17:00:19 #startmeeting FESCO (2014-05-07) 17:00:19 Meeting started Wed May 7 17:00:19 2014 UTC. The chair is mitr. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:19 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:00:21 #meetingname fesco 17:00:21 The meeting name has been set to 'fesco' 17:00:23 #chair abadger1999 dgilmore mattdm mitr notting nirik pjones t8m sgallagh mmaslano jwb 17:00:23 Current chairs: abadger1999 dgilmore jwb mattdm mitr mmaslano nirik notting pjones sgallagh t8m 17:00:25 #topic init process 17:00:27 Hello everyone 17:00:33 greetings 17:00:34 Hi 17:00:58 hi 17:01:13 hola 17:01:46 greetings, starfighter. 17:02:14 * notting is here 17:02:59 * mclasen lurks 17:03:41 That's 6(+), shall we start? 17:03:49 #topic #1221 Product working group activity reports 17:04:28 #info see activity reports starting at https://fedorahosted.org/fesco/ticket/1221#comment:51 17:04:35 Anything to discuss here? 17:05:45 seems no 17:06:03 #topic #1291 F21 System Wide Change: BerkeleyDB 6 - https://fedoraproject.org/wiki/Changes/BerkeleyDB_6 17:07:03 AFAICS our original concern has been resolved, but today questions about the plan (add versioning to the new major version) have been raised again 17:07:44 Proposal: BerkeleyDB 6 is approved 17:07:54 (that is: approved as written, we can revisit if the plan changed) 17:08:58 i'm a bit confused. the feature owner said on 4/15 that they were planning to contact upstream about symbol versioning 17:09:05 I'm okay with in in theory but it seems like the plan is evolving in unknown directions. 17:09:28 in the mail today they were asking for suggestions, and saying they were still planning to contact upstream about symbol versioning. can they just please do that? 17:09:39 notting: good catch 17:09:49 .fesco 1291 17:09:50 t8m: #1291 (F21 System Wide Change: BerkeleyDB 6 - https://fedoraproject.org/wiki/Changes/BerkeleyDB_6) – FESCo - https://fedorahosted.org/fesco/ticket/1291 17:10:00 i'm missing why this has to be done by f21 17:10:13 jwb: it doesnt 17:10:16 seems like it's a new version add, they're keeping the old version, so punt at this point until they get it cleared up 17:10:29 indeed. 17:10:46 if they can not cleanly implement the plan then they should hit the contingency plan 17:10:59 dgilmore, +1 17:11:10 that's my point though 17:11:14 it's not in-distro yet, is it? 17:11:20 if not, then don't even add it for f21 17:11:20 They have until 2014-07-22 (alpha freeze) to decide to revert, though, yes? 17:11:24 we do not have to reject the feature just now 17:11:33 t8m: right 17:11:36 Unless we think that the plan is not implementable in that timeframe. 17:12:00 abadger1999, and I don't think we think that :) 17:12:06 abadger1999: sure, but... why bother thinking that? Either it is and they succeed... 17:12:49 pjones: only possibly because rollback of partial implementation would be a messy 17:13:10 notting, which is why i'm saying don't add it at all until the plan is clearly doable 17:13:36 fair enough 17:13:39 pjones: . so I'm unclear as to why we're talking about reverting it at this point in the schedule. 17:14:09 abadger1999: yeah, I don't think we need to be talking about /reverting/, but it is in /ChangesReadyForFesco/, so we need to talk about if they can do it at all and if we should finally approve the thing or not. 17:14:14 abadger1999, it's not a revert at this point 17:14:27 a revert would be messy as notting says 17:15:18 proposal: suggest owner contact upstream now about symbol versioning, and report back by next week? 17:15:29 * mattdm is here -- sorry, previous thing ran late 17:16:33 I could live with that.... I'm thinking that we've done a non-upstream versioning change before so it's something that clearly can be done and we don't strictly need to block on a precise plan, but a little prodding to get the upstream discussion working might be helpful 17:16:34 and if upstream doesn't agree to symbol versioning by next week, feature owner should make a proposal about how they're planning to proceed. 17:16:44 mitr, +1 17:16:55 notting: +1 I think 17:17:07 t8m: is that a +1 to notting's proposal? 17:17:10 notting+abadger1999, +1 17:17:16 notting: +1 17:17:20 mitr, yep with the abadger1999's addition 17:17:31 notting: +1 17:17:49 notting: abadger1999: +1 17:18:21 abadger1999: +1 17:18:39 #agreed Please contact upstream now about symbol versioning, and report back by next week; if upstream doesn't agree, have a proposed approach (+7) 17:18:55 (I hope I'm not misrepresenting anybody) 17:19:04 #topic #1297 F21 System Wide Change: Workstation: Enable Software Collections - https://fedoraproject.org/wiki/Changes/Workstation_Enable_Software_Collections 17:19:06 .fesco 1297 17:19:07 mitr: #1297 (F21 System Wide Change: Workstation: Enable Software Collections - https://fedoraproject.org/wiki/Changes/Workstation_Enable_Software_Collections) – FESCo - https://fedorahosted.org/fesco/ticket/1297 17:19:20 i have no updates on this. 17:19:30 I pinged mclasen 17:19:35 I'm here 17:19:40 tada :) 17:19:57 mclasen was poking cschalle to fill this out a while ago... 17:20:00 I've looked over the two feature pages, and I think we're really going after the same goal here 17:20:16 make it easy to use and work with software collections in fedora 17:20:44 mclasen: sure. what is the source of the scls you are proposing 17:20:51 and what is to be enabled by default? 17:21:29 https://fedorahosted.org/fesco/ticket/1297#comment:3 has some specific questions 17:21:31 are there other sources beside softwarecollections.org ? 17:21:42 mclasen: shipped in fedora 17:21:47 Fedora itself will have a few scls for F21. 17:21:55 I'm not an expert on software collections by any measure... 17:21:56 mclasen, is it fair to say Workstation is looking to enable whatever scl-utils in fedora enables by default? 17:22:02 if the intention is to enable softwarecollections.org by default that needs to be made clear 17:22:03 yeah, basically 17:22:44 so i think the answer is "install scl-utils in workstation and have it use whatever is enabled by default in fedora" 17:22:50 mclasen: yeah basically to what exactly? 17:22:50 mclasen: "softwarecollections.org" and "collections in Fedora" are _not_ the same thing, so "yeah" is not clarifying 17:23:14 ok, that was what I was asking about earlier, sorry 17:23:41 proposal: merge this into https://fedoraproject.org/wiki/Changes/SCL, and add a note about workstation there (and possible to the release notes) 17:23:48 there is already such a note for cloud 17:23:49 hmm... softwarecollections.org is essentially copr repos... so it doesn't seem like enablig by default would fit under the current polocy. I think enabling from devassistant would, though. 17:24:03 what does 'enable by default' even mean ? 17:24:11 we have 4 people thinking 4 different things at this point 17:24:11 the typical scl command is 'scl enable', right ? 17:24:14 I do not think scl-utils is a software installer. 17:24:33 so you enable them when you use them 17:24:34 mattdm: That could work but may end up just enabling playground AFAICT 17:24:38 so talking about repos in the context of scl0utils doesn't make a lot of sense. 17:24:46 the jist of this feature is "WORKSTATION WANTS TO USE SCLs WHATEVER THAT MEANS TELL US PLEASE" 17:25:24 mclasen: So scl enable can be used to setup your environment variables to make use of a specific scl instead of system packages. 17:25:33 mclasen: "enabled by default" are your words in the Change so we're basically asking _you_ what that means... 17:25:40 jwb right, which is why I think merging into the previous change makes sense 17:25:48 s/system/packages/packages installed in the main system directrory hierarchy/ 17:25:57 mattdm: +1 that seems most practical if there are no more specific desires 17:26:07 mattdm: i think merging into the SCL change makes sens 17:26:08 mclasen: The scl has to be installed by some other mechanism. 17:26:08 sense 17:26:18 mattdm, +1 17:26:36 mclasen: Probably the way that people will install them is via yum. 17:27:17 mattdm: +1 17:27:58 mclasen does that suggestion make sense to you? (note that the previous change is alreay accepted, if that makes any difference) 17:28:02 sounds good to me too, fwiw; as we gain more experience with software collections, it'll probably become clearer what we actually wanted... 17:28:37 mattdm, mclasen: Note that _if_ the /SCL change ends up Playground-only, it would either mean Workstation not having SCLs "by default", or Workstation pulling in all of Playground by default. 17:29:21 is there a way to have the same SCLs available that are available for CentOS & RHEL? 17:29:30 for once, i think Workstation isn't trying to drive the content definiton here 17:29:30 ... and having Playground enabled in a project by default would be rather likely to break something. 17:29:51 without pulling in playground, I mean, but installable via Yum 17:30:08 e.g., ruby193, python27, python33, etc. 17:30:30 "the same" doesn't make sense if the default python is already 3.3 17:30:33 jzb: Only if someone rebuilds them for Fedora-N. 17:30:39 jwb: sure 17:30:51 jwb: actually... it does. 17:30:59 mitr: I think it is clear by now that 'enable by default' was an unfortunate choice of words; the goal is to have a straightforward way to use software collections - it is perfectly fine (and in fact expected) if you have to install something first 17:31:06 I hope SCL change does not end up Playground-only - I'd expect some scls more "mature" than playground level 17:31:14 because the scl-python33 will be a different stack than the main system python3.x 17:31:23 t8m: Me too but that eventuality is AFAICT very much on the table 17:31:31 mitr, I know 17:31:38 mclasen: OK... 17:31:50 mclasen *nod* 17:32:05 jzb: Also -- not by default; the user would have to enable some other repository to get the packages. 17:32:12 I've seen 4 people in agreement with mattdm's idea of merging them. mattdm: care to make it a formal proposal? 17:32:21 mattdm: +1 17:32:59 mitr: I thought that was formal enough? 17:33:15 i'm +1 to the merge 17:33:49 proposal (restated for the record): merge workstation SCL change into https://fedoraproject.org/wiki/Changes/SCL, and add a note about workstation there 17:34:08 mattdm: works for me... 17:34:15 #agreed merge workstation SCL change into https://fedoraproject.org/wiki/Changes/SCL, and add a note about workstation there (+6) 17:34:38 #topic #1303 Add exception for enabling nss-pam-ldapd's nslcd to start by default in obsoleting-install cases 17:34:40 .fesco 1303 17:34:42 mitr: #1303 (Add exception for enabling nss-pam-ldapd's nslcd to start by default in obsoleting-install cases) – FESCo - https://fedorahosted.org/fesco/ticket/1303 17:35:14 I was +1 in the ticket, restating for the record 17:35:22 +1 looks reasonable. 17:35:29 makes sense - +1 17:36:02 +1 17:36:16 +1 17:36:47 +1 17:36:49 +1 17:36:58 #agreed Exception for enabling nss-pam-ldapd's nslcd approved (+7) 17:37:04 #topic #1304 replacing reSIProcate in Fedora 20 17:37:06 .fesco 1304 17:37:07 mitr: #1304 (replacing reSIProcate in Fedora 20) – FESCo - https://fedorahosted.org/fesco/ticket/1304 17:37:22 * abadger1999 was +1 in ticket 17:37:39 +1 this looks like a reasonable exception 17:37:55 +1 17:37:55 As a VoIP geek, it's got my +1 17:38:47 +1 17:38:55 it should sit in updates-testing for a couple of weeks and not have a low auto-karma threshold 17:38:59 +1 17:39:19 +1 17:39:31 perhaps have autokarma turned off 17:39:48 +1, also noting pocock is one of the upstream authors 17:40:03 #agreed reSIProcate update policy exception approved (+7) 17:40:26 #info Please let the packages stay in updates-testing for an extended period, and consider turning off autokarma 17:40:34 #topic #1305 F21 System Wide Change: Application Installer Continued - https://fedoraproject.org/wiki/Changes/AppInstallerContinued 17:40:35 jsmith: please refrain from voting, as voting is restricted to elected fesco members and others voting leads to count confusion. 17:40:36 .fesco 1305 17:40:37 mitr: #1305 (F21 System Wide Change: Application Installer Continued - https://fedoraproject.org/wiki/Changes/AppInstallerContinued) – FESCo - https://fedorahosted.org/fesco/ticket/1305 17:41:11 pjones: No worries 17:41:35 notting: Were there other questions besides my note on relnotes? IMHO they need improving but aren't a blocker to our decision. 17:41:44 mitr: no, just those 17:42:16 proposal: approve, ask for help from docs team to improve relnotes 17:42:24 to which I am +1 17:42:34 +1 17:43:08 I don't think it's strictly necessary to get external help but good enough, +1 17:43:34 +1 17:43:59 I am still waiting on patches to integrate creating the metadata needed into createrepo 17:44:24 today we have no way to make the metadata needed 17:44:32 mattdm: +1 17:45:24 dgilmore: that's accounted for in the contingency plan AFAICS 17:45:38 Was choice of backends discussed to death on the list? 17:45:39 mitr: its not listed at all 17:45:58 dgilmore: "If application metadata extraction and hosting is not getting implemented by F21" ... 17:46:04 mitr: its not listed as a dependency or in the contingency plan 17:46:29 dgilmore looks like you are talking about https://fedorahosted.org/rel-eng/ticket/5721 ? 17:46:30 abadger1999: I have privately reached out to jzeleny and he didn't raise any objections, and again it's in the contingency plan 17:47:09 mattdm: yes, and I talked with Richard at devconf and laid out what has to happen 17:47:23 (it was discussed to death^W^Wto lack of followups in a fesco ticket https://fedorahosted.org/fesco/ticket/1148 ) 17:47:25 mattdm: he commited to doing teh work but i have seen no progress. 17:47:42 dgilmore it looks kind of like that ticket is waiting for a reply? 17:47:58 mattdm: its not. 17:48:12 mattdm: the reply was in person 17:48:26 mattdm: we sat down and discussed what we need and why 17:49:21 dgilmore: So if nobody does that work, it won't get done for F21. Is that good enough for FESCo? 17:49:32 mitr: k. So +1 contingent on jzeleny/packaging team doesn't have any objection to the backend. 17:49:39 #agreed Application Installer Continued was approved (+6) 17:49:42 mitr: I guess. 17:49:51 dgilmore: may be worth adding a note to that effect, and maybe include a rough summary of the decision 17:49:53 ok. it might be helpful to give a quick poke in the ticket. but, as long as it's all worked out... what mitr said. 17:50:20 #topic #1306 F21 System Wide Change: Wayland - https://fedoraproject.org/wiki/Changes/Wayland 17:50:23 .fesco 1306 17:50:24 mitr: #1306 (F21 System Wide Change: Wayland - https://fedoraproject.org/wiki/Changes/Wayland) – FESCo - https://fedorahosted.org/fesco/ticket/1306 17:50:59 so very +1 17:51:03 +! 17:51:05 +1 17:51:08 +1 17:51:16 +1 17:51:27 +1 17:52:07 +1 17:52:50 #agreed Wayland change approved (+6) 17:52:58 #topic #1307 F22 System Wide Change: Default Local DNS Resolver - https://fedoraproject.org/wiki/Changes/Default_Local_DNS_Resolver 17:53:01 .fesco 1307 17:53:02 mitr: #1307 (F22 System Wide Change: Default Local DNS Resolver - https://fedoraproject.org/wiki/Changes/Default_Local_DNS_Resolver) – FESCo - https://fedorahosted.org/fesco/ticket/1307 17:53:03 Note, this is targeted at F22 17:53:06 sorry, I was away for a bit - so, Richard has done quite a lot of work on the tools 17:53:10 * sgallagh is here (better late than never) 17:53:34 * abadger1999 late +1 to wayland 17:53:44 mclasen: we seem to have approved it despite his efforts ;) 17:53:52 #agreed Wayland change approved (+7) 17:55:23 I am not sure that we need to approve F22 changes just now, but +1 anyway as I'd like to see better DNSSEC support in Fedora as soon as possible 17:55:27 ok, so, dns resolver.... should we just wait until f22? 17:55:49 It sounds like it could get release noted as a tech preview. 17:55:53 (for f21) 17:55:56 No, I think this is one of those long-term tasks that we should green-light (or not) early 17:55:58 mattdm: i think we should wait until we have branched f21 17:55:59 I have unanswered concerns about how this would work with Fedora Cloud (let alone docker) 17:56:02 I think if anybody has major objections it'd be good to bring those up now so people aren't wasting time on it 17:56:03 I think having early discussion of large-scale plans like this isgenerally beneficial 17:56:34 there's the 'something needs to be done about the docker case' problem. it would be nice to have the plan idea in hand for that 17:56:41 mattdm: I agree, having final approval blocked on having a Cloud/Docker plan would make sense 17:56:52 discussing and talkingabout it resolving conflicts is great. we should wait to vote until they can actually implement I think 17:57:00 annnnd, I need to run to another meeting. 17:57:07 * mattdm runs 17:57:52 To separate concerns: 17:57:54 Proposal 1: FESCo doesn't have significant objections to the idea as such, or its use on physical machines and VMs 17:57:56 notting, the plan could be to just not have it in cloud product enabled by default - the current default way would have to work still 17:57:56 Proposal 2: Deferring approval of the change for a plan that accommodates containers, Docker in particular 17:58:20 +1/+1 17:58:21 t8m: That's not great for DNSSEC... 17:58:29 mitr, if that plan could be status quo for containers I would agree 17:58:29 * mitr is +1/+1 for the record 17:58:46 t8m: Let me reword that then... 17:59:07 Proposal 2: Deferring appproval fo the change for a plan that explicitly describes what, if anything, will be done for containers, Docker in particular 17:59:23 mitr, I know, but we should not block DNSSEC for the other products/installations by not having solution for containers 17:59:51 mitr, I can be +1/+1 then 18:00:04 +1/+1 18:00:15 I can be +1 on both with the new language. 18:00:19 +1/+1 again 18:00:19 mitr: +1 to both 18:00:23 +1/+1 18:00:39 #agreed FESCo doesn't have significant objections to the idea as such, or its use on physical machines and VMs (+7) 18:00:51 #agreed Deferring appproval fo the change for a plan that explicitly describes what, if anything, will be done for containers, Docker in particular (+7) 18:01:03 #topic #1302 Fedora Core OS Product 18:01:05 .fesco 1302 18:01:06 mitr: #1302 (Fedora Core OS Product) – FESCo - https://fedorahosted.org/fesco/ticket/1302 18:01:24 Viking-Ice: Were you withdrawing this? 18:01:42 Jóhann has said in the ticket that "This is not something that needs to be discussed" [today]; skip? 18:02:35 sgallagh, obviously since I'm leaving the project + it was just a point in what eventually needs to be done ( from my pov ) with regards to the wg as in the base wg to be split into corewg 18:02:51 and basewg 18:03:17 mitr: +1 to skip. 18:03:31 Proposal: Close this ticket 18:03:42 sgallagh, +1 18:04:05 yeah just close it 18:04:17 sgallagh: +1 18:04:20 #info Will close the ticket based on Jóhann's request 18:04:27 #topic Next week's chair 18:04:29 Anybody? 18:05:08 mitr: I haven't done it in a while. I'll take it 18:05:17 #info sgallagh will chair next week's meeting 18:05:19 sgallagh: thanks 18:05:24 #topic Open Floor 18:05:30 Are there any topics for open floor? 18:05:33 did we forget https://fedorahosted.org/fesco/ticket/1250#comment:22 ? 18:05:48 seems so 18:05:56 oops, my fault. Sorry. 18:06:03 #topic #1250 F21 Self Contained Changes 18:06:06 I would like mattdm clarifiy what I was refering to in that ticket 18:06:07 .fesco 1250 18:06:08 mitr: #1250 (F21 Self Contained Changes) – FESCo - https://fedorahosted.org/fesco/ticket/1250 18:06:25 Viking-Ice: Which ticket? 18:06:31 core os 18:06:59 my disturbing trend 18:07:06 That is: 18:07:08 Remote Journal Logging - ​https://fedoraproject.org/wiki/Changes/Remote_Journal_Logging discussed at ​https://lists.fedoraproject.org/pipermail/devel/2014-April/197980.html 18:07:10 Cache Logical Volumes - ​https://fedoraproject.org/wiki/Changes/Cache_Logical_Volumes discussed at ​https://lists.fedoraproject.org/pipermail/devel/2014-April/198692.html 18:07:12 MariaDB 10.0 - ​https://fedoraproject.org/wiki/Changes/MariaDB10 discussed at ​https://lists.fedoraproject.org/pipermail/devel/2014-April/198796.html 18:07:14 Shogun - ​https://fedoraproject.org/wiki/Changes/shogun discussed at ​https://lists.fedoraproject.org/pipermail/devel/2014-April/198795.html 18:07:16 (skipping Docker Cloud) 18:07:43 * dgilmore is +1 to those self contained changes 18:07:43 Viking-Ice: Please discuss that with him outside the meeting. He also left a little while ago: " annnnd, I need to run to another meeting." 18:08:17 +1 as well 18:08:18 sgallagh, or fesco can just have him clarify that on the ticket 18:08:31 since he's an official fesco representive 18:08:43 +1 to all of them as well 18:08:56 mitr: +1 to those 18:09:31 +1 18:10:26 #agreed Self-Contained Changes: Remote Journal Logging, Cache Logical Volumes, MariaDB 10.0, Shogun approved (+5) 18:10:45 #info Docker Cloud skipped per mattdm's request 18:10:56 Returning to ... 18:10:58 #topic Open Floor 18:11:10 er, late +1 to that. 18:11:16 (but whatever) 18:11:26 #agreed Self-Contained Changes: Remote Journal Logging, Cache Logical Volumes, MariaDB 10.0, Shogun approved (+6) 18:12:19 Anything more for open floor? 18:13:17 * mitr will close the meeting in 2 minutes if not 18:15:23 Thanks, mitr! 18:15:37 #endmeeting