ReleaseEngineering/Meetings/2008-apr-28

Fedora Release Engineering Meeting :: Monday 2008-04-28

Fedora 9

• over 300 packages in dist-f9 that are not f9-final
• 50-55 bugs on the blocker list that aren't in MODIFIED
• serious blocker: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=443445
• Cannot install packages from repositories from which RPM-GPG-KEYs have not been installed
• See log for more details

Release Candidate

• start spinning RCs on Thursday or Friday of this week

IRC Transcript

-!- f13 changed the topic of #fedora-meeting to: Fedora Release Engineering Meeting - ROll Call		<a href="#t13:03" class="time">13:03</a>
rdieter	yo	<a href="#t13:03" class="time">13:03</a>
f13	ping: spot rdieter lmacken	<a href="#t13:04" class="time">13:04</a>
notting	rdieter: whoops, knew i forgot someone. sorry.	<a href="#t13:04" class="time">13:04</a>
* spot is here		<a href="#t13:04" class="time">13:04</a>
* wwoods in the house like furniture		<a href="#t13:04" class="time">13:04</a>
skvidal	wwoods: nailed to the floor and covered with rubber?	<a href="#t13:05" class="time">13:05</a>
-!- f13 changed the topic of #fedora-meeting to: rel-eng meeting: F9!		<a href="#t13:05" class="time">13:05</a>
f13	so... I've been gone for a few days, how's F9?	<a href="#t13:05" class="time">13:05</a>
warren	notting pointed out over 300 packages in dist-f9 that are not f9-final.  Should we be concerned?	<a href="#t13:05" class="time">13:05</a>
warren	is the buildroot f9-final or dist-f9?	<a href="#t13:06" class="time">13:06</a>
f13	warren: buildroot is f9-final	<a href="#t13:06" class="time">13:06</a>
warren	k	<a href="#t13:06" class="time">13:06</a>
f13	I sent out a call to arms a week ago to help with this, a more targetted look	<a href="#t13:06" class="time">13:06</a>
notting	we've got ~50-55 bugs on the blocker list that aren't in MODIFIED	<a href="#t13:06" class="time">13:06</a>
f13	what I really want is to look through the blocker list, see what has been closed since the freeze, and which of those packages are in teh 300 list	<a href="#t13:06" class="time">13:06</a>
wwoods	F9 is... worrying me a bit. I've been monitoring the blocker list	<a href="#t13:06" class="time">13:06</a>
wwoods	AFAICT: in the past week, we've closed 20-30 blockers	<a href="#t13:07" class="time">13:07</a>
wwoods	and added the same number	<a href="#t13:07" class="time">13:07</a>
notting	re: those 300 packages. i was going to send mail with the list. should i take the time to do individual mails to the package maintainers, or just dump to -devel-list?	<a href="#t13:07" class="time">13:07</a>
f13	wwoods: are they actual "blockers"?  I haven't done any incoming review of them	<a href="#t13:07" class="time">13:07</a>
f13	notting: hrm, I assume you'd rather just dump to the list right?	<a href="#t13:08" class="time">13:08</a>
notting	f13: yes :)	<a href="#t13:08" class="time">13:08</a>
wwoods	on the bright (?) side I think we found our traditional "oh shit oh shit" last-minute nastybug	<a href="#t13:08" class="time">13:08</a>
wwoods	bug 443445	<a href="#t13:08" class="time">13:08</a>
buggbot	Bug <a href="https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=443445">https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=443445</a> low, urgent, ---, Robin Norwood, ASSIGNED , Cannot install packages from repositories from which RPM-GPG-KEYs have not been installed.	<a href="#t13:08" class="time">13:08</a>
f13	wwoods: which is funny as "last minute" given that we had signed packages over a week ago	<a href="#t13:08" class="time">13:08</a>
wwoods	I'd say.. probably half the bugs on the blocker list are actually blockerworthy	<a href="#t13:08" class="time">13:08</a>
wwoods	well, okay, maybe there's still room for another last-minute doombug	<a href="#t13:10" class="time">13:10</a>
* rdieter covers ears, la la la		<a href="#t13:11" class="time">13:11</a>
notting	f13: also, when sending this mail, should we move the pkgs to updates-candidate?	<a href="#t13:11" class="time">13:11</a>
wwoods	but in my estimation that's the only bug on the list that is both a) a non-negotiable release blocker, and b) the correct fix would definitely take longer than we have	<a href="#t13:12" class="time">13:12</a>
f13	notting: getting to that later	<a href="#t13:12" class="time">13:12</a>
wwoods	pretty much everything else is, you know.. a one-line fix. it just takes a week to figure out which line to change.	<a href="#t13:13" class="time">13:13</a>
* wwoods exaggerating and oversimplifying		<a href="#t13:13" class="time">13:13</a>
f13	heh	<a href="#t13:14" class="time">13:14</a>
wwoods	speaking of signed packages: did someone sign the 3 unsigned packages (control-center, gnome-desktop, gnome-settings-daemon)	<a href="#t13:14" class="time">13:14</a>
f13	ok, I'll try to roll my sleeves up and do some review/culling today	<a href="#t13:14" class="time">13:14</a>
f13	wwoods: I do believe notting took care of that	<a href="#t13:14" class="time">13:14</a>
notting	yes	<a href="#t13:14" class="time">13:14</a>
wwoods	kcool.	<a href="#t13:14" class="time">13:14</a>
f13	(and also, wasn't that a great test of what the tools will do when an unsigned package is encountered?)	<a href="#t13:14" class="time">13:14</a>
wwoods	heh. yes!	<a href="#t13:14" class="time">13:14</a>
wwoods	clever move, that	<a href="#t13:14" class="time">13:14</a>
warren	do we know how unsigned packages are making it into the mirrors?	<a href="#t13:15" class="time">13:15</a>
warren	those 3 weren't the only unsigned	<a href="#t13:16" class="time">13:16</a>
* poelcat was just going to ask about those :)		<a href="#t13:16" class="time">13:16</a>
f13	warren: the mash output will tell us when there is something unsigned.	<a href="#t13:16" class="time">13:16</a>
warren	f13: was mash doing so a few days ago?	<a href="#t13:16" class="time">13:16</a>
notting	those 3 was because they were tagged on saturday (?) and neither f13 nor I got around to signing them before rawhide went out	<a href="#t13:16" class="time">13:16</a>
wwoods	a few days ago we weren't requiring signed packages	<a href="#t13:16" class="time">13:16</a>
warren	notting: and the case where ltsp went out unsigned is a mystery?	<a href="#t13:16" class="time">13:16</a>
f13	wwoods: yea we were	<a href="#t13:16" class="time">13:16</a>
notting	warren: yeah	<a href="#t13:16" class="time">13:16</a>
wwoods	only after we changed fedora-release	<a href="#t13:17" class="time">13:17</a>
f13	wwoods: right, which was nearly a week ago	<a href="#t13:17" class="time">13:17</a>
wwoods	warren: the basic problem here is that the f9 repo is actually just a redirect to rawhide	<a href="#t13:17" class="time">13:17</a>
warren	For the reason of those 3 packages I'm afraid of tagging anything myself now.	<a href="#t13:18" class="time">13:18</a>
warren	That's why my participation on rel-eng list dropped off	<a href="#t13:18" class="time">13:18</a>
wwoods	so if a package gets tagged without being signed it goes out unsigned	<a href="#t13:18" class="time">13:18</a>
wwoods	which is a problem	<a href="#t13:18" class="time">13:18</a>
warren	There is no guarantee that f13 or notting will be around	<a href="#t13:18" class="time">13:18</a>
wwoods	so, yeah, the tag has to be done by someone with the signing key	<a href="#t13:18" class="time">13:18</a>
* wwoods not tagging anything unless a signer is around anymore		<a href="#t13:18" class="time">13:18</a>
f13	or just don't do a fire-forget	<a href="#t13:18" class="time">13:18</a>
warren	wwoods: there are two separate issues, that, and ltsp which was signed but went out unsigned anyway.	<a href="#t13:18" class="time">13:18</a>
wwoods	yeah I don't know anything about that	<a href="#t13:19" class="time">13:19</a>
f13	warren: what day did it go out unsigned?	<a href="#t13:19" class="time">13:19</a>
warren	f13: I don't know, only noticed it yesterday	<a href="#t13:19" class="time">13:19</a>
notting	f13: it has cached signatures from 04-18. dunno if it had signed packages, because apparently calling write-signed-rpms rewrites them	<a href="#t13:19" class="time">13:19</a>
f13	warren: what arch?	<a href="#t13:20" class="time">13:20</a>
warren	f13: x86_64 for sure, I didn't look at others	<a href="#t13:20" class="time">13:20</a>
f13	checking the mash archives, it doesn't appear to have been sinced since before 4/20 at least	<a href="#t13:21" class="time">13:21</a>
warren	if it is happening for this one package, perhaps it is happening to others/	<a href="#t13:21" class="time">13:21</a>
f13	warren: I don't think so.	<a href="#t13:21" class="time">13:21</a>
f13	but we can quickly check	<a href="#t13:22" class="time">13:22</a>
f13	$ find /pub/fedora/linux/development/ -name "*.rpm"|xargs rpmdev-checksig |grep -v 4f2a6fd2	<a href="#t13:22" class="time">13:22</a>
f13	that's running on releng1, it'll tell us what things are unsigned.  I only expect ltsp	<a href="#t13:23" class="time">13:23</a>
notting	f13: hey, ok if i clean out a couple months of ancient rawhide trees?	<a href="#t13:23" class="time">13:23</a>
f13	notting: yes	<a href="#t13:23" class="time">13:23</a>
f13	ok, we're actually finding a bit more unsigned, I"ll clean these up today	<a href="#t13:24" class="time">13:24</a>
wwoods	are there signers other than f13 and notting?	<a href="#t13:25" class="time">13:25</a>
warren	f13: what type of unsigned?  tagged but not signed, or like ltsp?	<a href="#t13:25" class="time">13:25</a>
f13	warren: it wouldn't be in the development/ directory if it weren't tagged.	<a href="#t13:26" class="time">13:26</a>
notting	f13: i wonder if it's generally unsigned, or write-signed-rpms failed somewhere	<a href="#t13:26" class="time">13:26</a>
f13	warren: likely it's another case where the cached signature exists, but the written copy doesn't for reasons of signed_unsigned being a steaming pile of shit.	<a href="#t13:26" class="time">13:26</a>
notting	haha. markmc already sent the list of new stuff to fedora-devel	<a href="#t13:26" class="time">13:26</a>
f13	notting: nice	<a href="#t13:27" class="time">13:27</a>
wwoods	doh! I'm on the list	<a href="#t13:28" class="time">13:28</a>
f13	hah, somehow I am too	<a href="#t13:28" class="time">13:28</a>
f13	well nwo that we've talked the signing issue to death, lets move on a bit	<a href="#t13:30" class="time">13:30</a>
-!- f13 changed the topic of #fedora-meeting to: rel-eng meeting: dist-f9-updates-candidate		<a href="#t13:30" class="time">13:30</a>
notting	f13: it's created, but unpopulated (and not the default)	<a href="#t13:30" class="time">13:30</a>
f13	we really need to start directing builds to go there, and make bodhi start accepting update requests	<a href="#t13:30" class="time">13:30</a>
f13	however that means we would then have to do two steps when tagging things, move-pkg from dist-f9-updates-candidate to dist-f9, and then tag for f9-final	<a href="#t13:31" class="time">13:31</a>
f13	well, I guess we don't /have/ to, but it does make things cleaner	<a href="#t13:32" class="time">13:32</a>
lmacken	so shall I add F9 to bodhi, and allow for submissions ?	<a href="#t13:32" class="time">13:32</a>
wwoods	people have already tried to build f-9 updates	<a href="#t13:32" class="time">13:32</a>
f13	lmacken: only once we've setup dist-f9-updates-candidate with allt he pending things in dist-f9	<a href="#t13:33" class="time">13:33</a>
lmacken	ok	<a href="#t13:33" class="time">13:33</a>
notting	f13: shall i move over all these things?	<a href="#t13:33" class="time">13:33</a>
f13	and, you'll have to keep it from showing up in teh push lists so that we can continue pusing f7/8 updates	<a href="#t13:33" class="time">13:33</a>
lmacken	f13: yep, there is a 'locked' boolean for each release that does just that	<a href="#t13:33" class="time">13:33</a>
f13	notting: sure, just remember we have to do move-pkg when tagging now	<a href="#t13:33" class="time">13:33</a>
warren	move-pkg from dist-f9-updates-candidate to dist-f9?	<a href="#t13:34" class="time">13:34</a>
f13	warren: yes, and then tag-pkg f9-final	<a href="#t13:35" class="time">13:35</a>
warren	f13: is there any time of night we should be able to tag until without fear?	<a href="#t13:36" class="time">13:36</a>
notting	when did we move rawhide to?	<a href="#t13:37" class="time">13:37</a>
warren	notting: I was referring to "will f13 or notting sign stuff"	<a href="#t13:37" class="time">13:37</a>
f13	warren: just make sure one of us is around to sign.	<a href="#t13:38" class="time">13:38</a>
notting	warren: right, but that's still a relevant data point	<a href="#t13:38" class="time">13:38</a>
notting	midnight eastern should be fine	<a href="#t13:38" class="time">13:38</a>
warren	ok	<a href="#t13:39" class="time">13:39</a>
notting	and possibly later if i'm paying attention	<a href="#t13:39" class="time">13:39</a>
notting	oh good. packages that abort in %post . *sigh*	<a href="#t13:39" class="time">13:39</a>
f13	WHAM	<a href="#t13:40" class="time">13:40</a>
notting	wwoods: one more blocker for you!	<a href="#t13:40" class="time">13:40</a>
wwoods	expletives!	<a href="#t13:40" class="time">13:40</a>
wwoods	oh yeah: the mouse capplet needs a package which wasn't submitted for review until saturday	<a href="#t13:41" class="time">13:41</a>
wwoods	ha ha ha!	<a href="#t13:41" class="time">13:41</a>
notting	f13: ok, will send mail about the shuffle to devel-announce & -evel	<a href="#t13:41" class="time">13:41</a>
wwoods	who do we poke to get that through with the quickness?	<a href="#t13:42" class="time">13:42</a>
f13	wwoods: any packager can review it	<a href="#t13:42" class="time">13:42</a>
f13	wwoods: including you!  (:	<a href="#t13:42" class="time">13:42</a>
wwoods	aw dammit	<a href="#t13:43" class="time">13:43</a>
-!- f13 changed the topic of #fedora-meeting to: rel-eng meeting: release candidates		<a href="#t13:43" class="time">13:43</a>
f13	The schedule has us spinning RCs as of Thursday/Friday of this week	<a href="#t13:44" class="time">13:44</a>
f13	The box that I use to do composes was supposed to get moved into the Boston lab, and subsequently moved early over to the new building and thus be available over the weekend	<a href="#t13:44" class="time">13:44</a>
f13	As of yet, I cannot reconnect to said box	<a href="#t13:44" class="time">13:44</a>
f13	jeremy was handling this, and I do not know what the expected state of things is, so I'm not in panic mode yet	<a href="#t13:45" class="time">13:45</a>
wwoods	what happens if it turns out the box is on fire or something?	<a href="#t13:48" class="time">13:48</a>
wwoods	is there a backup plan?	<a href="#t13:48" class="time">13:48</a>
f13	yeah	<a href="#t13:48" class="time">13:48</a>
notting	we find a different box :)	<a href="#t13:49" class="time">13:49</a>
f13	all I need is an x86_64 machine that can run mock	<a href="#t13:49" class="time">13:49</a>
wwoods	I assume it's just "grab another machine with fat disks"	<a href="#t13:49" class="time">13:49</a>
wwoods	yeah	<a href="#t13:49" class="time">13:49</a>
f13	I think we have... a few of those.	<a href="#t13:49" class="time">13:49</a>
wwoods	'kay	<a href="#t13:49" class="time">13:49</a>
f13	hey guess what?  more corrupted packages!	<a href="#t13:49" class="time">13:49</a>
f13	paraview needs a rebuild.	<a href="#t13:49" class="time">13:49</a>
notting	*sigh*	<a href="#t13:49" class="time">13:49</a>
warren	How dangerous is it to add the entire list of new packages to f9-final?	<a href="#t13:50" class="time">13:50</a>
notting	warren: i wouldn't	<a href="#t13:50" class="time">13:50</a>
pingou	!	<a href="#t13:51" class="time">13:51</a>
f13	warren: uh, no.	<a href="#t13:52" class="time">13:52</a>
f13	pingou: that's not necessary in this meeting.	<a href="#t13:52" class="time">13:52</a>
pingou	sorry I have just seen that <f13> wwoods: any packager can review it but if you need help why not ask around ? I believe you have other things to do than a review no ?	<a href="#t13:52" class="time">13:52</a>
warren	I wasn't suggesting it, just asking about the danger.	<a href="#t13:53" class="time">13:53</a>
wwoods	yeah, I'll probably just ask around in #fedora-devel	<a href="#t13:53" class="time">13:53</a>
f13	warren: thats a rather giant pile of untested packages and we've got 4 working days until RC	<a href="#t13:53" class="time">13:53</a>
wwoods	okay so candidates will get built later this week and we should rock on down the blocker list until then	<a href="#t13:57" class="time">13:57</a>
wwoods	anything else we need to cover? I asked all the questions I've got, I think	<a href="#t13:58" class="time">13:58</a>
rdieter	wwoods: can I have a sec (after meeting is ok)?	<a href="#t13:59" class="time">13:59</a>
wwoods	rdieter: sure	<a href="#t14:00" class="time">14:00</a>
notting	lmacken: want me to poke you when i'm done moving stuff to updates-candidate?	<a href="#t14:00" class="time">14:00</a>
f13	I think that's it	<a href="#t14:00" class="time">14:00</a>

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