18:01:15 <paragan> #startmeeting FESCO (2015-10-14)
18:01:15 <zodbot> Meeting started Wed Oct 14 18:01:15 2015 UTC.  The chair is paragan. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:01:15 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
18:01:15 <paragan> #meetingname fesco
18:01:15 <zodbot> The meeting name has been set to 'fesco'
18:01:15 <paragan> #chair ajax dgilmore number80 jwb nirik paragan rishi thozza sgallagh
18:01:15 <paragan> #topic init process
18:01:15 <zodbot> Current chairs: ajax dgilmore jwb nirik number80 paragan rishi sgallagh thozza
18:01:23 <rishi> .hello rishi
18:01:24 <zodbot> rishi: rishi 'Debarshi Ray' <debarshir@redhat.com>
18:01:28 <sgallagh> .hello sgallagh
18:01:29 <zodbot> sgallagh: sgallagh 'Stephen Gallagher' <sgallagh@redhat.com>
18:01:51 <jkurik> .hello jkurik
18:01:54 <zodbot> jkurik: jkurik 'Jan Kurik' <jkurik@redhat.com>
18:01:56 <jwb> hi
18:02:18 <nirik> morning
18:02:58 <paragan> Hi all
18:03:11 <paragan> dgilmore, number80 ajax around for this meeting?
18:03:13 <ajax> hi
18:04:01 <paragan> #topic #615 Strategy for services that do not have systemd native unit files
18:04:01 <paragan> .fesco 615
18:04:01 <paragan> https://fedorahosted.org/fesco/ticket/615
18:04:03 <zodbot> paragan: #615 (Strategy for services that do not have systemd native unit files) – FESCo - https://fedorahosted.org/fesco/ticket/615
18:04:05 <number80> o/
18:04:58 <paragan> I think proposal is already in the ticket
18:05:07 <jwb> 0
18:05:15 <nirik> which proposal? what comment?
18:05:17 <paragan> proposal: https://fedorahosted.org/fesco/ticket/615#comment:69 along with updating the SysVInitScript related packaging guidelines.
18:06:12 <dgilmore> hola
18:06:49 <ajax> one sec while i dig up a list of packages...
18:06:53 <nirik> I guess I can be ok with that...
18:07:27 <nirik> might mean the end of 'network', but I guess they could ask for an exception if they want to keep it around.
18:07:46 <paragan> yes
18:08:14 <number80> +1
18:08:22 <paragan> if above proposal given by me looks okay can we have votes please?
18:08:26 <dgilmore> sgallagh is wrong in the proposal as to when the branch point is
18:08:27 <number80> (I had to dig it but makes sense)
18:08:35 <rishi> paragan: I am +1 to sgallagh 's propsal in comment 69.
18:08:36 <nirik> also, the iptables package and it's iptables-services might need an exception
18:08:38 <dgilmore> but I am okay with branch point
18:08:39 <sgallagh> dgilmore: Right, I realized that later, but yeah.
18:08:57 <sgallagh> Hang on, though. I'm going to post the proposal in here so it's clear.
18:09:34 <sgallagh> One question: I originally proposed F25, but I'm amenable to making it F24 if folks are good with that.
18:10:09 <nirik> can we have 2 datapoints before deciding that?
18:10:19 <dgilmore> nirik: sure
18:10:20 <nirik> when is that and how many packages are left on the list? :)
18:10:55 <number80> and it needs a heads-up on announce-list
18:11:15 <nirik> devel-announce?
18:11:22 <dgilmore> nirik: f24 branch point is feb 2
18:11:52 <dgilmore> nirik: f25 would be around early July
18:12:01 <nirik> ok.
18:12:16 <sgallagh> Proposal: http://fpaste.org/279216/44846322/
18:12:35 <sgallagh> But yes, let's get the list of remaining packages.
18:12:48 <nirik> might put it in the channel... fpastes expire...
18:13:07 <sgallagh> Proposal:
18:13:07 <sgallagh> Any package in Fedora that provides a SYSV init script instead of a systemd service unit will be retired from the distribution at the branch point of Fedora 24. Exceptions will be considered on a case-by-case basis if submitted through a FESCo ticket at least three weeks prior to the branch point. FESCo will coordinate with the Change Wrangler to provide periodic reminders of this deadline mailed out. FESCo explicitly does not advoc
18:13:27 <sgallagh> (Did that get split by my client or truncated?)
18:13:31 <nirik> Why remove the docs? that seems poor...
18:13:32 <jwb> truncated
18:13:33 <nirik> yes.
18:13:35 <dgilmore> sgallagh: truncated
18:13:42 <gholms> It ended at advocate
18:13:46 <jwb> "FESCo explicitly does not advoc"
18:14:00 <sgallagh> FESCo explicitly does not advocate the removal of the sysvinit compatibility functionality from systemd itself, as we do not want to negatively impact third-party software that may still be relying upon it. We will also remove all packaging documentation around the use of SYSV init scripts.
18:14:15 <jwb> i'm still 0 on the whole thing
18:14:32 <dgilmore> I think the docs need to stay until RHEL 7 is EOL
18:14:43 <nirik> I can be +1 if we drop the docs thing... IMHO we should just put a disclaimer on it, but people still might find the docs of use.
18:14:43 <zbyszek> sgallagh: I think the documentation should be kept around, just prefixed with a banner.
18:14:51 <sgallagh> OK, I'm fine with revising that to state that they are not recommended for use in Fedora.
18:14:55 <ajax> eh.  with my employer's hat on, i'd like to get everything corresponding to a package in rhel7 done, but i suspect we're pretty close to that anyway
18:14:58 <sgallagh> zbyszek: Fine by me
18:14:58 <paragan> yes docs should not be removed
18:15:06 <dgilmore> nirik: docs will be needed for EPEL 7
18:15:14 <dgilmore> so they have to stay
18:15:24 <dgilmore> but give a bit fat warning
18:15:28 <nirik> dgilmore: well, possibly. :) if it's a new epel7 package it could use systemd units
18:15:34 <ajax> (still grepping all the specs for _initrddir vs _unitdir)
18:15:41 <nirik> but yes, I have pointed people to those docs before for rhel stuff.
18:16:00 <sgallagh> OK, amended proposal:
18:16:00 <sgallagh> Any package in Fedora that provides a SYSV init script instead of a systemd
18:16:00 <sgallagh> service unit will be retired from the distribution at the branch point of
18:16:00 <sgallagh> Fedora 24. Exceptions will be considered on a case-by-case basis if submitted
18:16:01 <sgallagh> through a FESCo ticket at least three weeks prior to the branch point. FESCo
18:16:02 <sgallagh> will coordinate with the Change Wrangler to provide periodic reminders of this
18:16:03 <sgallagh> deadline mailed out. FESCo explicitly does not advocate the removal of the
18:16:04 <sgallagh> sysvinit compatibility functionality from systemd itself, as we do not want to
18:16:05 <sgallagh> negatively impact third-party software that may still be relying upon it. We
18:16:06 <sgallagh> will also update all packaging documentation around the use of SYSV init
18:16:07 <sgallagh> scripts to indicate that its use in Fedora is not recommended, but that it may
18:16:08 <sgallagh> be useful for EPEL.
18:16:08 <dgilmore> nirik: sure, but packages can not switch, and people willneed to understand why things are, as tehy are
18:16:16 <nirik> dgilmore: yep.
18:16:28 <nirik> sure, +1 to that.
18:16:30 <dgilmore> even if i can not spell
18:16:34 <paragan> +1
18:16:40 <sgallagh> +1
18:16:50 <rishi> +1
18:17:29 <dgilmore> sgallagh: it will still be necessary for Fedora until F23 is EOL
18:17:38 <dgilmore> assuming we adopt it for f24
18:17:43 <dgilmore> the docs that is
18:18:02 <dgilmore> I would like to see teh potentially effected packages
18:18:39 <ajax> +1
18:18:49 <jwb> that's 5
18:19:12 <dgilmore> but I am generally +1 I just wonder if there are cases we should give an exception for up front, like iptables
18:19:40 <paragan> #agreed to sgallagh's modified proposal based on https://fedorahosted.org/fesco/ticket/615#comment:69 (7, 1, 0)
18:19:52 <nirik> ha. also, our livecd's use sysvinit scripts. ;)
18:19:56 <sgallagh> dgilmore: The proposal includes the mechanism for requesting exceptions. I think it's fine to let the maintainer come to us.
18:20:18 <dgilmore> sgallagh: okay
18:20:18 <nirik> dgilmore: looks like sigul is one, we may need to fix it or get it an exception.
18:20:26 <ajax> packages with specs matching _initrddir but not _unitdir: http://paste.fedoraproject.org/279219/44846797/
18:20:29 <dgilmore> nirik: we will need an exception
18:20:33 <dgilmore> I believe
18:20:48 <paragan> #info sgallagh to work on updating SySV init packaging guidelines update
18:20:49 <ajax> all of 51 packages, most of them fairly fringe
18:20:57 <nirik> libreport is a biggie I think...
18:20:59 <rishi> dgilmore: What's up with iptables?
18:21:19 * nirik sighs at opentracker. Will need fixing
18:21:37 <paragan> let's move to next ticket
18:21:42 <paragan> #topic #1486 FESCo elections for F24/F25
18:21:43 <paragan> .fesco 1486
18:21:43 <paragan> https://fedorahosted.org/fesco/ticket/1486
18:21:44 <zodbot> paragan: #1486 (FESCo elections for F24/F25) – FESCo - https://fedorahosted.org/fesco/ticket/1486
18:22:07 <nirik> rishi: it has a iptables-services subpackge that is the old sysvinit scripts for starting/stopping/saving things.
18:22:27 <paragan> as suggested in https://fedorahosted.org/fesco/ticket/1486#comment:3
18:22:30 * nirik is fine with whatever jkurik wants to do here to track elections. Or whoever is going to run them.
18:22:46 <paragan> proposal: Have one ticket tracking elections for all governing bodies, opened in Council's ticketing system as Council is the top-level governance body.
18:22:47 <jwb> nirik, same
18:22:49 <dgilmore> I am with nirik on this
18:23:02 <jwb> paragan, +!
18:23:03 <jwb> er
18:23:05 <jwb> +1
18:23:12 <rishi> nirik: Can't we drop it? The reason I ask is because dgilmore made it sound like making iptables systemd-only is a lost cause.
18:23:16 <paragan> +1 to my proposal
18:23:19 <number80> +1 (no need to discuss this further from my PoV)
18:23:26 <dgilmore> +1
18:23:34 <nirik> sure, +1 whatever works
18:23:37 <rishi> +1
18:23:56 <sgallagh> +1
18:24:17 <dgilmore> rishi: I was talking about sigul
18:24:34 <nirik> rishi: people still use it. I don't think it's possible to convert, just the way that the old scripts worked. IMHO it should just be excepted.
18:24:41 <paragan> ajax, your vote?
18:25:13 <ajax> +1
18:25:17 <paragan> #agreed Have one ticket tracking elections for all governing bodies, opened in Council's ticketing system as Council is the top-level governance body. (9, 0, 0)
18:25:32 <jkurik> ok, thanks. I will inform Council that we are going to track elections in the Council's tracking system
18:25:46 <paragan> jkurik, thanks
18:25:49 <paragan> #topic #1489 F24 System Wide Change: NetworkManager 1.2
18:25:50 <paragan> .fesco 1489
18:25:50 <paragan> https://fedorahosted.org/fesco/ticket/1489
18:25:51 <zodbot> paragan: #1489 (F24 System Wide Change: NetworkManager 1.2) – FESCo - https://fedorahosted.org/fesco/ticket/1489
18:25:55 <rishi> nirik: I see, ok. I never looked inside those scripts. Thanks.
18:26:09 <nirik> +1
18:26:14 <sgallagh> +lots
18:26:20 <paragan> +1
18:26:28 <ajax> ooh, tui for vpn, +1
18:26:31 <number80> +1
18:26:32 <rishi> +1
18:26:56 <jwb> +!
18:27:02 * jwb curses his shift key
18:27:23 <sgallagh> jwb: YOU MISSED. NOW MINE IS STUCK.
18:28:04 <paragan> #agreed Approved F24 System Wide Change: NetworkManager 1.2 (8, 0, 0)
18:28:16 <paragan> #topic #1490 restore MPI Requires Provides as F23 change
18:28:16 <paragan> .fesco 1490
18:28:16 <paragan> https://fedorahosted.org/fesco/ticket/1490
18:28:19 <zodbot> paragan: #1490 (restore MPI Requires Provides as F23 change) – FESCo - https://fedorahosted.org/fesco/ticket/1490
18:29:03 <number80> since it's done and already approved for F24 ...
18:29:05 <nirik> this is unfortunate...
18:29:15 <number80> +1
18:29:17 <nirik> but we should advertise it if it's done.
18:29:20 <paragan> As requested in the ticket description
18:29:23 <paragan> proposal: Mark the Change https://fedoraproject.org/wiki/Changes/RpmMPIReqProv as completed and include in https://fedoraproject.org/wiki/Releases/23/ChangeSet
18:30:27 <rishi> If it is done, it is done. We should add it back. +1
18:30:31 <dgilmore> why did we drop it from f23?
18:30:36 <rishi> nirik: Why "unfortunate" ?
18:30:45 <nirik> dgilmore: they never answered that it was done, or replied to any progress...
18:30:53 <paragan> dgilmore, its not marked as MODIFIED at time of deadline
18:30:55 <nirik> rishi: that they didn't tell us it was done. ;)
18:31:06 <paragan> well it was completed one week later I guess
18:31:23 <sgallagh> I'm really not sure why this is worth advertising anyway. At a glance it doesn't seem useful except to other packagers.
18:31:53 <dgilmore> I think its unfortunate that tehy failed to communicate but I think we shoudl promote it for f24
18:32:12 <dgilmore> i think i should work on my spelling
18:32:13 * nirik is +1 to just advertising it
18:32:17 <ajax> +1 i guess
18:32:19 <zbyszek> sgallagh: It is user visible, it allows automatic installation of dependencies to work properly for MPI packages.
18:32:19 <number80> sgallagh: very useful for scientists who have to deal with MPI (yes, it's a niche)
18:33:10 <sgallagh> number80: I didn't say it wasn't useful. I said it doesn't look like something an end-user would care about. Just a packager.
18:33:14 <jwb> +1.  let's not let process stand in the way of work actually being done.
18:33:27 <sgallagh> But I'm 0 on this. I'm not really fond of the end-run on the Change process
18:33:42 <paragan> dgilmore your vote please?
18:33:44 <number80> *nods* about the latter
18:33:46 <dgilmore> I am 0 to -1
18:34:00 <dgilmore> i guess 0
18:34:04 <paragan> thanks
18:34:26 <paragan> #agreed Mark the Change https://fedoraproject.org/wiki/Changes/RpmMPIReqProv as completed and include in https://fedoraproject.org/wiki/Releases/23/ChangeSet (7, 2, 0)
18:34:41 <jkurik> paragan: I will do this
18:35:23 <paragan> jkurik, thanks
18:35:27 <paragan> #topic #1491 clarifications/improvements for new bundling policy
18:35:28 <paragan> .fesco 1491
18:35:28 <paragan> https://fedorahosted.org/fesco/ticket/1491
18:35:29 <zodbot> paragan: #1491 (clarifications/improvements for new bundling policy) – FESCo - https://fedorahosted.org/fesco/ticket/1491
18:35:50 <nirik> I'm ok deferring this if folks want more discussion... (although I will be gone next week)
18:36:15 <jwb> defer
18:36:27 <dgilmore> I think defer also
18:36:34 <paragan> I too think defer
18:36:35 <ajax> haven't had the chance to go over this yet, so defer
18:36:42 <nirik> I think we may want to also make a wiki page for this policy so we have a thing to point people at and propose changes to.
18:36:46 <ajax> (been busy writing actual tools for this)
18:37:41 <sgallagh> Yeah, let's discuss this in the ticket some more
18:37:46 <nirik> I will note as I did in the ticket we have a thing that already has checksums for all source files if it helps ay of these efforts
18:37:53 <number80> yeah
18:38:07 <sgallagh> nirik: Yeah, that was good to know
18:38:14 <rishi> Yeah, I haven't had time to read the follow-up thread on fedora-devel either.
18:38:19 <rishi> Let's defer.
18:38:27 <paragan> #agreed defer ticket 1491 for next week meeting
18:38:29 <number80> rishi: lucky you
18:38:29 <nirik> there was a GSoC to make a web interface to it, but not sure where that is
18:38:30 <sgallagh> nirik: Does that include the resultant binaries, or only the sources?
18:38:44 <nirik> just the sources. it's done at upload time
18:38:50 <sgallagh> ok
18:39:09 <sgallagh> That's very good to know about, though
18:39:55 <nirik> rpm / koji should have the binaries... at least any shipped
18:40:32 <paragan> #topic Next week's chair
18:40:47 <sgallagh> nirik: Actually, I was thinking of the hashes of that output. But htat probably still needs to happen in RPM itself
18:40:55 <sgallagh> If we want to de-duplicate, I mean
18:41:15 <sgallagh> I may not be able to attend next week.
18:41:18 <paragan> Who want to chair for next week meeting?
18:41:27 <nirik> I will not be here next week.
18:42:05 <rishi> Ok, I will do it.
18:42:13 <paragan> rishi, thanks
18:42:14 <number80> thanks
18:42:29 <paragan> #info rishi to chair next to next week
18:42:39 <paragan> #topic Open Floor
18:42:51 <paragan> anyone want to discuss anything here?
18:43:50 <number80> I guess no
18:44:12 <dgilmore> do we want to have a timeline on conversion of everything to dnf?
18:44:19 <dgilmore> or do we not care?
18:44:32 <sgallagh> We care
18:44:43 <number80> we do
18:44:49 <paragan> what is remaining to be converted?
18:44:59 <dgilmore> all the releng tools
18:45:36 <dgilmore> we are working on setting priorites and roadmaps for the releng tools at the moment
18:45:44 <sgallagh> So that's like, what, a couple hours' work? ;-)
18:45:48 <nirik> har
18:45:48 <dgilmore> two big things we have is python3 dnf
18:45:54 <dgilmore> we have not started on either
18:46:24 <dgilmore> sgallagh: sure, and the sky is green and the grass blue
18:46:24 <nirik> dgilmore: for koji would that be a new 1.x release or be work for 2.0?
18:46:44 <dgilmore> nirik: there is talk that 2.x will be python3 only
18:46:52 <nirik> seems reasonable.
18:46:56 <dgilmore> but it may mean porting parts of koji
18:47:13 <dgilmore> the main library and the cli
18:47:36 <dgilmore> I just ask because we switched the rawhide buildroot last week
18:47:42 <dgilmore> we have hit a few bugs
18:47:50 <nirik> yeah...
18:48:27 <nirik> I think we all agree it would be great to get there, but hard to set a deadline when we don't know how much work it would be and if anyone is commited to doing it.
18:48:53 <paragan> is there any plans for Koji 2.x release? when will it be available?
18:49:09 <nirik> it's still in the wishlist/design phase as far as I know.
18:49:11 <dgilmore> paragan: it is just in requirements gathering phase
18:49:17 <dgilmore> so its very unkonwn
18:49:29 <paragan> oh okay
18:49:29 <number80> I would be fine with a previsional timeline (no hard deadline)
18:50:06 <number80> so people would know what's in progress, and remaining tasks
18:50:35 <dgilmore> we (releng) have been changing how we do things, part of it is setting out stimelines and tracking what we have on our plate.  We hope that soon we will have something that will show people what we are doing
18:51:35 <dgilmore> I would like to know if we (FESCo or fedora community) is going to have some push to prioritise removal of yum or python2 that we need to ensure the work is a higher priority or not
18:52:58 <nirik> it's really hard to say without seeing the lisrt
18:53:02 <paragan> I am not sure about python2 removal
18:53:17 <dgilmore> some of the code we use has not been touched in over 3 years. I am trying to get owners for it to take it forward
18:53:35 <nirik> I mean yes, we want this done, but if it's at the cost of other features we want more, then we would prioritize accordingly.
18:53:35 <ajax> i don't personally have an agenda on that point i suppose
18:53:55 <ajax> i do think it's disingenuous to say fedora is "python3 by default" if we can't bootstrap our own tools without python2
18:54:03 <ajax> but, okay, fine, so we don't say that.
18:55:46 <dgilmore> okay. I will keep it high, but try to keep a balance
18:55:56 <dgilmore> the priority high that is
18:58:37 <paragan> okay so if nothing more to discuss here we can close in 2 minutes
18:58:39 <dgilmore> I have nothing else
18:58:45 <paragan> dgilmore, thank you for bringing this topic here and let us aware about releng plans and issues you are facing. Hope to see those resolved soon.
18:59:04 <dgilmore> paragan: we will see. there is a lot to do.
19:00:16 <paragan> #endmeeting