ReleaseEngineering/Meetings/2007-jul-30

Fedora Release Engineering Meeting :: Monday 2007-07-30

F8 Test 1 Status

• So far, not so good
• https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=F8Test1
• Blocker evaluation
• f13

1. 247863 is not; but I do believe we can find somebody with time to bang out a fix for this in short order. So while not a blocker, I'd really like to see this fixed as it is very visible. 1. 249597 is totally blocker (insta-reboot issue) 1. 249908 (adding extra repo at install time) not a blocker, just a /really/ nice to have. 1. 250047 is fixed with new policy, just need to clicky the buttons 1. 250130 is no pata support on ppc mac minis (potentially more ppc32 hardware).

• going to spin again this evening with the things we've fixed so far and sync that up too

Summary: Decision as to whether we need to slip Fedora 8 Test 1 will be made tomorrow.

Fedora 8 Test 1 - What to deliver

• Release Engineering will produce the following for Fedora 8 Test1

1. Fedora installable spin

• i386
• x86_64
• ppc
• source

1. Fedora {i386,x86_64} Live image 1. KDE {i386,x86_64 Live} image 1. Everything tree will be provided by standard Rawhide ("development") locations

IRC Transcript

f13 changed the topic of #fedora-meeting to: Fedora Release Engineering - Fedora 8 Test 1		<a href="#t13:01" class="time">13:01</a>
f13	notting: poelcat: warren: rdieter: wwoods: jwb_gone: spot: ping	<a href="#t13:02" class="time">13:02</a>
* notting is here		<a href="#t13:02" class="time">13:02</a>
* spot is here		<a href="#t13:02" class="time">13:02</a>
f13	jeremy is on vacation,	<a href="#t13:02" class="time">13:02</a>
rdieter	here (still)	<a href="#t13:02" class="time">13:02</a>
f13	ooh, /me notices rdieter here.	<a href="#t13:02" class="time">13:02</a>
* wwoods here		<a href="#t13:02" class="time">13:02</a>
f13	so Test 1 status.  Doom-o-meter is scary high	<a href="#t13:03" class="time">13:03</a>
f13	Dell; still b0rked.	<a href="#t13:03" class="time">13:03</a>
* warren here		<a href="#t13:03" class="time">13:03</a>
f13	ppc ide; b0rked.	<a href="#t13:03" class="time">13:03</a>
wwoods	not just confined to Dell anymore - someone reported the same problem on a toshiba laptop	<a href="#t13:03" class="time">13:03</a>
warren	Any bug # for the Dell issue?  Do we have ANY idea what's going on?	<a href="#t13:03" class="time">13:03</a>
f13	wheee	<a href="#t13:03" class="time">13:03</a>
wwoods	warren: x86 setup code was rewritten post-2.6.22	<a href="#t13:04" class="time">13:04</a>
f13	adding external repos at install time; b0rked (clumens is working on it now)	<a href="#t13:04" class="time">13:04</a>
warren	wwoods, so in-kernel definitely?	<a href="#t13:04" class="time">13:04</a>
notting	f13: in the ui, or in the backend (does ks work?)	<a href="#t13:04" class="time">13:04</a>
wwoods	warren: definitely. bug #249597	<a href="#t13:04" class="time">13:04</a>
f13	firstboot; b0rked in multiple ways (non-fatal traceback; won't launch if rhgb doesn't run)	<a href="#t13:04" class="time">13:04</a>
f13	notting: haven't tried ks.  In the UI there is a traceback when it starts to add package objects to the transaction set.	<a href="#t13:05" class="time">13:05</a>
f13	We didn't have an installable rawhide until yesterday, we /still/ don't have an installable ppc rawhide, but those 4 users wouldn't notice anyway.	<a href="#t13:05" class="time">13:05</a>
notting	do we have installable ppc test trees/	<a href="#t13:06" class="time">13:06</a>
f13	yes	<a href="#t13:06" class="time">13:06</a>
wwoods	Can I propose that the pre-Test1 freeze for each release from now on be, like, two weeks?	<a href="#t13:06" class="time">13:06</a>
wwoods	Because damn.	<a href="#t13:06" class="time">13:06</a>
f13	on reducto and on koji.fp.o/rel-eng/trees/	<a href="#t13:06" class="time">13:06</a>
notting	wwoods: surely part of the solution is to start testing pre-freeze	<a href="#t13:07" class="time">13:07</a>
f13	wwoods: not a /horrible/ idea, but sucks as developers to have non-critical devel/ builds stagnate.	<a href="#t13:07" class="time">13:07</a>
wwoods	notting: indeed, indeed	<a href="#t13:07" class="time">13:07</a>
f13	notting: yes.  I thought I was doing testing, but obviously not sufficient enough	<a href="#t13:07" class="time">13:07</a>
f13	I was keeping the tool working, rather than testing the output (which is also important)	<a href="#t13:07" class="time">13:07</a>
f13	that and changing the toolchain after the freeze Did Not Help.	<a href="#t13:07" class="time">13:07</a>
f13	On the plus side, on Friday the Live images stuff still looked pretty good	<a href="#t13:08" class="time">13:08</a>
f13	including installation from them.	<a href="#t13:08" class="time">13:08</a>
wwoods	yeah we really need to be doing nightly install tests on rawhide.	<a href="#t13:08" class="time">13:08</a>
wwoods	it is unavoidable. Once we cram t1 out the door I'm gonna set that up again.	<a href="#t13:08" class="time">13:08</a>
f13	right now clumens may have fixed nfs-iso method, and is working on the adding repos thing.	<a href="#t13:08" class="time">13:08</a>
wwoods	(err work on setting that up again)	<a href="#t13:09" class="time">13:09</a>
f13	after that I may have him raise the bar for doing boot.iso gui http installs as I can reliably freeze a box with 256 megs	<a href="#t13:09" class="time">13:09</a>
f13	that's not a high priority though so any new anaconda fun will go in front of that.	<a href="#t13:09" class="time">13:09</a>
f13	oh wait, we should have him poke at firstboot instead.	<a href="#t13:10" class="time">13:10</a>
wwoods	firstboot traceback is a s-c-securitylevel bug	<a href="#t13:10" class="time">13:10</a>
f13	wwoods: there are other problems	<a href="#t13:10" class="time">13:10</a>
wwoods	looks like a fairly easy patch but nobody seems to have gotten around to it	<a href="#t13:10" class="time">13:10</a>
f13	like, it can't launch an X of it's own if rhgb didn't launch	<a href="#t13:10" class="time">13:10</a>
wwoods	really? great.	<a href="#t13:10" class="time">13:10</a>
f13	and not for the same reasons we just fixed rhgb for	<a href="#t13:11" class="time">13:11</a>
warren	f13, do we have any affected Dell machines in the office?	<a href="#t13:11" class="time">13:11</a>
f13	warren: cebbert has one.	<a href="#t13:11" class="time">13:11</a>
f13	which is rather fitting.	<a href="#t13:11" class="time">13:11</a>
warren	good	<a href="#t13:11" class="time">13:11</a>
wwoods	oh cebbert is in the "boston" office? cool	<a href="#t13:12" class="time">13:12</a>
f13	yeah	<a href="#t13:12" class="time">13:12</a>
f13	so we need a couple decisions	<a href="#t13:12" class="time">13:12</a>
f13	we need to figure out which of these problems are full-stop release blockers, and if so, what our deadline is to have them fixed before we introduce slippage.	<a href="#t13:12" class="time">13:12</a>
wwoods	<a href="https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=F8Test1">https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=F8Test1</a> for those of you keeping score at home.	<a href="#t13:12" class="time">13:12</a>
f13	I think we can easily come up with the point of no return for slippage, and that would be having it fixed tonight.	<a href="#t13:13" class="time">13:13</a>
f13	to spin trees early morning, and even that is cutting it close.	<a href="#t13:13" class="time">13:13</a>
f13	(also we need some consensus on what all we're going to deliver, but we'll get to that)	<a href="#t13:13" class="time">13:13</a>
wwoods	so, examining the old standby <a href="http://fedoraproject.org/wiki/QA/ReleaseCriteria">http://fedoraproject.org/wiki/QA/ReleaseCriteria</a>	<a href="#t13:14" class="time">13:14</a>
warren	sorry back, connection died	<a href="#t13:15" class="time">13:15</a>
wwoods	test release bare-minimum is: boot and install on all arches, yum upgrade works, no data corruptors, etc.	<a href="#t13:16" class="time">13:16</a>
wwoods	and currently we don't boot on a whole mess of machines	<a href="#t13:16" class="time">13:16</a>
f13	yes, that is definitely a blocker	<a href="#t13:16" class="time">13:16</a>
wwoods	technically, though, firstboot is not part of our current release criteria	<a href="#t13:16" class="time">13:16</a>
f13	bug 247863 is not; but I do believe we can find somebody with time to bang out a fix for this in short order.  So while not a blocker, I"d really like to see this fixed as it is very visible.	<a href="#t13:16" class="time">13:16</a>
wwoods	I'd suggest we might change the criteria in that area though	<a href="#t13:17" class="time">13:17</a>
f13	249597 is totally blocker (insta-reboot issue)	<a href="#t13:17" class="time">13:17</a>
f13	249908 (adding extra repo at install time) not a blocker, just a /really/ nice to have.	<a href="#t13:17" class="time">13:17</a>
f13	250047 is fixed with new policy, just need to clicky the buttons	<a href="#t13:18" class="time">13:18</a>
f13	and 250130 is no pata support on ppc mac minis (potentially more ppc32 hardware).  While technically could be considered a blocker, I think the vast amounts of users that would be effected by this can just suffer.	<a href="#t13:18" class="time">13:18</a>
f13	does anybody else have any other opinions regarding the above bugs?	<a href="#t13:19" class="time">13:19</a>
notting	well, we're waiting in 249597. so, i'm not sure any of the others need to be moved off the blocker as long as that's the case	<a href="#t13:19" class="time">13:19</a>
wwoods	249908 is final-blocker, not a test-blocker.. technically ppc is still a first-class arch so 250130 would probably be a blocker	<a href="#t13:20" class="time">13:20</a>
f13	notting: no, I'm not moving any off, but I just want to set the expectation that if the insta-reboot gets fixed, we're not going to hold up on any of the other nonblockers	<a href="#t13:20" class="time">13:20</a>
wwoods	but yeah, it's all moot because 249597 is a full-fledged stop-the-presses blocker	<a href="#t13:20" class="time">13:20</a>
f13	(there is a reason the tracker is just 'F8Test1' and not 'F8Test1Blockers'	<a href="#t13:20" class="time">13:20</a>
f13	but obviously if we can't fix 249597 today, we're slipping.	<a href="#t13:21" class="time">13:21</a>
wwoods	so let's talk firstboot - we've got *two* bugs now? can't start its own X, and the s-c-securitylevel traceback?	<a href="#t13:21" class="time">13:21</a>
wwoods	f13: right.	<a href="#t13:21" class="time">13:21</a>
f13	wwoods: yes.  Arguably the traceback isn't a firstboot bug, just one triggered /by/ firstboot.	<a href="#t13:22" class="time">13:22</a>
f13	I should hope it would be pretty straight forward to fix s-c-securitylevel.  Obviously the maintainer isn't paying attention to anything (perhaps time to ring somebody's bell?)	<a href="#t13:22" class="time">13:22</a>
wwoods	right, yes, it's not a firstboot but, but it prevents firstboot from running	<a href="#t13:23" class="time">13:23</a>
wwoods	err "firstboot bug, but..."	<a href="#t13:23" class="time">13:23</a>
f13	wwoods: no it doesn't	<a href="#t13:23" class="time">13:23</a>
wwoods	aroo? it did for me	<a href="#t13:23" class="time">13:23</a>
f13	wwoods: you click "ok" or whatever on the error popup and firstboot continues on happily	<a href="#t13:23" class="time">13:23</a>
wwoods	huh! didn't know that.	<a href="#t13:23" class="time">13:23</a>
wwoods	mea culpa then	<a href="#t13:23" class="time">13:23</a>
f13	it's still pug ugly and we'll get a shitton of email/bugreports/nastygrams about it so we should fix it	<a href="#t13:24" class="time">13:24</a>
wwoods	yeah, I'd like to poke the s-c-securitylevel maintainer about it	<a href="#t13:24" class="time">13:24</a>
f13	failing that, notting can you pinch-hit here?	<a href="#t13:24" class="time">13:24</a>
wwoods	it seems like we'll be slipping for the dell bug anyway so we might have a window to fix the firstboot problems	<a href="#t13:24" class="time">13:24</a>
notting	f13: for? twoerner?	<a href="#t13:25" class="time">13:25</a>
notting	afaik, he's around somewhere	<a href="#t13:25" class="time">13:25</a>
f13	notting: if that's the owner of s-c-securitylevel yes.  I'm a bit peeved that this hasn't been fixed yet, but I don't know what his current workload is.  What would have been useful is dropping a note that he couldn't fix it in time, get somebody else to look at it.	<a href="#t13:25" class="time">13:25</a>
f13	hey good, clumens duped the problem when using an added repo at install time.  There is potential for a fix there.	<a href="#t13:28" class="time">13:28</a>
notting	f13: have you tracked him down on irc?	<a href="#t13:28" class="time">13:28</a>
f13	notting: no, I've bene chasing other fires.	<a href="#t13:28" class="time">13:28</a>
wwoods	what about the firstboot-cant-start-X thing?	<a href="#t13:30" class="time">13:30</a>
wwoods	do we have someone on that?	<a href="#t13:30" class="time">13:30</a>
notting	bug#?	<a href="#t13:30" class="time">13:30</a>
f13	so that's where we're at.  I'm going to poke a bit at firstboot to look at why it can't launch X ( at least get some logging out of it )	<a href="#t13:31" class="time">13:31</a>
wwoods	f13: is that filed?	<a href="#t13:31" class="time">13:31</a>
f13	wwoods: I don't think so.  I probably should once I figure out how it's actually not happening	<a href="#t13:33" class="time">13:33</a>
wwoods	f13: righteous. I'm sure I'll get mail when you add it to the tracker. Lemme know if you need help reproducing.	<a href="#t13:34" class="time">13:34</a>
f13	k	<a href="#t13:34" class="time">13:34</a>
wwoods	I'm probably gonna mess around with PPC installs some more - the tree on koji.fp.o is up-to-date, right?	<a href="#t13:34" class="time">13:34</a>
f13	and it looks like twoerner is actually working on s-c-sl right now.	<a href="#t13:34" class="time">13:34</a>
wwoods	cool!	<a href="#t13:34" class="time">13:34</a>
f13	wwoods: yes, that's what I spun yesterday and there were no package changes since then.	<a href="#t13:35" class="time">13:35</a>
f13	I'm going to spin again this evening with the things we've fixed so far and sync that up too	<a href="#t13:35" class="time">13:35</a>
f13 changed the topic of #fedora-meeting to: Fedora Release Engineering - Fedora 8 Test 1 - What to deliver		<a href="#t13:36" class="time">13:36</a>
f13	So for Fedora 7, we did a different content set each test release as we were feeling our way toward the final set which was the Fedora spins, and a set of Everything repos	<a href="#t13:36" class="time">13:36</a>
f13	Ideally, from a 'getting bits spun and in place' for the test release we'd just do the Fedora spin and rely upon rawhide as our 'Everything' tree.	<a href="#t13:37" class="time">13:37</a>
f13	(and for our debuginfo tree)	<a href="#t13:37" class="time">13:37</a>
wwoods	sounds reasonable.	<a href="#t13:37" class="time">13:37</a>
warren	"Everything" is a checkbox during anaconda?	<a href="#t13:37" class="time">13:37</a>
f13	we'd still probably have to spin source isos for the FEdora spin for those pesky legal reasons but...	<a href="#t13:37" class="time">13:37</a>
f13	warren: not there currently	<a href="#t13:38" class="time">13:38</a>
f13	warren: probably want to land the ability to use mirrorlists in anaconda add repos first	<a href="#t13:38" class="time">13:38</a>
f13	and there are some interesting side effects to adding a second repo that duplicates a lot of the content	<a href="#t13:38" class="time">13:38</a>
f13	like, what repo is yum going to pull the packages from?	<a href="#t13:38" class="time">13:38</a>
f13	seth has got a 'priorities' like plugin that I need to test that we'd want to use in anaconda to ensure that the main repo (install method) is used for all possible packages, only going to add-on repos for things not available in the main repo	<a href="#t13:39" class="time">13:39</a>
* rdieter likes the sound of that.		<a href="#t13:39" class="time">13:39</a>
f13	yeah, can be quite the useful little plugin.  Originally I wanted it for mock	<a href="#t13:40" class="time">13:40</a>
f13	but I'm finding other places where it would be really useful.	<a href="#t13:40" class="time">13:40</a>
f13	Does anybody see a reason why we should deliver anything more than 'Fedora installable spin i386,x86_64,ppc + source, Fedora i386,x86_64 Live image, and KDE i386,x86_64 Live image'	<a href="#t13:41" class="time">13:41</a>
wwoods	sounds fine	<a href="#t13:42" class="time">13:42</a>
wwoods	Someday we should talk about delivering just live media, in two sizes (CD and DVD). But not now.	<a href="#t13:42" class="time">13:42</a>
rdieter	the only possible addition would be Everything spin? (or something else?)	<a href="#t13:43" class="time">13:43</a>
f13	wwoods: that has deep hurt until you can do upgrades from Live	<a href="#t13:43" class="time">13:43</a>
f13	rdieter: yeah, Everything is what's missing.	<a href="#t13:43" class="time">13:43</a>
f13	<@twoerner> jk_meeting: 1.7.0-4_fc8 is building	<a href="#t13:43" class="time">13:43</a>
f13	*ROCK*	<a href="#t13:43" class="time">13:43</a>
f13	Ok, so	<a href="#t13:44" class="time">13:44</a>
f13 changed the topic of #fedora-meeting to: Fedora Release Engineering - Fedora 8 Test 1 - Open Discussion'		<a href="#t13:44" class="time">13:44</a>
poelcat	f13: what was summary on last topic?	<a href="#t13:44" class="time">13:44</a>
f13	anything else related to Test1 that anybody would like to discuss?	<a href="#t13:44" class="time">13:44</a>
f13	poelcat: rel-eng will deliver for Test1 'Fedora installable spin i386,x86_64,ppc + source, Fedora i386,x86_64 Live image, and KDE i386,x86_64 Live image'	<a href="#t13:44" class="time">13:44</a>
poelcat	f13: thanks!	<a href="#t13:45" class="time">13:45</a>
wwoods	f13: right, that's the "someday"	<a href="#t13:45" class="time">13:45</a>
wwoods	heh	<a href="#t13:45" class="time">13:45</a>
f13	poelcat: with a side note "Everything tree will be provided by standard Rawhide locations"	<a href="#t13:45" class="time">13:45</a>
f13	If there's nothing else I'd like to call the meeting and get back to work.	<a href="#t13:45" class="time">13:45</a>
f13	Any objections?	<a href="#t13:45" class="time">13:45</a>
wwoods	we should have that as a checkbox in the "additional repo" list for test2 and on	<a href="#t13:45" class="time">13:45</a>
f13	wwoods: as stated before, we need a few things first; support for mirror lists; priority plugin to prevent you from skipping local media and downloading everything off the net	<a href="#t13:46" class="time">13:46</a>
f13	wwoods: not totally unreasonable, and I'd consider throwing up a Feature page for this	<a href="#t13:47" class="time">13:47</a>
wwoods	or some other way of tracking it, yeah	<a href="#t13:47" class="time">13:47</a>
poelcat	f13: what was recap for first topic?... wait and see how tonight looks?	<a href="#t13:47" class="time">13:47</a>
wwoods	we're blocked on the Dell bug, which won't be fixed tonight, so we'll be using this window to try to get fixes in for the other tracked issues	<a href="#t13:48" class="time">13:48</a>
f13	poelcat: yes, decision on slipping will be made tomorrow.	<a href="#t13:48" class="time">13:48</a>
* poelcat attempts to publish minutes in near-real-time for a change		<a href="#t13:49" class="time">13:49</a>
f13	poelcat: awesome.  Thank you again so much for this.	<a href="#t13:50" class="time">13:50</a>
poelcat	f13: you're welcome :)	<a href="#t13:50" class="time">13:50</a>

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