02:55 <cstratak> that makes sense. I can create the builds and bodhi updates today, however what would be the procedure after that?
17:03:45 <nirik> https://qa.fedoraproject.org/blockerbugs/milestone/26/alpha/buglist
17:03:59 <sgallagh> cstratak: We'd probably file the Change BZ as a freeze-exception
17:04:02 <nirik> need some bug describing the change and then propose it...
17:04:05 <sgallagh> And grant it on FESCo's authority
17:04:16 <Rathann> https://qa.fedoraproject.org/blockerbugs/propose_bug
17:04:33 <cstratak> I already have another bugzilla proposing the change, give me one second
17:04:42 <dgilmore> .fesco 1126
17:04:43 <zodbot> dgilmore: Issue #1126: Need a procedure for tracking FESCo release blockers - fesco - Pagure - https://pagure.io/fesco/issue/1126
17:04:55 <cstratak> https://bugzilla.redhat.com/show_bug.cgi?id=1404918
17:05:32 <Rathann> mbriza: just for the record, is https://fedoraproject.org/wiki/Changes/ArmSupportInFmw in a testable state right now?
17:06:04 <mbriza> Rathann: yes and no... there are testable windows and mac builds, the testing version is not imported to fedora yet
17:06:20 <mbriza> Rathann: i'm going to release 4.1 next week and import it
17:06:25 <Rathann> ok
17:06:33 <dgilmore> mbriza: if you need official builds please file a releng issue
17:06:48 <sgallagh> Proposal: FESCo declares that inclusion of the C.UTF8 locale python Change is an approved Freeze Exception for F26 Alpha
17:07:04 <mbriza> dgilmore: ok... there will be a change to how it is built for both platforms, do you use my scripts for building it?
17:07:22 <dgilmore> mbriza: I have to tweak the mac one
17:07:22 <mbriza> dgilmore: from what i remember you don't do so for mac so i assume i'll have to file another issue for that too, right?
17:07:36 <sgallagh> Actually, amended:  Proposal: FESCo declares that inclusion of the C.UTF8 locale python Change is approved both for Fedora 26 and as a Freeze Exception for F26 Alpha
17:07:36 <nirik> sgallagh: +1
17:07:36 <dgilmore> mbriza: we should fix that
17:07:46 <jforbes> sgallagh: +1
17:07:53 <Rathann> sgallagh: +1
17:08:03 <dgilmore> +1
17:08:08 <sgallagh> +1
17:08:10 <mbriza> dgilmore: ok - we'll need to add xz-libs  to both builds, i'll put some details about how i do it into the releng issue
17:08:20 * nirik really needs to get the macos builders configured and setup in phx2. The todo list never seems to grow shorter tho.
17:08:55 <Rathann> anyone else wants to vote?
17:09:23 <Rathann> jwb, maxamillion ?
17:09:39 <jwb> +1
17:10:01 <cstratak> so I proposed a freeze exception
17:10:07 <cstratak> for the c locale
17:10:32 <Rathann> #agreed FESCo declares that inclusion of the C.UTF8 locale python Change is approved both for Fedora 26 and as a Freeze Exception for F26 Alpha (+1:6, 0:0, -1:0)
17:10:48 <Rathann> ok, that looks like the end of the agenda
17:11:17 <Rathann> #topic Next week's meeting chair
17:11:23 <Rathann> any volunteers?
17:11:45 <sgallagh> I haven't done it in a while.
17:12:03 <Rathann> #action sgallagh to run the next meeting
17:12:08 <Rathann> thank you
17:12:18 <Rathann> #topic Open floor
17:13:03 <Rathann> does anyone have anything else to discuss?
17:13:21 <Rathann> if not, I'll end the meeting in about 2 minutes
17:15:22 <Rathann> #endmeeting