QA/Meetings/20121126

Attendees

• adamw (135)
• tflink (87)
• Viking-Ice (21)
• kparal (20)
• maxamillion (10)
• jreznik_n9 (6)
• Martix (6)
• zodbot (6)
• satellit_e (3)
• robatino (2)
• pschindl (2)
• mkrizek (1)
• Cerlyn_ (1)

Agenda

• Previous meeting follow-up
• Fedora 18 Beta review / Final planning
• Test case / criteria revision
• Open floor

Previous meeting follow-up

• tflink to continue to evaluate fedup status and propose significant bugs for blocker status - this was done, fedup has rough edges but is basically working
  • Docs team was taking care of fedup documentation, tflink was to make sure something would be in place for Beta

Fedora 18 Beta review / Final planning

• Common bugs needed updating as always
• tflink was worried about state of fedup, adamw said he would try to keep an eye on forums
• Agreed to go ahead and start holding Final blocker review meetings immediately
• Final TC1 was scheduled for 2012-12-11 at the time
• Agreed to build smoketest images regularly until TC1 point

Test case / criteria revision

• Some discussion about fedup requirements for final release
  • Agreed tentatively that GUI and progress monitoring must be done for fedup for final
• Agreed to make jskladan check the final criteria / test cases for revision candidates

Open floor

• tflink proposed signing smoketest images, after discussion, general agreement that it was not worth the effort
• Some confusion over the exact proposed final mechanism for fedup image generation and provision etc
• Another idea from tflink: use zsync for smoketest images to reduce bandwidth load - some technical discussion, not at a point worth making decisions about

Action items

• tflink to ensure some kind of upgrade documentation is ready for beta availability tomorrow
• tflink to brief #fedora ops and fedora-user-list regulars on fedup
• adamw to co-ordinate with anaconda team on TC1 date planning
• jskladan to review final criteria and test cases for obvious revision candidates
• viking-ice or tflink to try and get a fedup design document out of wwoods

IRC Log

adamw	#startmeeting Fedora QA meeting	16:01
zodbot	Meeting started Mon Nov 26 16:01:56 2012 UTC. The chair is adamw. Information about MeetBot at http://wiki.debian.org/MeetBot.	16:01
zodbot	Useful Commands: #action #agreed #halp #info #idea #link #topic.	16:01
adamw	#meetingname fedora-qa	16:01
zodbot	The meeting name has been set to 'fedora-qa'	16:01
adamw	#topic roll call	16:02
* kparal is here		16:02
* tflink is here		16:02
* Cerlyn_ is here		16:02
* maxamillion is here		16:02
* adamw is barely here		16:02
* satellit_e listening		16:02
tflink	adamw: that's what coffee's for :)	16:02
adamw	mmm, glorious coffee.	16:02
* mkrizek is here		16:03
* jreznik_n9 is not available today... day off		16:03
maxamillion	adamw: +1	16:03
adamw	.fire jreznik	16:03
zodbot	adamw fires jreznik	16:03
adamw	#topic previous meeting follow-up	16:03
adamw	alright, a nice batting practice fastball for tflink here...	16:04
tflink	uh oh	16:04
adamw	...just as soon as i can get to the wiki.	16:04
adamw	"tflink to continue to evaluate fedup status and propose significant bugs for blocker status"	16:05
tflink	oh, yeah. that is done for the time being	16:05
tflink	fedup still has some rough edges but in general, it seems to be working	16:05
adamw	we've managed to keep the victims' families away from the press, so all is well	16:06
kparal	will we have some upgrade documentation in time for Beta release?	16:06
tflink	yeah, that's on my todo list for today	16:06
jreznik_n9	rough edges...	16:06
adamw	#info "tflink to continue to evaluate fedup status and propose significant bugs for blocker status" - this was done, fedup has rough edges but is basically working	16:06
tflink	either will, I or the docs team will get something done before beta is released tomorrow	16:06
jreznik_n9	we also need a todo list for final	16:07
tflink	finishing the test plan and more test cases would be good, too	16:07
adamw	kparal: i mailed docs list to ask them to look after it, jack reed said he will do it	16:07
jreznik_n9	to be on the same side of force	16:07
adamw	"I will keep in touch with Will and tflink to ensure fedup is documented	16:07
adamw	accurately."	16:07
* pschindl is late, but here		16:07
adamw	not sure if he meant for beta, though	16:07
adamw	#info docs team taking care of fedup documentation, but not sure if they'll be ready for beta	16:07
adamw	#action tflink to ensure some kind of upgrade documentation is ready for beta availability tomorrow	16:08
adamw	since we're already on it anyway...	16:08
tflink	yeah, that was the impression I got too. not sure if the docs would be done for beta release	16:08
kparal	in the CommonBugs page I referenced http://fedoraproject.org/wiki/Upgrading a few times	16:08
adamw	#topic Fedora 18 Beta review / Final planning	16:08
adamw	oh, you started on commonbugs? awesome, thanks	16:08
* tflink makes note to update that page		16:09
adamw	#info commonbugs needs updating for various beta bugs, kparal already made a start	16:09
adamw	for anyone else who wants to help out - the page is https://fedoraproject.org/wiki/Common_F18_bugs , see the comments for instructions and style guide, and compare to existing entries	16:10
jreznik_n9	especially nfs thing has to be mentioned	16:10
* jreznik_n9 could go through the page tmrw in the morning for final review, touches		16:11
adamw	please, everyone who has something to contribute, do :)	16:11
adamw	and make sure any significant bugs are tagged with the CommonBugs keyword	16:11
adamw	anything else anyone is worried about for beta?	16:12
tflink	I'm a little worried about fedup for beta but a lot of that is due to the change from preupgrade/DVD	16:14
tflink	I think there are going to be a lot of questions	16:14
adamw	i think that was pretty inevitable	16:14
adamw	i'll try and stay on top of the forums, any help welcome	16:14
adamw	the other thing we could do is make sure the folks from #anaconda and user-list are briefed	16:14
adamw	do you want to do that, once we have the docs in place, or should i?	16:15
tflink	briefed? on where the docs are?	16:15
tflink	I can do the coordination on fedup docs, though	16:15
tflink	do/help with	16:15
adamw	yeah, just let them know fedup is landing, here are the docs, you might want to try it out so you know what you're talking about to help people, here's who to ask if you get stuck	16:16
adamw	that kinda thing	16:16
tflink	ok, I think that finding a few #fedora regulars would be wise as well	16:18
adamw	d'oh, that's what I meant, not #anaconda obviously :)	16:18
adamw	#info tflink to brief #fedora ops and fedora-user-list regulars on fedup	16:18
adamw	grr	16:18
adamw	#undo	16:19
zodbot	Removing item from minutes: <MeetBot.items.Info object at 0xa3f77d0>	16:19
adamw	#action tflink to brief #fedora ops and fedora-user-list regulars on fedup	16:19
adamw	ok, anything else on beta?	16:19
tflink	other than bracing myself for final testing, nothing from me :)	16:19
maxamillion	heh ... I apparently hit "pageup" and irssi scrolled up but then I didn't notice and was staring at an earlier part of the convo thinking to myself "this meeting seems slower than usual" .... /me facepalms (sorry for the offtopic, just had to share)	16:20
adamw	that's the kind of top-quality staff we at rh insist on, folks	16:20
maxamillion	>.>	16:20
maxamillion	\o/	16:20
adamw	(says adamw from his dining table, in dressing gown, with cat at his side)	16:21
maxamillion	adamw: I don't even have a good excuse	16:21
adamw	ok, so for final planning	16:21
adamw	since i know everyone loves them so much, i was thinking we could indulge ourselves and start blocker review this week	16:21
tflink	yeah, I was thinking the same thing	16:21
adamw	i mean i know it barely seems worth it with the tiny final blocker list - http://qa.fedoraproject.org/blockerbugs/milestone/18/final/buglist - but still...	16:22
adamw	(dies)	16:22
tflink	just to at least start getting through the huge list we have thus far	16:22
adamw	so who's ready for some exciting blocker review on wednesday?	16:22
maxamillion	o/	16:22
tflink	maxamillion: says a guy who doesn't regularly show up for the review meetings :-D	16:23
maxamillion	tflink: I know :( ... I've been terrible at that this release	16:23
adamw	i think he probably just facepalmed on the keyboard again.	16:23
tflink	how do you facepalm on the keyboard?	16:24
maxamillion	silly $dayjob keeps getting in the way	16:24
adamw	tflink: it takes skills, to be sure.	16:24
adamw	okay, the other thing we have is the Final TC date	16:24
adamw	#info first Final blocker review meeting this Wednesday, 11-28	16:25
tflink	do we still have test cases and/or criteria to revise for final?	16:25
adamw	yes, i have a topic coming up for that next	16:25
tflink	ok	16:25
adamw	mumblegrumblepeoplewhodon'treadtheagendamumble	16:25
tflink	agenda?	16:25
tflink	:-D	16:25
adamw	=)	16:25
adamw	so final tc1 is currently scheduled for 12-11, two weeks from now	16:26
Martix	when final TC images starts releasing?	16:26
adamw	with go/no-go on 01-01	16:26
adamw	assuming jreznik's space is correct, anyhow	16:26
Martix	adamw: ok	16:26
tflink	new years day? whose bright idea was that?	16:26
adamw	fesco?	16:26
adamw	blame fesco!	16:26
tflink	sounds like a plan to me	16:26
Martix	lol	16:27
Martix	01-01? certainly No-Go	16:27
tflink	I'd like to get testable images started as soon as we have another anaconda build	16:27
tflink	whether that's in the format of smoke images or TCs	16:27
adamw	so that'd give us three weeks for validation. we could move it up yet further just to make us all even more sick of validation. but it might be better, i guess, to look at that after we do one blocker review meeting at least, and co-ordinate with anaconda team, see what their plans are	16:27
Martix	tflink: +1	16:27
adamw	it might be a solid plan to have smokes all the way, tough	16:28
tflink	on a side note, work is progressing somewhat on the image building project from GSoC	16:28
adamw	#agreed we should build at least smoketest images throughout final phase, from now onwards	16:28
tflink	I don't think it's going to be done before F18 is over with but I'm hoping to start using it for smoke images soon	16:28
adamw	awesome!	16:28
adamw	#info imagebuilder may be usable for Final smoketest images	16:29
tflink	ideally, I'll be able to get something set up to watch for anaconda builds and trigger new smoke images but that's more than likely to be "crazy fantasy land"	16:29
adamw	#action adamw to co-ordinate with anaconda team on TC1 date planning	16:29
adamw	okay, anything else on 18 beta/final before we go on to test case / criteria stuff?	16:30
* Viking-Ice scrolls up		16:31
* adamw gives viking-ice a minute		16:32
Viking-Ice	let's carry on	16:33
adamw	okey dokey, yell if you think of anything	16:34
adamw	#topic Test case / criteria revision	16:34
adamw	so i just put this in as a kind of catch-all topic	16:34
adamw	i know I still suck and haven't figured out the partitioning criteria, sorry.	16:34
tflink	.fire adamw	16:35
zodbot	adamw fires adamw	16:35
maxamillion	awwww, sadface	16:35
tflink	one question I have is whether we're going to block final if there's no fedup gui	16:35
adamw	I'M FREE!	16:35
* adamw dances off into distance, clicking heels		16:35
tflink	wait, that's how it works?	16:35
Viking-Ice	tflink, I would say so yes	16:36
Viking-Ice	no gui no release	16:36
adamw	#info kparal and jskladan have been cleaning up several test cases for final	16:36
adamw	i'm inclined to side with viking-ice on that one	16:36
adamw	though not sure if we want to rope in fesco, or what wwoods' take is	16:36
Viking-Ice	tflink, just feed wwoods more alcohol	16:36
tflink	do we want to refine the upgrade criteria for upgrades so that there is less ambiguity when issues arise	16:36
Viking-Ice	;)	16:36
tflink	ie, fedup theme, visable progress during upgrade, hot-fixability post-release etc.	16:36
adamw	tflink: i'm not so sure if i'd want to go that route as it's not like we'll be rewriting the gui at every release, but that's just mho	16:37
tflink	er, s/visable progress/easily monitor-able progress/	16:37
tflink	I'm less worried about the gui than I am about plymouth integration and progress monitoring	16:37
adamw	yeah, progress monitoring seems like a big thing to have too.	16:38
Viking-Ice	progress monitoring is a must have	16:38
* tflink will propose as a final blocker		16:38
adamw	let's make a wishlist	16:38
tflink	wishlist or a musthave list?	16:38
adamw	#agreed general consensus that we expect a) progress monitoring and b) GUI for fedup to be complete for final release	16:38
* adamw handwaves		16:38
Viking-Ice	I dont think it has to show which package is being upgraded thou	16:39
adamw	always leave room to back out later!	16:39
adamw	no, just some sort of 'i'm alive, don't reboot me'.	16:39
Viking-Ice	ack	16:39
adamw	and vague 0-100 of some kind.	16:39
tflink	Viking-Ice: that would be nice but I agree that it's probably not worth blocking for	16:39
adamw	oop, sorry, i didn't propose, but sounds like it's okay.	16:39
tflink	this might be better for open floor, but I've been wondering about signing more of the test images - smoke builds and fedup stuff	16:40
adamw	let's make it open floor yeah	16:40
Viking-Ice	tflink, honestly the whole anaconda showing which package is being installed never made sense to me	16:40
Viking-Ice	if the install fails it fails regardless which package was being installed	16:40
adamw	Viking-Ice: in Beta RC1 it doesn't, actually - at least on live	16:40
adamw	it gives you a percentage	16:40
adamw	which is neat.	16:41
adamw	okay, back on criteria!	16:41
tflink	eh, I like seeing which package is being installed - it's a quick indication if the process hangs but I guess it's personal preference	16:41
tflink	either way, not worried about it :)	16:41
adamw	does anyone want to take an action item to review the final criteria and test cases for particularly obvious revision candidates?	16:41
adamw	or should i just assign it to jskladan in his absence? :)	16:42
tflink	that'll teach him not to show up :-P	16:43
* kparal agrees		16:43
adamw	eeeeexcellent	16:43
adamw	#action jskladan to review final criteria and test cases for obvious revision candidates	16:43
adamw	can you let him know, kparal? thanks	16:43
kparal	adamw: sure	16:43
kparal	pschindl: can you remind me that tomorrow?	16:44
kparal	pschindl: don't forget to delegate this process further	16:44
tflink	another action for josef?	16:45
adamw	some chicken farmer in swaziland is going to wake up with a todo item for this tomorrow, isn't he	16:45
tflink	#action jskladan to remind pschindl to remind kparal to let jskladan know that he got an action item	16:46
adamw	i like it!	16:46
pschindl	kparal: I will remind you ;-)	16:46
adamw	rhl: lean, mean, efficiency machine	16:46
tflink	the circle is now complete ...	16:46
adamw	okay, that sounds good	16:47
adamw	anything else before we go on to open floor?	16:47
Martix	ack	16:47
adamw	#topic open floor	16:48
satellit_e	https://bugzilla.redhat.com/show_bug.cgi?id=880300 easy way to format USB with no swap and / partitions (non LVM)	16:48
adamw	fire away with crazy ideas now!	16:48
tflink	I've been wondering about signing more of the test images that we're using	16:49
adamw	satellit: i saw that. to be honest it doesn't make any sense. it's as simple to make non-LVM partitions as any other kind.	16:49
adamw	#topic open floor - image signing	16:49
tflink	I'm just not sure it's worth the effort	16:49
adamw	tflink: this comes under 'desirable but not critical' for me, especially if it'd make the process any heavier	16:49
satellit_e	ok	16:49
adamw	i really would hope that no-one is installing smoketest images and then actually *using* them for anything	16:49
tflink	yeah, that's mostly why I'm wondering if it would be worth any effort	16:50
Viking-Ice	agreed I think signing smokes is useless	16:50
Viking-Ice	smokes are kinda use once discard	16:50
tflink	fedup testing is more arguable, though	16:50
adamw	so i'd say don't waste more than minimal time on it, and especially not worth it if it makes it more complex/time-consuming to build smokes	16:50
adamw	we have a bug for fedup signing, right?	16:51
tflink	yeah, that's as much a releng issue, though	16:51
tflink	if we did do it, I'd wonder what key to use	16:51
tflink	I can sign them myself but that doesn't seem quite right and there's no way I'm putting my gpg key on a regular infra machine	16:52
adamw	another darn thing.	16:52
tflink	I'll keep it in mind but I suspect that there are going to be higher priorities for final	16:52
adamw	so when you say fedup signing, signing what exactly? there are various bits, right? are we talking the upgrade.img here?	16:53
tflink	the upgrade.img, yeah	16:53
adamw	ok	16:53
adamw	well i guess the process i'd expect there is that it gets generated by releng and signed by releng, but i may be in cuckoo land	16:53
tflink	I suppose that if we're not testing stuff only in git, it doesn't matter a whole lot, though	16:53
tflink	I'm tempted to not do it unless there is enough complaining to justify the effort or we need to start testing some sort of verification	16:54
Viking-Ice	do we have any page that outlines the upgrade.img and it's process?	16:55
tflink	building or using?	16:55
tflink	upgrade documentation is on my todo list for today, already have an action item or two	16:55
Viking-Ice	both fedup progress in general how it (supposed to ) work etc	16:56
tflink	I'm planning to update the testing page once I have some basic usage docs in place	16:56
adamw	we already did some planning for documenting the user side of things	16:56
tflink	that's what I've been trying to use to record progress, what to expect etc.	16:57
adamw	that was before you came in	16:57
* tflink has another item for open floor once we're done with this		16:57
Viking-Ice	we need documentation on the whole process not just the user side of things	16:57
tflink	Viking-Ice: what do you mean "the whole process"?	16:58
tflink	that's rather vague and I'm not 100% sure what exactly you're getting at	16:58
Viking-Ice	how the upgrade process is supposed to work	16:58
Viking-Ice	from a - z	16:58
tflink	how is that different from user-facing docs?	16:58
Viking-Ice	has to do with making criteria decision for the process	16:59
tflink	I might just be slow this morning but I'm still not following you	17:00
adamw	i think he means almost a design document	17:00
Viking-Ice	until we have those docs we cant adjust any criteria surrounding it	17:00
adamw	here's how the Final Finished Fedup Process is envisioned to work - no temporary locations, no UI hacks	17:00
adamw	right?	17:00
Viking-Ice	since no one in the QA community will have other then vague idea how it works	17:00
tflink	I'm not against the idea but I'm not sure I'm really qualified to write stuff like that	17:01
tflink	since I only have vague, seldom-mentioned-in-irc plans to go off of	17:01
adamw	yeah it seems like it'd be best coming from wwoods	17:02
adamw	i agree it seems like a good thing to have though	17:02
adamw	do you want an action item to try and get one out of wwoods?	17:02
tflink	I can ask him about it, sure	17:02
adamw	i meant viking-ice, but either way	17:03
adamw	Viking-Ice: do you want to take it? or give it to tflink?	17:03
adamw	#action viking-ice or tflink to try and get a fedup design document out of wwoods	17:05
adamw	alrighty	17:05
adamw	#topic open floor	17:05
adamw	anything else for open floor, anyone?	17:05
tflink	yeah, zsync - http://zsync.moria.org.uk/	17:06
tflink	if I figure out how to get that to work for the smoke images, would people use it?	17:06
tflink	there is no fedora package for it due to bundled zlib	17:06
adamw	is that like deltaiso? or what?	17:06
adamw	i use deltaisos all the time	17:07
tflink	a more generic form of deltaiso	17:07
tflink	and probably a bit more efficient	17:07
kparal	tflink: robatino uses that for all TC/RC images	17:07
tflink	yeah, that's who I heard about it from	17:07
kparal	tflink: http://dl.fedoraproject.org/pub/alt/stage/deltaisos/zsync/	17:07
adamw	well count me in favour of anything deltaiso-y	17:07
kparal	zsync is still not acceptable in Fedora	17:08
kparal	but there's a rsync update that unbundles zlib	17:08
tflink	but the documentation isn't 100% clear and I wanted to get an idea of how much it would be used before putting time into it	17:08
kparal	that's a step in the right direction	17:08
tflink	I thought that didn't work with zsync	17:08
robatino	basically rsync except putting the load on the client instead of the server	17:08
adamw	#info tflink looking at zsync for smoketest images	17:08
tflink	but I haven't checked the review bug for zsync in the last week, I could be mis-remembering	17:08
kparal	tflink: basically first we need to push rsync update to Fedora that unbundles zlib, then we have to change zsync to unbundle it as well, and then we can accept zsync info Fedora	17:09
tflink	ok, it sounds like I was thinking of a different patch	17:10
* kparal links https://bugzilla.redhat.com/show_bug.cgi?id=495310		17:10
tflink	either way, we could still put zsync packages w/ bundled zlib in a side repo for the short term	17:10
adamw	ok, we've over an hour	17:10
kparal	tflink: opensuse has them, their rpm works great	17:10
tflink	kparal: do you have a link off hand?	17:11
adamw	do we need the technical zsync discussion here, or can we wrap up? any other topics?	17:11
tflink	nothing from me	17:11
kparal	tflink: http://software.opensuse.org/package/zsync?search_term=zsync	17:11
tflink	we can take the zsync details elsewhere, I was mostly wondering about interest	17:11
* kparal thinks zsync is great		17:12
adamw	+1	17:12
kparal	tflink: it doesn't work well for Live images, just netinst/DVD	17:12
adamw	ok, setting fuse for Professor X minutes	17:12
tflink	I wonder why that is	17:12
kparal	tflink: you might also want to look at http://kparal.wordpress.com/2009/09/01/zsync-transfer-large-files-efficiently/	17:12
robatino	i thought it was interesting that ubuntu's live downloads work well with rsync/zsync	17:14
kparal	it's 3 years old measurement, things might have changed	17:14
adamw	alright! zsync discussion to -qa please	17:15
adamw	thanks for coming folks	17:15
adamw	#endmeeting	17:15

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!