ReleaseEngineering/Meetings/2008-mar-24

From FedoraProject

< ReleaseEngineering | Meetings
Revision as of 17:17, 18 February 2014 by Holmja (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Fedora Release Engineering Meeting :: Monday 2008-03-24

IRC Transcript

f13has changed the topic to: Fedora Release Engineering Meeting - Roll Call<a href="#t13:01" class="time">13:01</a>
f13 ping: notting jeremy spot jwb warren lmacken wwoods poelcat rdieter<a href="#t13:01" class="time">13:01</a>
* jeremy waves<a href="#t13:01" class="time">13:01</a>
jwb i'm here-ish<a href="#t13:02" class="time">13:02</a>
* notting is here<a href="#t13:02" class="time">13:02</a>
f13just a sec, flipping networks<a href="#t13:03" class="time">13:03</a>
f13whoh that's odd<a href="#t13:03" class="time">13:03</a>
* lmacken is hereish<a href="#t13:03" class="time">13:03</a>
f13going from wired + vpn to wireless + vpn happened quick enough that my ssh terminals through the vpn didn't die.<a href="#t13:03" class="time">13:03</a>
jwbthat happens a lot for me<a href="#t13:04" class="time">13:04</a>
f13usually I go from wired no vpn to wireless + vpn<a href="#t13:04" class="time">13:04</a>
* f13 gives folks another couple minutes<a href="#t13:06" class="time">13:06</a>
f13welp<a href="#t13:07" class="time">13:07</a>
f13has changed the topic to: Fedora Release Engineering - Beta Status<a href="#t13:08" class="time">13:08</a>
f13Beta was cut, and made available to mirrors.<a href="#t13:08" class="time">13:08</a>
jwbwith surprising jigdo files<a href="#t13:08" class="time">13:08</a>
f13mirrors are picking it up, I don't have an estimate yet on how many have them, but many do<a href="#t13:08" class="time">13:08</a>
f13yes, we're offering jigdo as an alternative download method for the beta, we'll see how that goes and may continue doing so for releases.<a href="#t13:09" class="time">13:09</a>
f13We also unfroze rawhide Thursday night<a href="#t13:09" class="time">13:09</a>
jwbdo we have metrics in place to track how often it's used?<a href="#t13:09" class="time">13:09</a>
jwbjigdo that is<a href="#t13:09" class="time">13:09</a>
* warren here<a href="#t13:09" class="time">13:09</a>
f13thankfully rawhide has been installable every day since<a href="#t13:09" class="time">13:09</a>
lmackenwith some exceptions (ntfs resizing)<a href="#t13:09" class="time">13:09</a>
f13jwb: all the links to jigdo files will go through mirror manager so theoretically we can capture that info<a href="#t13:09" class="time">13:09</a>
warrenf13, that only works if you happen to get the same vpn address<a href="#t13:09" class="time">13:09</a>
warrenf13, which is entirely by chance<a href="#t13:10" class="time">13:10</a>
f13lmacken: yeah, mostly there have been images available which is good.<a href="#t13:10" class="time">13:10</a>
* spot is here<a href="#t13:10" class="time">13:10</a>
* wwoods hereish<a href="#t13:10" class="time">13:10</a>
f13I'm meeting with mmcgrath later this afternoon to ensure we have a web target to point to from the release announcement, and perhaps static release notes<a href="#t13:10" class="time">13:10</a>
f13as well as ensuring info about jigdo makes it somewhere<a href="#t13:10" class="time">13:10</a>
f13tomorrow I'll flip the bit on the master server at 1345 UTC and mirrors will pick it up shortly after that<a href="#t13:11" class="time">13:11</a>
f13oh I'll be uploading isos to torrent today too, which I seemingly completely forgot about :/<a href="#t13:11" class="time">13:11</a>
sticksterf13: quaid has been setting up the beta relnotes static page this a.m.<a href="#t13:11" class="time">13:11</a>
jwbcool<a href="#t13:12" class="time">13:12</a>
f13goot goot<a href="#t13:12" class="time">13:12</a>
f13post-beta we need to keep an eye on the stability of things, as we'd like tomake a decision near the end of this week regarding slipping the final schedule to take into account earlier slips<a href="#t13:13" class="time">13:13</a>
f13personally I'd rather not slip, but a lot of that is selfish personal reasons<a href="#t13:13" class="time">13:13</a>
f13anything else on Beta before I move on to post-beta topics?<a href="#t13:13" class="time">13:13</a>
jwbdid people think beta went smoother than alpha?<a href="#t13:14" class="time">13:14</a>
f13oh I'm tracking at a high level what I'm doing to "release" beta in a tomboy note, that will make it into an SOC of sorts<a href="#t13:14" class="time">13:14</a>
f13staging the beta was far smoother for me<a href="#t13:14" class="time">13:14</a>
jwbthe quality seems greatly improved from alpha as well.  it's almost as if we're making progress or something<a href="#t13:15" class="time">13:15</a>
f13right<a href="#t13:15" class="time">13:15</a>
f13I like the new qa page for testing stuff, I think we covered a better amount of ground, and jds2001 really helped out this time too<a href="#t13:16" class="time">13:16</a>
f13ok<a href="#t13:17" class="time">13:17</a>
13:17--- f13 has changed the topic to: Fedora Release Engineering - Post-Beta
f13the biggest thing looming post-beta is the ability to pre-branch for F-9<a href="#t13:18" class="time">13:18</a>
f13I just found a tomboy note from when we did the mass-branch, which made use of pkgdb and such.  I'm reviwing it and poking at hte current state of things to figure out what knobs we need to twist to make branching a possiblity.  I at least need to prepare the collections in koji<a href="#t13:18" class="time">13:18</a>
f13I'll send a report about it out later this afternoon<a href="#t13:19" class="time">13:19</a>
wwoodsyeah, much love to jds2001 for helping with installer testing<a href="#t13:19" class="time">13:19</a>
wwoodswe're getting the efforts and plans organized enough that other people can help. amazing!<a href="#t13:19" class="time">13:19</a>
f13The other looming thing is we're supposed to deliver a rawhide snapshot release this Friday<a href="#t13:19" class="time">13:19</a>
BobJensenjds2001 is awesome<a href="#t13:19" class="time">13:19</a>
nottingf13: first time we've tried that, yes?<a href="#t13:20" class="time">13:20</a>
f13yes<a href="#t13:20" class="time">13:20</a>
nottingwhat's it involve? isos? torrent only? a tree? etc.<a href="#t13:20" class="time">13:20</a>
f13the snapshots have been a vague schedule item without much details as to what they will be and how they will be delivered<a href="#t13:20" class="time">13:20</a>
wwoodsIIRC it involves putting isos on a torrent<a href="#t13:20" class="time">13:20</a>
f13so we have some room to define that ourselves.<a href="#t13:20" class="time">13:20</a>
nottingalso, when do we get betas of live isos?<a href="#t13:20" class="time">13:20</a>
wwoodsrawhide's the tree<a href="#t13:20" class="time">13:20</a>
f13now that we have stage2 in rawhide every day in the from of boot.iso, I'm not so sure a full tree for snapshot makes a ton of sense, whereas live images really do still<a href="#t13:21" class="time">13:21</a>
f13notting: 'betas of live isos'?  Those were made and are in the tree with the rest of the beta content<a href="#t13:21" class="time">13:21</a>
wwoodsI fetched some beta liveCD images off jeremy's machine last week; dunno if they've been synched to RDU anywhere<a href="#t13:21" class="time">13:21</a>
sticksterjds2001++<a href="#t13:21" class="time">13:21</a>
nottingf13: for non desktop/kde. or is that up to the spin maintainers?<a href="#t13:22" class="time">13:22</a>
f13notting: up to the spin maker I suppose<a href="#t13:22" class="time">13:22</a>
f13how to handle those spins has also been rather... vague.<a href="#t13:23" class="time">13:23</a>
jeremynotting: I'm somewhat holding off based on the discussion about how we decide what to release also.  and FEL isn't going to have a release this time around<a href="#t13:23" class="time">13:23</a>
wwoodsThere isn't a Beta dir in /mnt/redhat/released/F-9 yet, so.. dunno if it's in RDU yet<a href="#t13:23" class="time">13:23</a>
f13wwoods: I'm rsyncing to that right now.<a href="#t13:23" class="time">13:23</a>
wwoodshot.<a href="#t13:23" class="time">13:23</a>
jwbodd that my mirror has it before you do<a href="#t13:23" class="time">13:23</a>
f13jwb: /mnt/redhat/releases is mostly an afterthought for longterm storage<a href="#t13:23" class="time">13:23</a>
jwbah<a href="#t13:24" class="time">13:24</a>
f13jwb: the rdu "mirror" has it, but it's restricted access<a href="#t13:24" class="time">13:24</a>
f13so I can do "choose your own adventure" isos, as well as Live isos for the snapshot.<a href="#t13:25" class="time">13:25</a>
f13the problem is, if I do isos, I'm going to have to do source isos too (really we should for the live images too, but we aren't...)<a href="#t13:25" class="time">13:25</a>
f13it means more sync time and bits that will sit there without being used much, but *shrug*<a href="#t13:26" class="time">13:26</a>
f13I was thinking we just start on Friday with Friday's rawhide content, get it composed, uploaded to torrent and turn the torrent on as soon as all the bits are in place.  That'll probably be late afternoon on Friday, any objections?<a href="#t13:26" class="time">13:26</a>
f13I would also suggest we keep only one snapshot active on the torrent, remove the previous one when putting up the next one<a href="#t13:27" class="time">13:27</a>
jeremyif we do fridays, then if we miss, it's monday<a href="#t13:27" class="time">13:27</a>
f13jeremy: so Thursdays?<a href="#t13:27" class="time">13:27</a>
jeremyyeah, that's what I was thinking<a href="#t13:27" class="time">13:27</a>
nottingare these going to be 'snapshot, you get all pieces', or are we doing verification?<a href="#t13:28" class="time">13:28</a>
jeremywell, I had thought wednesdays at one point, but I talked myself out of it ;)<a href="#t13:28" class="time">13:28</a>
f13notting: I think it's mostly you get all the pieces<a href="#t13:28" class="time">13:28</a>
f13I'll do some verification as I make the images, but I don't want to get into beating ourselves up on having beta++ quality images each week<a href="#t13:28" class="time">13:28</a>
nottingare we jigdoing the snapshots as well?<a href="#t13:30" class="time">13:30</a>
f13notting: I thought about that, but not sure it'll help.<a href="#t13:30" class="time">13:30</a>
wwoodsthe snapshots should get some very basic smoketests but they won't get the full suite<a href="#t13:30" class="time">13:30</a>
jeremygiven how quickly packages churn in rawhide, ... yeah, not sure jigdo would help<a href="#t13:30" class="time">13:30</a>
wwoodsin theory they should get through the tier1 tests in the install test plan<a href="#t13:30" class="time">13:30</a>
f13notting: we'd have to have the isos available by http then, and you'd lose some because rawhide will move on and we weren't going to have an exploaded tree anywhere.<a href="#t13:30" class="time">13:30</a>
wwoodsbut in theory tier1 is supposed to be completely automated<a href="#t13:31" class="time">13:31</a>
nottingtre<a href="#t13:31" class="time">13:31</a>
nottingf13: true<a href="#t13:31" class="time">13:31</a>
f13although, if mirrors were ok with us abusing them a bit more for snapshots, we could drop them on the phx netapp<a href="#t13:31" class="time">13:31</a>
f13since we have control now<a href="#t13:31" class="time">13:31</a>
warrenmuwhahhaha<a href="#t13:31" class="time">13:31</a>
* jeremy thinks torrent only and reduce the number of moving pieces<a href="#t13:32" class="time">13:32</a>
f13but maybe we should wait and see what kind of uptake we see on torrents and how useful they are<a href="#t13:32" class="time">13:32</a>
f13yeah, less moving parts better<a href="#t13:32" class="time">13:32</a>
f13so lets try to get in touch with eachother on Thursday to cover our snapshot bases.<a href="#t13:33" class="time">13:33</a>
f13jwb: I'll still push the buttons for ppc isos and put them up too, but I'm not really going to do much verification of them.  I'll be looking to you and your team for that<a href="#t13:34" class="time">13:34</a>
f13one last thing on post-beta, again for selfish reasons, if I could get some extra eyes/effort on findign and fixing broken deps and such that would be awesome.  I'd like to keep the tree in a sane state going forward<a href="#t13:36" class="time">13:36</a>
f13I'm going to hook up that cron job to just repoclosure the static-repos hourly so that we get up to the hour info on whats broken<a href="#t13:36" class="time">13:36</a>
f13it won't be hooked into FAS2 or anything like that for assigning blame but it will be informative.<a href="#t13:36" class="time">13:36</a>
f13anything else on Post-Beta before we move on?<a href="#t13:37" class="time">13:37</a>
warrenuh, wouldn't jigdo be PERFECT for snapshots?<a href="#t13:37" class="time">13:37</a>
warrenunless you're talking snapshots of live<a href="#t13:37" class="time">13:37</a>
f13warren: jigdo needs the files available on http mirrors, which teh files will get shuffled out due to rawhide moving on<a href="#t13:38" class="time">13:38</a>
jeremywarren: only if you grab it with that specific day's rawhide<a href="#t13:38" class="time">13:38</a>
jeremywarren: tomorrow, the kernel at least is gone.  and probably ooo. :)<a href="#t13:38" class="time">13:38</a>
warrenhmm<a href="#t13:38" class="time">13:38</a>
f13warren: it also needs the material on http for filling in the missing bits, and they won't be available via http, just torrent.<a href="#t13:38" class="time">13:38</a>
f13a combo of jigdo + torrent would be interesting<a href="#t13:38" class="time">13:38</a>
13:39--- f13 has changed the topic to: Fedora Release Engineering - releng1 changes
f13Now that beta is out, we're (I and the infrastructure team) will be transitioning releng1 over to RHEL5<a href="#t13:40" class="time">13:40</a>
f13this well have effect upon bodhi and rawhide creation<a href="#t13:40" class="time">13:40</a>
f13but at the same time we'll be making some improvements to rawhide creation to hopefully make it a little less hacky and a bit more stable<a href="#t13:40" class="time">13:40</a>
f13and start using some of the cool tools that skvidal has written for us like repodiff<a href="#t13:40" class="time">13:40</a>
f13I don't have an exact date for this, but it'll be soon, but won't conflict with anything we're trying to do for the tree/snapshots/whatever.<a href="#t13:41" class="time">13:41</a>
f13any thoughts/concerns here?<a href="#t13:42" class="time">13:42</a>
jeremynope<a href="#t13:42" class="time">13:42</a>
f13cool.<a href="#t13:42" class="time">13:42</a>
13:42--- f13 has changed the topic to: Fedora Release Engineering - Open Floor
nottingf13: is it 'transitioning releng1 to rhel5', or 'transitioning releng1 to releng2, which happens to be rhel5'?<a href="#t13:42" class="time">13:42</a>
lmacken1->2<a href="#t13:43" class="time">13:43</a>
f13notting: I'm not sure how mmcgrath wants to handle that.  We'll use releng2 to verify things for the move.<a href="#t13:43" class="time">13:43</a>
f13but he did some things for my releng1 disk space, not sure how the act of releng2 becomes the new releng1 will happen<a href="#t13:43" class="time">13:43</a>
f13I doubt it'll be an anaconda upgrade of releng1 though (:<a href="#t13:43" class="time">13:43</a>
lmackenideally we'll have multiple bodhi instances on the app servers, and it will kick off mashes on releng2<a href="#t13:44" class="time">13:44</a>
f13cool<a href="#t13:44" class="time">13:44</a>
f13anything else?<a href="#t13:48" class="time">13:48</a>
f13alright, Ill end the meeting and get working on torrents and cvs.<a href="#t13:50" class="time">13:50</a>

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