From Fedora Project Wiki

Fedora Release Engineering Meeting :: Monday 2007-10-22

F8 Readiness Discussion

IRC Transcript

Idea.png f13 changed the topic of #fedora-meeting to: Release Engineering Meeting<a href="#t13:02" class="time">13:02</a>
f13jwb: notting: jeremy: wwoods: warren: rdieter: poelcat: spot: ping<a href="#t13:03" class="time">13:03</a>
* notting is here<a href="#t13:04" class="time">13:04</a>
warrenyo<a href="#t13:04" class="time">13:04</a>
* spot is here<a href="#t13:04" class="time">13:04</a>
* jeremy is around<a href="#t13:04" class="time">13:04</a>
Idea.png f13 changed the topic of #fedora-meeting to: Release Engineering Meeting - Release Readiness<a href="#t13:05" class="time">13:05</a>
f13so I invited a bunch of folks to crash the meeting today, to discuss release readiness.  Something I want to do on a nearly daily bases as we head toward Fedora 8<a href="#t13:05" class="time">13:05</a>
f13Do we have any guests out there?<a href="#t13:06" class="time">13:06</a>
* poelcat here<a href="#t13:07" class="time">13:07</a>
* wwoods almost here (need beverage)<a href="#t13:08" class="time">13:08</a>
f13hrm, looks like nobody is joining us.<a href="#t13:08" class="time">13:08</a>
f13jrb: cebbert davej ping<a href="#t13:08" class="time">13:08</a>
f13well...<a href="#t13:10" class="time">13:10</a>
f13We're planning for a Nov 8 release<a href="#t13:10" class="time">13:10</a>
f13In my mind that means we have to have a Gold tree by Nov 1, maybe the 2nd if it's early in the day<a href="#t13:10" class="time">13:10</a>
* jwb is late<a href="#t13:10" class="time">13:10</a>
f13that means we have this week to really pound out bugs and validate install paths<a href="#t13:10" class="time">13:10</a>
f13We have a slight issue with disk space in the colo and getting signed packages written out.  We need to do some garbage collection first<a href="#t13:11" class="time">13:11</a>
wwoodspretty much, yeah<a href="#t13:11" class="time">13:11</a>
warrenpound<a href="#t13:11" class="time">13:11</a>
f13so I don't plan on starting to write out signed packages until this weekend<a href="#t13:11" class="time">13:11</a>
warrenf13, be sure to remind me when signing happens<a href="#t13:11" class="time">13:11</a>
f13warren: I will<a href="#t13:11" class="time">13:11</a>
f13I'm pretty confident in the compose tools so it really should be all about the bits.<a href="#t13:12" class="time">13:12</a>
nottinghow slight is slight?<a href="#t13:12" class="time">13:12</a>
f13we've got 250~ G free<a href="#t13:13" class="time">13:13</a>
ljuwaidahis this the kde meeting?<a href="#t13:13" class="time">13:13</a>
jwbljuwaidah, no<a href="#t13:13" class="time">13:13</a>
f13ljuwaidah: no.<a href="#t13:13" class="time">13:13</a>
f13notting: <a href="https://admin.fedoraproject.org/cacti/graph.php?action=view&rra_id=all&local_graph_id=280">https://admin.fedoraproject.org/cacti/graph.php?action=view&rra_id=all&local_graph_id=280</a><a href="#t13:13" class="time">13:13</a>
ljuwaidahwhere is it then?<a href="#t13:13" class="time">13:13</a>
f13ljuwaidah: you missed it.<a href="#t13:14" class="time">13:14</a>
f13off by an hour perhaps<a href="#t13:14" class="time">13:14</a>
ljuwaidahisn't it at 5PM utc?<a href="#t13:14" class="time">13:14</a>
f13right now is 5pm UTC correct.<a href="#t13:15" class="time">13:15</a>
jwbbut it's not the KDE meeting<a href="#t13:15" class="time">13:15</a>
nottingljuwaidah: ... on tuesdays<a href="#t13:15" class="time">13:15</a>
lmackenljuwaidah: tuesdays<a href="#t13:15" class="time">13:15</a>
f13ljuwaidah: you're off by a day<a href="#t13:15" class="time">13:15</a>
ljuwaidahOh...<a href="#t13:15" class="time">13:15</a>
ljuwaidahsorry :D<a href="#t13:15" class="time">13:15</a>
ljuwaidahhahahahaha<a href="#t13:15" class="time">13:15</a>
f13so anyway...<a href="#t13:16" class="time">13:16</a>
f13that's where we're at from the compose side of things.<a href="#t13:16" class="time">13:16</a>
ljuwaidahcan i stay?<a href="#t13:16" class="time">13:16</a>
nottingf13: any insight into the errors from anaconda-devel-list?<a href="#t13:16" class="time">13:16</a>
f13I'm going to drop fedora-release in today I think, provided that mdomsch hooked up the mirrorlist cgi to return f8 requests to rawhide for now.<a href="#t13:16" class="time">13:16</a>
f13notting: no, and they didn't repeat last night<a href="#t13:16" class="time">13:16</a>
f13I can start signing and importing if I have sufficient disk space ona  box somewhere, or else I"d have to abuse /mnt/koji for the signing even, not just writing out.<a href="#t13:17" class="time">13:17</a>
f13ljuwaidah: you're welcome to stay.<a href="#t13:17" class="time">13:17</a>
ljuwaidahf13: thanks :)<a href="#t13:18" class="time">13:18</a>
f13wwoods: what do we have setup QA wise?<a href="#t13:18" class="time">13:18</a>
poelcathas the libIlmImf.so.4, this is not available problem related to kde been fixed?<a href="#t13:19" class="time">13:19</a>
poelcati still see it updating x86_64<a href="#t13:19" class="time">13:19</a>
f13rdieter: ?<a href="#t13:20" class="time">13:20</a>
wwoodsf13: the usual junk, mostly just KATE install tests. I've got some upgrade tests lined up for F7->F8<a href="#t13:21" class="time">13:21</a>
wwoodsthere's no automated testing for Live images yet, so those will need attention<a href="#t13:21" class="time">13:21</a>
f13wwoods: do we have a tree testing page?  Going to send out the "OMG TEST IT NOW" emails?<a href="#t13:21" class="time">13:21</a>
wwoodsI'll be doing that stuff today<a href="#t13:21" class="time">13:21</a>
f13ok.<a href="#t13:21" class="time">13:21</a>
wwoodsBlocker list is still at 51 bugs, but quite a few of those are questionable<a href="#t13:22" class="time">13:22</a>
wwoodsthings that will need attention: PulseAudio (esp. in KDE), NM, new compiz..<a href="#t13:23" class="time">13:23</a>
cebbertf13: belated pong<a href="#t13:23" class="time">13:23</a>
wwoodsoh. what's going on with GRUB menu entries? I keep seeing bugs about it failing to drop old entries on upgrade, or failing to add entries for Windows<a href="#t13:24" class="time">13:24</a>
jeremywwoods: the latter was a bug in booty that I fixed just after test3<a href="#t13:24" class="time">13:24</a>
davejf13: ugh, little more warning than 30 mins next time please?<a href="#t13:24" class="time">13:24</a>
jeremywwoods: the former still needs looking at (and is on the blocker list, as we should really fix it because it looks really bad)<a href="#t13:24" class="time">13:24</a>
f13davej: sorry, it just occured to me that we could double book this meeting time.<a href="#t13:24" class="time">13:24</a>
wwoodsthere's also no update on the traditional Post Freeze Boot Hang (bug 338551)<a href="#t13:24" class="time">13:24</a>
davej(especially as my fetchmail only runs every 30 minutes..)<a href="#t13:25" class="time">13:25</a>
f13davej: of course if I waited for people to give me feedback, we'd probably have our first meeting next year.<a href="#t13:25" class="time">13:25</a>
f13wwoods: I think a kernel went in last night that fixes that.<a href="#t13:25" class="time">13:25</a>
wwoodsjeremy: is that one gonna be a tricky fix? I hate messing with grub etc. this close to release<a href="#t13:25" class="time">13:25</a>
f13- Disable polling before registering netdevice in e100.<a href="#t13:25" class="time">13:25</a>
jwbi'd like to know what the hell is going on with my backlight buttons, but i don't consider it anywhere close to a release blocker<a href="#t13:26" class="time">13:26</a>
jeremywwoods: shouldn't be tricky other than setting upgrades takes too damned long :)<a href="#t13:26" class="time">13:26</a>
wwoodsoh, I noticed that we have a repo-edit button when pirut fails to start now.. that should make bug 315671 (pirut/pup demand DVD) non-blocker<a href="#t13:27" class="time">13:27</a>
nottingjeremy: fwiw, on a FC6 upgrade i did, grubby created two new entries and did not remove the old one<a href="#t13:27" class="time">13:27</a>
jeremywwoods: or even closed as the fix is "let you go in and disable the repo" :)<a href="#t13:28" class="time">13:28</a>
jeremynotting: yeah, I reproduced it when I did an upgrade a week or so ago.  but I then didn't set up a second upgrade to test with<a href="#t13:28" class="time">13:28</a>
wwoodsjeremy: oh, anaconda allowing x86_64 upgrades on i386 systems (222424) - any plan for that one?<a href="#t13:28" class="time">13:28</a>
jeremywwoods: we _should_ be giving a warning now.  if we're not, then that's the fix<a href="#t13:28" class="time">13:28</a>
poelcatf13: what package should I file a bug against for the libIlmImf.so problem?<a href="#t13:29" class="time">13:29</a>
wwoodsjeremy: right, but you mention wanting a long-term solution, and I didn't know if you wanted to keep the bug open for that. If not, I'll retest and close<a href="#t13:29" class="time">13:29</a>
f13poelcat: what is the full error output?<a href="#t13:29" class="time">13:29</a>
wwoodsf13: do we have config files etc. for making RC images?<a href="#t13:29" class="time">13:29</a>
jeremywwoods: nah<a href="#t13:29" class="time">13:29</a>
wwoodsor, better yet, RC images?<a href="#t13:29" class="time">13:29</a>
jeremywwoods: we can't do an RC until we have signed packages<a href="#t13:30" class="time">13:30</a>
f13wwoods: I don't want to call them "RC" until packages are signed, but I don't envision changing the config files from as shipped in current pungi package (and fedora-release once I tag it)<a href="#t13:30" class="time">13:30</a>
wwoodswell, not literal RC images/media, but just some DVD images I can test with (since we have media-based tests and bugs)<a href="#t13:31" class="time">13:31</a>
wwoodskcool, I'll run with what's in pungi<a href="#t13:31" class="time">13:31</a>
f13wwoods: the config in the pungi package is suitable.<a href="#t13:31" class="time">13:31</a>
f13I was going to make a pile of them today but anaconda boomage<a href="#t13:31" class="time">13:31</a>
jeremylivecd is still hovering a little too close for comfort to the size (and ended up going over with the build I did this morning), so probably going to end up taking something out and shooting it<a href="#t13:31" class="time">13:31</a>
poelcatf13: ERROR with rpm_check_debug vs depsolve:<a href="#t13:31" class="time">13:31</a>
poelcatPackage kdegraphics needs libIlmImf.so.4, this is not available.<a href="#t13:31" class="time">13:31</a>
poelcatPackage kdebase needs libIlmImf.so.4, this is not available.<a href="#t13:31" class="time">13:31</a>
poelcatupdating OpenEXR-libs so I'll file it there I guess<a href="#t13:32" class="time">13:32</a>
f13poelcat: I don't see those broken deps in the rawhide report<a href="#t13:32" class="time">13:32</a>
poelcathmm  I see it on my box :)<a href="#t13:32" class="time">13:32</a>
f13poelcat: I bet new installs don't see this.  Would be interesting to see if F7 -> F8 sees it though<a href="#t13:32" class="time">13:32</a>
jeremyf13: that's easy enough to test.  one sec<a href="#t13:33" class="time">13:33</a>
f13davej: cebbert: how are we looking kernel side?<a href="#t13:36" class="time">13:36</a>
jwbi've been running rawhide kernels for a while now.  seem pretty stable<a href="#t13:37" class="time">13:37</a>
jwbacross i686, x86_64, ppc64<a href="#t13:37" class="time">13:37</a>
f13there is an e100 crasher I've heard about<a href="#t13:37" class="time">13:37</a>
jeremyf13: it looks like a lot of the kde packages now have split out -libs packages.  which means they're going to need upgrade munging<a href="#t13:37" class="time">13:37</a>
nottingjeremy: ugh. that's going to break yum upgrades too<a href="#t13:38" class="time">13:38</a>
jeremyrdieter: can you file a bug against anaconda with how things were split so that we can actually have upgrades work<a href="#t13:38" class="time">13:38</a>
jeremynotting: yes, yes it is<a href="#t13:38" class="time">13:38</a>
f13le sigh<a href="#t13:38" class="time">13:38</a>
* jeremy notes that the fact that the x86_64 live image is multilib makes testing this incredibly easy...<a href="#t13:38" class="time">13:38</a>
davejf13: I still haven't got confirmation that the e100 problem is fixed properly (though upstream ACKd the patch I did)<a href="#t13:38" class="time">13:38</a>
davejf13: otherwise - not too shabby. I'm going to kick off a bunch of installs today and see if I can shake anything out.  On the whole though, we're in better shape than we were this time at F7<a href="#t13:39" class="time">13:39</a>
cebbertf13: e100 bug is 321521<a href="#t13:39" class="time">13:39</a>
wwoodshow about the other stuff on the kernel blocker list? actually worth blocking over?<a href="#t13:39" class="time">13:39</a>
jwbi know that ppc kernels stopped working on ppc32 CHRP recently<a href="#t13:40" class="time">13:40</a>
jwbbut nobody cares<a href="#t13:40" class="time">13:40</a>
jwbthey're broken in F7 update kernels anyway<a href="#t13:40" class="time">13:40</a>
davejI seem to have gained a ppc imac, so I could give that a test later too. (or I could do something useful with my time)<a href="#t13:40" class="time">13:40</a>
ljuwaidahbrb<a href="#t13:40" class="time">13:40</a>
f13hah<a href="#t13:41" class="time">13:41</a>
jwbdavej, that's not 32-bit CHRP<a href="#t13:41" class="time">13:41</a>
cebbertwwoods: I think most of those can be solved by disabling DMA for libata devices, at least that will get the install working<a href="#t13:41" class="time">13:41</a>
davejno, I know.<a href="#t13:41" class="time">13:41</a>
* nirik notes that both his ppc32 and x86_64 rawhide test boxes have been just fine with current rawhide kernels. <a href="#t13:41" class="time">13:41</a>
jwbdavej, you're welcome to send that imac to me.  i'd make use of it<a href="#t13:41" class="time">13:41</a>
jwbnirik, 32bit CHRP is like a 43p-150 RS/6000<a href="#t13:41" class="time">13:41</a>
f13Im' pretty sure we've made no claims to supporting that hardware<a href="#t13:41" class="time">13:41</a>
cebbertwwoods: do we have a link to <a href="https://fedoraproject.org/wiki/KernelCommonProblems">https://fedoraproject.org/wiki/KernelCommonProblems</a> in the release notes?<a href="#t13:41" class="time">13:41</a>
nirikjwb: yeah, I had a bebox that I think was a CHRP box... it's been a long time tho, so I might be misremembering.<a href="#t13:41" class="time">13:41</a>
jwbf13, not true.  it was supported in FC-6 and F7 on release<a href="#t13:42" class="time">13:42</a>
wwoodscebbert: no, but we definitely should<a href="#t13:42" class="time">13:42</a>
davejjwb: noted for when I need that space :)<a href="#t13:42" class="time">13:42</a>
f13jwb: but was that intentional or by accident?<a href="#t13:42" class="time">13:42</a>
jwbf13, i know because i sent one to nasrat for it way back when<a href="#t13:42" class="time">13:42</a>
jwbf13, intentional<a href="#t13:42" class="time">13:42</a>
jwbf13, anyway, there's a bug open and dwmw2 is looking at it<a href="#t13:42" class="time">13:42</a>
f13interesting<a href="#t13:42" class="time">13:42</a>
f13do we even setup the booting correct for chrp in the boot images?<a href="#t13:43" class="time">13:43</a>
davejwwoods/cebbert: hmm, is it really release note material?   it may give an overly negative impression.. "here's a list of a dozen things that might go wrong"<a href="#t13:43" class="time">13:43</a>
davejas opposed to the actual "common problems we know about in F8" link we already have<a href="#t13:43" class="time">13:43</a>
wwoodserr, hmm<a href="#t13:43" class="time">13:43</a>
jwbf13, you have to enter linux32 or something explicitly in yaboot but it worked<a href="#t13:43" class="time">13:43</a>
nirikany chance the last of the ps3 blockers could get fixed so f8 works out of the box for installs? (I know there probibly aren't many ps3 installs, but it would be nice. ;)<a href="#t13:43" class="time">13:43</a>
wwoodswell, we should.. yes. the Common Bugs link should mention the KernelCommonProblems page<a href="#t13:43" class="time">13:43</a>
jwbf13, but in reality, there are very few people that even have these machines still.<a href="#t13:43" class="time">13:43</a>
f13nirik: that reminds me to test that thing today<a href="#t13:44" class="time">13:44</a>
cebbertwwoods: but maybe we should note that these aren't necessarily known to be present in F8<a href="#t13:44" class="time">13:44</a>
nirikf13: if you could also update the wiki page on exactly how to do the install that would be cool too. I have been meaning to try here, but haven't yet.<a href="#t13:45" class="time">13:45</a>
* rdieter looks in...<a href="#t13:45" class="time">13:45</a>
wwoodscebbert: right. hmm. probably it just needs to be mentioned as "if you're having problems that are't mentioned in this list, here are a few places you can look for de3bugging help"<a href="#t13:46" class="time">13:46</a>
rdieterjeremy: fyi, <a href="https://bugzilla.redhat.com/show_bug.cgi?id=339981">https://bugzilla.redhat.com/show_bug.cgi?id=339981</a><a href="#t13:46" class="time">13:46</a>
f13nirik: bah.  Still can't use the network device.<a href="#t13:47" class="time">13:47</a>
nirikf13: is there a bug on that? just anaconda not using the right device name?<a href="#t13:47" class="time">13:47</a>
f13nirik: I'm not sure if there is  abug.  If I don't find one I'll file one today<a href="#t13:47" class="time">13:47</a>
f13jrb: ping; anything from you RE F8 release?<a href="#t13:48" class="time">13:48</a>
davejwwoods: yeah, that sounds good<a href="#t13:48" class="time">13:48</a>
jeremyrdieter: thanks; added to the blocker list<a href="#t13:48" class="time">13:48</a>
wwoodswho's responsible for iscsi stuff? we've got 2 iscsi bugs on the blocker list<a href="#t13:49" class="time">13:49</a>
f13thats probably pjones<a href="#t13:49" class="time">13:49</a>
jrbf13: not really; I'm mostly just logging this today. )-;<a href="#t13:49" class="time">13:49</a>
f13jrb: ok.<a href="#t13:49" class="time">13:49</a>
f13Is there anything anybody else would like to talk about re the F8 release?<a href="#t13:49" class="time">13:49</a>
f13(we should have our release notes soon)<a href="#t13:50" class="time">13:50</a>
wwoodsf13: should we block for the three packages still apparently in need of rebuilding for ppc32 selinux junk?<a href="#t13:50" class="time">13:50</a>
jwbwwoods, which 3 packages?<a href="#t13:50" class="time">13:50</a>
f13wwoods: If they're in the default install path yes, otherwise probably not.<a href="#t13:51" class="time">13:51</a>
wwoodsjwb: syslog-ng, redhat-lsb, perl-BerkeleyDB<a href="#t13:51" class="time">13:51</a>
wwoodsredhat-lsb actually appears to be pulled in on default installs<a href="#t13:51" class="time">13:51</a>
f13hrm, those seem somewhat important.<a href="#t13:52" class="time">13:52</a>
jwbwwoods, do they simply not build at all, or are they just in need of a rebuild?<a href="#t13:52" class="time">13:52</a>
f13well, syslog-ng isn't.<a href="#t13:52" class="time">13:52</a>
f13jwb: they're in need.  They were built during a time that glibc on ppc32 was making everything need execmem<a href="#t13:52" class="time">13:52</a>
* jrb reads<a href="#t13:52" class="time">13:52</a>
f13so if you use those with selinux enabled you will get denials.<a href="#t13:52" class="time">13:52</a>
wwoodsalthough, actually, running lsb_release on ppc64 doesn't throw errors<a href="#t13:52" class="time">13:52</a>
wwoodsweird<a href="#t13:52" class="time">13:52</a>
* davej eyes the number of packages that haven't been rebuilt since fc6<a href="#t13:52" class="time">13:52</a>
jwbf13, ok.  so who owns then?<a href="#t13:52" class="time">13:52</a>
jwber, them<a href="#t13:52" class="time">13:52</a>
f13wwoods: it's not guarenteed that they'll fail.<a href="#t13:53" class="time">13:53</a>
jeremyredhat-lsb got built iirc<a href="#t13:53" class="time">13:53</a>
jeremyyeah, 3 oct<a href="#t13:53" class="time">13:53</a>
f13herm, yeah<a href="#t13:53" class="time">13:53</a>
f13that bug can be closed<a href="#t13:53" class="time">13:53</a>
wwoodsconfirmed, just built it<a href="#t13:54" class="time">13:54</a>
f13mclasen: hey, toss you to the fire, but do you have anything to comment on regarding the push to F8 final?<a href="#t13:56" class="time">13:56</a>
mclasenno comments; dcbw built another nm snapshot last night<a href="#t13:57" class="time">13:57</a>
lmackenf13: I'll be doing a bodhi upgrade shortly, which will then allow people to start adding F8 updates.<a href="#t13:57" class="time">13:57</a>
f13lmacken: thanks.<a href="#t13:57" class="time">13:57</a>
mclasenI don't know if stransky is still trying to get the "plugin installation sucks" bug in<a href="#t13:57" class="time">13:57</a>
f13mclasen: yeah, we'll tag that one.<a href="#t13:57" class="time">13:57</a>
davejnot sure just how good/bad the wireless drivers are this time around, could probably do with making a list of whats been tested/known working/broken<a href="#t13:58" class="time">13:58</a>
jeremydavej: there's a reasonable-ish list in my laptop list :)<a href="#t13:58" class="time">13:58</a>
davejjeremy: ah, good point.<a href="#t13:58" class="time">13:58</a>
jwbjeremy, what laptop list?<a href="#t13:59" class="time">13:59</a>
jeremyjwb: <a href="http://fedoraproject.org/wiki/JeremyKatz/Laptops">http://fedoraproject.org/wiki/JeremyKatz/Laptops</a><a href="#t13:59" class="time">13:59</a>
jwbhm... i should add my stuff there<a href="#t14:00" class="time">14:00</a>
f13Is there anything else?<a href="#t14:00" class="time">14:00</a>
f13ok, Meeting over.<a href="#t14:01" class="time">14:01</a>
davejf13: do we have an ETA on when we can expect to see an rc tree ?<a href="#t14:01" class="time">14:01</a>
f13davej: consider rawhide RC<a href="#t14:01" class="time">14:01</a>
davejok<a href="#t14:01" class="time">14:01</a>
f13(minus signed packages)<a href="#t14:01" class="time">14:01</a>
wwoodstomorrow's rawhide should be installable<a href="#t14:01" class="time">14:01</a>
davejwwoods: todays isn't?<a href="#t14:01" class="time">14:01</a>
wwoodstoday's needs an updates.img, but works otherwise<a href="#t14:01" class="time">14:01</a>
f13davej: with <a href="http://people.redhat.com/clumens/updates.img">http://people.redhat.com/clumens/updates.img</a> today is installable.<a href="#t14:02" class="time">14:02</a>

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