ReleaseEngineering/Meetings/2007-jul-02

From FedoraProject

Jump to: navigation, search

Contents

Fedora Release Engineering Meeting :: Monday 2007-07-02

Fedora 5 shutdown

Decision: warren wrapping things up

Fedora 8 Schedule Adjustments

  • Freeze dates do not line up on the correct week days
  • Supposed to be freezes on Tuesdays, and target the next Thursday (one week +2 days) for release
  • Make minor corrections +/-1 day

Decisions:

  • schedule dates will be adjusted to follow format of Freeze on Tuesday, release on following Thursday (week +2days). New schedule will be proposed to FESCo
  • GA changes by one day later to November 8th.
  • Let RHEL Engineering know

Fedora Signing Server

  • It is very clear to us that we need more than just me signing/pushing f7 updates however we also don't want to be giving more people the pass phrase to the gpg key. So we need some form of a signing server.
  • Draft spec for the signing server by f13 is here: http://fedoraproject.org/wiki/JesseKeating/SigningServerSpecDraft
  • f13 is requesting input on the spec (from those that haven't given it to me already)
  • make what is there more official
  • use it to evaluate some existing code (not open source yet) to determine if it's worth the up front cost of open sourcing it vs writing our own.

Decision: None at this time

perl-devel in build roots

  • perl-devel in the minimal build root was supposed to be a short term "hack" for getting Fedora 7 out the door, to be removed for Fedora 8. It is still there.
  • it is scheduled for removal today

Decision: removal of 'perl-devel' from minimal build roots is put on hold until a final vote comes out of Fedora Packaging Committee and ratified by FESCo

image composes in the colo

  • talk with mike/infrastructure and get things set so that we can do Official (tm) tree/live image composes on hardware in the colo (hosted data center)

Decision: topic added to next weeks meeting using this week to determine where we are for accomplishing this goal.

IRC Transcript

--- Log opened Mon Jul 02 11:24:43 2007
Idea.png Irssi: #fedora-meeting: Total of 63 nicks [1 ops, 0 halfops, 0 voices, 62 normal] <a href="#t11:24" class="time">11:24</a>
Idea.png Irssi: Join to #fedora-meeting was synced in 0 secs<a href="#t11:24" class="time">11:24</a>
Idea.png f13 changed the topic of #fedora-meeting to: RELENG-Meeting<a href="#t13:01" class="time">13:01</a>
jwbhere<a href="#t13:01" class="time">13:01</a>
* warren here<a href="#t13:01" class="time">13:01</a>
f13spot: jeremy: notting: wwoods: poelcat: ping<a href="#t13:01" class="time">13:01</a>
f13hrm, no Rex.<a href="#t13:01" class="time">13:01</a>
* jeremy is here<a href="#t13:01" class="time">13:01</a>
nottingyes?<a href="#t13:01" class="time">13:01</a>
f13notting: just role calling<a href="#t13:01" class="time">13:01</a>
* wwoods here<a href="#t13:03" class="time">13:03</a>
f13ok, lets get started.<a href="#t13:03" class="time">13:03</a>
Idea.png f13 changed the topic of #fedora-meeting to: RELENG-Meeting - Fedora 5 shutdown - JesseKeating<a href="#t13:03" class="time">13:03</a>
f13We're well on our way to shutting down FC5<a href="#t13:03" class="time">13:03</a>
f13<a href="http://fedoraproject.org/wiki/ReleaseEngineering/Fedora5Shutdown">http://fedoraproject.org/wiki/ReleaseEngineering/Fedora5Shutdown</a><a href="#t13:04" class="time">13:04</a>
warrenI had trouble getting repoclosure to behave<a href="#t13:04" class="time">13:04</a>
f13I think we're just waiting for warren to finish repoclosure to see how we stand, and send an announcement.<a href="#t13:04" class="time">13:04</a>
f13and "fix" the internal tools (this may have already been done)<a href="#t13:04" class="time">13:04</a>
f13and then we'll be done.<a href="#t13:05" class="time">13:05</a>
f13Any further discussion needed on this?<a href="#t13:05" class="time">13:05</a>
warrennot really<a href="#t13:05" class="time">13:05</a>
warrenI'll finish up whatever needs to be done<a href="#t13:05" class="time">13:05</a>
warrenI have the announcement almost writen<a href="#t13:05" class="time">13:05</a>
warrenwritten<a href="#t13:05" class="time">13:05</a>
wwoodssounds good<a href="#t13:06" class="time">13:06</a>
f13k, moving on.<a href="#t13:06" class="time">13:06</a>
Idea.png f13 changed the topic of #fedora-meeting to: RELENG-Meeting - Fedora 8 Schedule Adjustments - JesseKeating<a href="#t13:06" class="time">13:06</a>
f13so this is a bad on me thing.<a href="#t13:06" class="time">13:06</a>
f13I had previously been looking at the schedule and thinking that all the freeze days and stuff were right and not noticing that some freezes were on Wed, some on Monday, some on Friday<a href="#t13:07" class="time">13:07</a>
f13and the release days were on Wed.  This is not what we agreed upon late last release where we would do freezes on Tuesdays, and target the next Thursday (one week +2 days).<a href="#t13:07" class="time">13:07</a>
f13so I'd like to apply that to the schedule propose it to FESCo<a href="#t13:08" class="time">13:08</a>
f13that would shift TEst1 to Freeze July 24th, release Aug 2nd<a href="#t13:09" class="time">13:09</a>
f13Test2 is Freeze Aug 28th, release Sept 6th<a href="#t13:10" class="time">13:10</a>
f13test3 is Sept 25th freeze, Oct 4th release<a href="#t13:10" class="time">13:10</a>
warreneffects final too?<a href="#t13:11" class="time">13:11</a>
f13Final freeze oct 24th, release Nov 8th<a href="#t13:11" class="time">13:11</a>
f13er<a href="#t13:11" class="time">13:11</a>
f13Freeze Oct 23rd, release Nov 8th<a href="#t13:11" class="time">13:11</a>
f13these aren't really big changes, most are just a single day on one or the other side, most of them make the freeze a day earlier and the release a day later.<a href="#t13:12" class="time">13:12</a>
warrensounds good to me<a href="#t13:12" class="time">13:12</a>
warrenf13, one caveat<a href="#t13:12" class="time">13:12</a>
nottingwfm<a href="#t13:12" class="time">13:12</a>
f13(note that freeze doesn't always extend through till the release, freeze can be lifted earlier)<a href="#t13:12" class="time">13:12</a>
jeremywfm<a href="#t13:12" class="time">13:12</a>
warrenf13, notify RHEL engineering so perhaps their deadlines don't clash with ours<a href="#t13:13" class="time">13:13</a>
wwoodsyeah, these are still pretty minor changes but I'm sure they'll appreciate being notified<a href="#t13:13" class="time">13:13</a>
f13warren: can do.<a href="#t13:13" class="time">13:13</a>
f13Any opposed?<a href="#t13:13" class="time">13:13</a>
warrenf13, kreilly might be good.<a href="#t13:14" class="time">13:14</a>
warrenf13, +1<a href="#t13:14" class="time">13:14</a>
f13ok, We'll consider that passed and I'll present to FESCo<a href="#t13:14" class="time">13:14</a>
f13or maybe just mail to list to get it taken care of sooner<a href="#t13:14" class="time">13:14</a>
poelcatf13: summary?<a href="#t13:14" class="time">13:14</a>
f13poelcat: Decision, schedule dates will be adjusted to follow format of Freeze on Tuesday, release on following Thursday (week +2days).  New schedule will be proposed to FESCo<a href="#t13:15" class="time">13:15</a>
Idea.png f13 changed the topic of #fedora-meeting to: RELENG-Meeting - Fedora Signing Server - JesseKeating<a href="#t13:15" class="time">13:15</a>
poelcatf13: thanks!<a href="#t13:15" class="time">13:15</a>
f13It is very clear to us that we need more than just me signing/pushing f7 updates<a href="#t13:15" class="time">13:15</a>
f13however we also don't want to be giving more people the passphrase to the gpg key.  So we need some form of a signing server.  I created a draft spec for the signing server:<a href="#t13:16" class="time">13:16</a>
warrentimezone coverage (more than just American east) would be nice<a href="#t13:16" class="time">13:16</a>
f13<a href="http://fedoraproject.org/wiki/JesseKeating/SigningServerSpecDraft">http://fedoraproject.org/wiki/JesseKeating/SigningServerSpecDraft</a><a href="#t13:16" class="time">13:16</a>
f13warren: yes, that can be accomplished by having more than just me.<a href="#t13:16" class="time">13:16</a>
f13I'd like to get some initial input on the spec (from those that haven't given it to me already) so that I can make this more official, and use it to evaluate some existing code (not opensource yet) to determine if it's worth the up front cost of opensourcing it vs writing our own.<a href="#t13:17" class="time">13:17</a>
f13either way, having a general consensus of how this thing should operate and what goals we'd like to accomplish with it would be good.<a href="#t13:18" class="time">13:18</a>
f13so take time after the meeting to review and provide me with whatever feedback you can, and we'll vote on list or some such to make it "official" after the fact.  Sound good?<a href="#t13:18" class="time">13:18</a>
warrenare you asking implementation details yet?<a href="#t13:18" class="time">13:18</a>
f13warren: purposefully left out.<a href="#t13:19" class="time">13:19</a>
f13warren: I don't want people's minds clouded with such details.  I'm more after the general goals and high level design<a href="#t13:19" class="time">13:19</a>
jwblike i said before, it seems good to me<a href="#t13:20" class="time">13:20</a>
f13ok, lets move on in the interest of time.<a href="#t13:20" class="time">13:20</a>
Idea.png f13 changed the topic of #fedora-meeting to: RELENG-Meeting - perl-devel in buildroots - JesseKeating<a href="#t13:20" class="time">13:20</a>
f13This is the last named subject for today<a href="#t13:20" class="time">13:20</a>
f13perl-devel in the minimal buildroot was supposed to be a short term "hack" for getting Fedora 7 out the door, to be removed for Fedora 8.  I forgot to remove it early on, and noticed it recently<a href="#t13:21" class="time">13:21</a>
f13Robin sent an announcement that this would be going away, scheduled for removal today.<a href="#t13:21" class="time">13:21</a>
warrenI personally think we don't gain anything by removing it from the minimum buildroot.<a href="#t13:21" class="time">13:21</a>
f13There has been some discussion after that about this afterward, but I don't thikn it means we shoudln't remove it.<a href="#t13:22" class="time">13:22</a>
f13warren: silent BuildRequires in packages.<a href="#t13:22" class="time">13:22</a>
f13or rather missing.<a href="#t13:22" class="time">13:22</a>
jeremyf13: is perl-devel really any worse in that respect than gcc?<a href="#t13:22" class="time">13:22</a>
jeremy(aside: not sure this is really rel-eng's call in any case)<a href="#t13:22" class="time">13:22</a>
warrenf13, there isn't full agreement that perl should have been split at all.<a href="#t13:23" class="time">13:23</a>
f13jeremy: honestly I don't know, but the split of devel content vs runtime content has been made, we're just cleaning up the temporary work around.<a href="#t13:23" class="time">13:23</a>
warrenf13, perl-devel could be in the same category as gcc.  Where exactly do we draw this arbitrary line?<a href="#t13:23" class="time">13:23</a>
f13warren: there isn't full agreement on many if not most of the things that are done in Fedora.<a href="#t13:23" class="time">13:23</a>
jeremyf13: I think we should hold off on the change and push the question to the packaging committee as they're the people that should be making the call<a href="#t13:24" class="time">13:24</a>
jwbyes<a href="#t13:25" class="time">13:25</a>
f13to be fair, we'd be removing a change we did before.  We've always had 'perl' listed as a default package.  This is really a subject of the 'perl' package changed contents do we adjust buildroot .<a href="#t13:25" class="time">13:25</a>
f13IIRC this came up with python-devel a while ago and we didn't add python-devel to the buildroot<a href="#t13:25" class="time">13:25</a>
warrenSo we're going with jeremy's suggestion?<a href="#t13:26" class="time">13:26</a>
f13but I'm OK with passing it off to FPC as well.<a href="#t13:26" class="time">13:26</a>
f13they decide, we impliment.<a href="#t13:27" class="time">13:27</a>
warrenIsn't FPC where this originated?<a href="#t13:27" class="time">13:27</a>
warrenperhaps FESCo should vote on this?<a href="#t13:27" class="time">13:27</a>
f13FESCo gets teh chance to override any FPC (and releng) decision<a href="#t13:27" class="time">13:27</a>
warrenFPC normally only makes recommendations for FESCo to ratify.<a href="#t13:27" class="time">13:27</a>
f13they're presented at the beginning of each FESCo meeting, so nothing new needs to happen wrt FESCo<a href="#t13:28" class="time">13:28</a>
f13poelcat: DECISION: removal of 'perl-devel' from minimal buildroots is put on hold until a final vote comes out of Fedora Packaging Committee and ratified by FESCo<a href="#t13:28" class="time">13:28</a>
Idea.png f13 changed the topic of #fedora-meeting to: RELENG-Meeting - Open Discussion<a href="#t13:28" class="time">13:28</a>
f13anybody got anything they'd like to discuss?<a href="#t13:29" class="time">13:29</a>
drago01f13: mass rebuild?<a href="#t13:29" class="time">13:29</a>
f13drago01: FESCo decision.<a href="#t13:29" class="time">13:29</a>
warrenDid/will any major toolchain things happen necessitating a mass rebuild?<a href="#t13:29" class="time">13:29</a>
drago01f13: ok<a href="#t13:29" class="time">13:29</a>
f13warren: I've only caught wind of somethings that might<a href="#t13:30" class="time">13:30</a>
nottingwarren: possibly<a href="#t13:30" class="time">13:30</a>
drago01dunno but will we have gcc 4.2 in F8 ?<a href="#t13:30" class="time">13:30</a>
wwoodswas there a FESCo decision about how to propose/track new features?<a href="#t13:30" class="time">13:30</a>
nottingdrago01: no<a href="#t13:30" class="time">13:30</a>
f13drago01: question for the Tools team.<a href="#t13:30" class="time">13:30</a>
wwoodssince we'll be the ones responsible for shepherding / kicking things out<a href="#t13:30" class="time">13:30</a>
warrengcc 4.2 will break C++ ABI for the first time since gcc-3.3 IIRC.<a href="#t13:30" class="time">13:30</a>
f13wwoods: poelcat's proposal is out there, FESCo talked a bit about it.<a href="#t13:30" class="time">13:30</a>
* poelcat is working on 0.2 right now<a href="#t13:31" class="time">13:31</a>
jwbf13, hans wants us to talk about automatic update pushing<a href="#t13:31" class="time">13:31</a>
* jwb ducks<a href="#t13:31" class="time">13:31</a>
f13again?<a href="#t13:31" class="time">13:31</a>
wwoodsstill being refined and discussed, then. in the meantime should we add stuff to the feature list as before?<a href="#t13:31" class="time">13:31</a>
jwbf13, i think he missed that we talked about it<a href="#t13:31" class="time">13:31</a>
wwoodsI thought automatic update pushing a) required a signing server and b) was maybe already added to bodhi<a href="#t13:32" class="time">13:32</a>
f13wwoods: promotion, not pushing.<a href="#t13:33" class="time">13:33</a>
jwbwwoods, right promotion not pushing<a href="#t13:33" class="time">13:33</a>
f13wwoods: yes, just add to list as before.<a href="#t13:33" class="time">13:33</a>
jwbsorry, my mistake<a href="#t13:33" class="time">13:33</a>
wwoodsah<a href="#t13:33" class="time">13:33</a>
f13jwb: I replied to hans<a href="#t13:33" class="time">13:33</a>
jwbheh, me too<a href="#t13:33" class="time">13:33</a>
XulChriswhats the diff between promoting and pushing?<a href="#t13:33" class="time">13:33</a>
jwbXulChris, promotion means the request goes from updates-testing to updates<a href="#t13:34" class="time">13:34</a>
jwbXulChris, pushing means the packages actually get pushed to repos<a href="#t13:34" class="time">13:34</a>
mclasenof course, the bodhi ui does not use these terms...<a href="#t13:34" class="time">13:34</a>
f13XulChris: promotion is something the maintainer does, or automation does in behalf of the maintainer.  pushing is somethign rel-eng does, which involves signing the packages and such.<a href="#t13:34" class="time">13:34</a>
f13mclasen: yes, some terminology cleansing is needed.<a href="#t13:35" class="time">13:35</a>
XulChrisf13: ah k thx<a href="#t13:35" class="time">13:35</a>
jwbmclasen, we can change that<a href="#t13:35" class="time">13:35</a>
f13are there any other topics to discuss?<a href="#t13:35" class="time">13:35</a>
jeremywho's working on the 36 hour days? :)<a href="#t13:36" class="time">13:36</a>
jwbi am<a href="#t13:36" class="time">13:36</a>
jeremyexcellent.  keep us updated on the status<a href="#t13:36" class="time">13:36</a>
poelcatf13: what about freeze dates around rel-eng tools<a href="#t13:36" class="time">13:36</a>
jwbjeremy, will do.  currently blocked by that stupid speed of light issue<a href="#t13:37" class="time">13:37</a>
poelcatseems like someone mentioned that as something to do differently for F8?<a href="#t13:37" class="time">13:37</a>
jwbpoelcat, meaning pungi and livecd-tools, or other things only used by rel-eng like mash?<a href="#t13:37" class="time">13:37</a>
f13poelcat: well, they're a good idea, but I don't know if we're to the point of making policy on that.<a href="#t13:37" class="time">13:37</a>
mclasenjwb: update to light 2.0<a href="#t13:37" class="time">13:37</a>
f13especially as that would cut across into things like Anaconda which is a bit beyond our reach<a href="#t13:38" class="time">13:38</a>
jwbmclasen, ooh!  that rss feed must have gone by too quickly<a href="#t13:38" class="time">13:38</a>
poelcatf13: wouldn't that overall be a good thing if all that was in sync?<a href="#t13:38" class="time">13:38</a>
jeremypoelcat: I think for f7, we largely went by "freeze the tools feature-wise by feature freeze, bugfixes from there"<a href="#t13:39" class="time">13:39</a>
jeremypoelcat: which seems reasonable enough to me<a href="#t13:39" class="time">13:39</a>
jwbwell<a href="#t13:40" class="time">13:40</a>
f13poelcat: I think rel-eng should be aware of changes being made to things that effect us, like changes in Anaconda, koji, etc...<a href="#t13:40" class="time">13:40</a>
jwbi think one thing we should strive for this time is to make sure the pungi/livecd-tools that are used to spin the isos for release are the versions actually included in the release<a href="#t13:40" class="time">13:40</a>
f13jwb: sometimes that's hard to do<a href="#t13:40" class="time">13:40</a>
jwbf13, i said "strive for"  not "require"<a href="#t13:40" class="time">13:40</a>
f13true<a href="#t13:41" class="time">13:41</a>
jeremyjwb: if we're going to do that, then we need to move the configs out of the packages imho<a href="#t13:41" class="time">13:41</a>
jwbjeremy, yes.  is that a problem per se?<a href="#t13:42" class="time">13:42</a>
jeremyjwb: it's not a huge problem -- it's a change from what some people have wanted, though<a href="#t13:42" class="time">13:42</a>
jwbi can see both sides i guess<a href="#t13:42" class="time">13:42</a>
jeremyme too<a href="#t13:42" class="time">13:42</a>
jwbi don't think we need to decide anything right now.  maybe just keep it in the middle of our brains<a href="#t13:43" class="time">13:43</a>
f13sure<a href="#t13:43" class="time">13:43</a>
jeremyjwb: sounds good :)<a href="#t13:43" class="time">13:43</a>
jeremyf13: related note -- we should talk with mike/infrastructure and get things set so that we can do Official (tm) tree/live image composes on hardware in teh colo<a href="#t13:44" class="time">13:44</a>
nottingis there anything preventing that?<a href="#t13:44" class="time">13:44</a>
* warren has A/C repairman over, bbl<a href="#t13:45" class="time">13:45</a>
jeremynotting: preventing which?<a href="#t13:46" class="time">13:46</a>
nottingcomposes in phx<a href="#t13:46" class="time">13:46</a>
f13disk space and machines mostly<a href="#t13:46" class="time">13:46</a>
f13there is an RFR for systems I think.<a href="#t13:46" class="time">13:46</a>
mmcgrath~/me is here<a href="#t13:49" class="time">13:49</a>
f13jeremy: lets add that topic to next weeks meeting, and spend this week seeing where we are for accomplishing that goal.<a href="#t13:49" class="time">13:49</a>
jeremyf13: okie<a href="#t13:49" class="time">13:49</a>
f13anything else?<a href="#t13:50" class="time">13:50</a>
f13ok, meeting over.<a href="#t13:53" class="time">13:53</a>
Idea.png f13 changed the topic of #fedora-meeting to: <a href="http://fedoraproject.org/wiki/Communicate/FedoraMeetingChannel">http://fedoraproject.org/wiki/Communicate/FedoraMeetingChannel</a> -- Meetings often get logged -- see schedule in the wiki for next meeting<a href="#t13:53" class="time">13:53</a>
f13thanks all<a href="#t13:53" class="time">13:53</a>
--- Log closed Mon Jul 02 13:55:23 2007

Generated by irclog2html.py 2.3 by Marius Gedminas - find it at mg.pov.lt!