ctylerLet's start with a roll call and brief intro from the candidates12:00
* jds2001 hre12:00
* sharkcz here12:00
*** mdomsch has joined #fedora-townhall12:01
*** ChanServ sets mode: +v mdomsch12:01
* jwb is here12:01
* j-rod here12:01
ctylerIs Dominik around?12:01
* jds2001 Jon Stanley, current member running for re-election. I maintain about a dozen packages in Fedora and EPEL.12:01
*** ChanServ sets mode: -v mdomsch12:01
jds2001professional linux sysadmin for about 10 years give or take, going back to RH5.212:02
jwbi'm Josh Boyer.  I've been using Red Hat/Fedora since RH 7.1.  I've been a contributor (packagers) since around FC3, and i've argued on mailing lists for longer than that.  i'm also on FESCo, and usually part of rel-eng12:02
ctylerj-rod, sharkcz -- 1-2 line self-introductions?12:03
* sharkcz since March employed by RH, long time RHL, Fedora and CentOS user, maintainer of 20+ packaged, founder of Server SIG12:03
*** Rathann has joined #fedora-townhall12:03
*** ChanServ sets mode: +v Rathann12:03
ctylerRathann: 1-2 line self-introduction?12:03
j-rodJarod Wilson, RH employee for 2.5 years, RHEL kernel group, current fesco member12:03
j-rodmaintainer of some fedora packages, regular contributor to the fedora kernel12:04
jds2001I also lead the bug triage efforts, do QA, and am leading the QA portion of the feature process for F11.12:04
*** kiilerix has joined #fedora-townhall12:04
RathannDominik Mierzejewski, sysadmin for 5+ years (at a university supercomputing centre), Packaging Committee member12:04
ctylerThanks for the intros. I'm watching for questions in #fedora-townhall-public and will relay them to this channel.12:04
Rathannmaintainer of several dozen of packages, both in Fedora and RPMFusion12:05
ctylerFirst question is from notting:12:05
ctylerAre you interested in volunteering to chair fesco?12:05
Rathannnot at this time12:05
sharkczno, it is my first try for FESCo12:06
jds2001bpepple needs a well-deserved break :)12:06
ctylerAny expansion on those answers?12:06
jwbwhat jds2001 said12:06
jwbit's a thankless job, and we've had great Chairs in the past12:06
j-rodI'd really like to do more than I do on the Fedora side, and if time permitted it, I'd be happy to try to give bpepple that break too.12:07
jwbcontinuing the trend would be a fun challenge :)12:07
ctylerHere's Q2 from mdomsch: How do you feel about FESCo's role in overseeing the SIGs; what has worked, and what hasn't, and what will you do to improve that?12:07
RathannI'm not comfortable with being the chair on my first term12:07
j-rodknowing my current workload though, I wouldn't pick myself as the first choice for it though12:07
jwbi think FESCo needs to get a bit more interactive with the SIGs12:07
jds2001and we thank bpepple for his service in that role over the years.12:07
jwbwe've been mostly hands off so far, and it's worked for some and not so much for others12:08
j-rodwhat jwb said -- as evident from our discussion on secondary arches yesterday12:08
jwbthat's a good example12:08
jds2001I think FESCo needs to be proactive, currently, it's more of a reactive body (we see something broken and go help fix it)12:08
jwbSpins is another where i think we should interact more than we do12:08
Rathannwhat jwb said, and also I think we should get regular reports from SIG reps and/or organize meetings with them12:08
jwband i'll take the blame for not doing so already on Spins12:09
* jds2001 is working on that now.12:09
jds2001spins that is.12:09
sharkczI agree with jwb, get reports from SIGs etc.12:09
j-rodI'll take some on secondary arches -- I'm one of the members w/the most invested in them being successful...12:09
jwbby interactive, i don't mean 'authoritative'12:10
jwbi just mean interactive.  just wanted to clarify12:10
RathannI'm interested in working with the server SIG and the desktop maintainers to smooth out any rough edges12:10
sharkczI participate on s390x, so secondary arches are my second focus afther the server12:11
j-rodalong the same lines as the secondary arches, it'd be good to get at least monthly reports from all SIGs12:11
ctylerFollowup from mdomsch: please elaborate on secondary arch discussion and challenges faced by fesco12:11
jwbnot so much challenges12:11
jwbthe only thing we're lacking there is nobody has any clue as to what the status of them are12:11
j-rodand perhaps have specific fesco members responsible for interaction with a given subset of SIGs12:11
jds2001ia64 has pretty much died off, for various reasons. (I'm not even sure why).12:12
jwbso the current FESCo (me) is going to poke them for reprots12:12
jwber, reports12:12
Rathannj-rod: preferably overlapping subsets12:12
j-rodthat'd work12:12
jds2001so that's the current issue.  However, we had no idea until it was mentioned yesterday.12:12
jwbwell, for the most part12:12
jwbi had status on a few12:13
jds2001We should be more proactive on these issues, so that we can see them coming before there's a shipwreck.12:13
jds2001jwb: sorry, just speaking for myself.12:13
j-rodmore two-way communication could help motivate the secondary arch groups to do more12:13
j-rodand we might be able to help with current roadblocks, etc12:13
jwbaside from the SIGs themselves, we need to interlock with the koji devs12:13
jwbsince koji needs support for this12:13
jwbbut that is being actively worked12:13
ctylerNext q...12:14
ctylerQ3, also from mdomsch: is the size and makeup of FESCo proper?  How do you ensure the proper skills are present?12:14
RathannI used to be involved with sparc, maybe I'll revisit it12:14
jwbi think the size is proper12:14
jwbthe old FESCo with 13 members was too large12:14
jwbas for skill sets, well we're entirely elected12:15
jwbso we sort of have to be composed of people who want to be here12:15
Rathanntoo many cooks spoil the broth :)12:15
jwbif we are missing skills, we pull in subject experts when needed12:15
jds2001and the electorate really decides what the proper skills are and that they're present.  We're all packagers, we all do other stuff within Fedora.12:15
jwbexample, ivazquez for the python 2.6 stuff12:15
sharkczsize is OK and skill are distributed between the nominees12:15
jds2001feature owners for their features.12:15
jwbjds2001, yep12:16
j-rodI'm fine w/the current size, wasn't around when it was any other size12:16
jwbi was12:16
jwbit was a pain12:16
Rathannwe can only encourage people who we think are needed to accept nominations12:16
Rathannafter that it's up to the voters12:16
j-rodand I think we have a reasonable balance of skills12:16
* jds2001 not going to be hacking on kernels, for instance :)12:17
jwbthat's why we have j-rod ;)12:17
j-rodjwb: hey, you do your fair share of that too... :)12:17
jwbsome :)12:17
RathannI think we have the right mix, too12:18
ctylerContinuing with Q4, from again from mdomsch: what would you each like to accomplish as a FESCo member in the next year?12:18
jwbhe's reusing questions from the Board meetings12:18
jds2001comes back to fesco isn't a posistion for advancing personal platforms :)12:18
j-roddoesn't anyone else have any questions?!? ;)12:19
jwbjds2001, right12:19
jds2001however, i'd like to see less schedule slip, and that's entirely in fesco's control.12:19
jwbsort of12:19
jds2001I hope that the changes we made to the schedule this time around help effect that.12:19
j-rodI'd really like to do more on the secondary arch front12:19
jds2001F10 being an obvious exception.12:19
Rathannduring my time as fedora developer I noticed that some changes catch people by surprise, which means they're not communicated as well as they could be12:20
RathannI'd like to change that12:20
jwbj-rod, i would to, but you don't need to be on FESCo for that12:20
jds2001we == fesco, obviously proposed by releng.12:20
jwbso if i were going to change FESCo in the next year, it would be to make it a bit more interactive12:20
j-rodbut its related to the "fesco needs to interact w/sigs more" thing12:20
jwbacross the board12:20
sharkczI would like to see the server complement the desktop :-)12:20
j-rodI think jwb and I are reading the same play book...12:21
jwbkernel people have to stick together ;)12:21
*** ctyler changes topic to "FESCo Townhall meeting @ 17:00 UTC - Questions can be asked in #fedora-townhall-public"12:22
j-rod(on a semi-related note, RH is hiring for a fedora secondary arch QA person in China now...)12:22
Rathannalso I want to try and reach out to the community to get more testing of new features12:22
Rathannyou can never have enough testing :)12:22
jds2001Rathann: you're stealing from my playbook now :)12:23
Rathannwell, those weekly package review reports gave me an idea12:23
Rathannwe could do something similar for testers12:23
sharkczjds2001: I think there were some ideas presented on the QA meeting this week12:24
jds2001yeah, the age-old problem is how to get the data.12:24
*** Roobarb has left #fedora-townhall12:24
j-rodas a sub-group of that, I'd like to try to get a community fedora kernel qa group going12:24
jds2001the QA: namespace would help12:24
jds2001sharkcz: you mean the marathon QA/FESCo meeting this week :)12:24
jds2001that was like 3 hours of continous fedora meetings :)12:25
j-rodwith a focus being on testing kernels in updates-testing and beating on kernels just before release12:25
jds2001j-rod: we'd love to, test plans accepted :)12:25
jwb'run LTP'12:25
sharkczdivide to smaller areas and conquer :-)12:25
ctylerQ5, from (guess!) mdomsch: Describe FESCo's role in the Feature Process, how that's worked, and how you will improve it.12:26
jwbit's worked because poelcat makes it work12:26
jwbhe deserves all the credit for any success it's had12:27
j-rodI think we've already said it was a bit broken, and have implemented some new stuff for the F11 devel cycle12:27
jwbbut yes, there are some changes coming for F1112:27
jds2001in general, the feature process is a great tool.  it works for what it's designed to do.  Some shortcomings have been found, and they are iteratively workeed out.12:27
sharkczthe last fesco meeting presented some improvements12:27
j-rod"it" being fesco's role, not polecat's12:27
jds2001the QA portion that I'm working on, for instance.12:27
jwband i think FESCo needs to be a bit more discerning on what fits and what is acceptable12:27
ctylerThat raises the question of what you each see as fits/acceptable (?)12:28
Rathannmostly it seems to be working fine, but it needs more impact analysis and communicating the upcoming changes to the affected audience12:28
jds2001jwb: I think we did a better job of that last time than in the past.12:28
jwbjds2001, for the most part, yes12:28
RathannI see missing presto as a failure here12:29
jwbctyler, complete Feature pages, quantifiable test plans, backup solutions, etc12:29
jwbRathann, presto is definitely something we might look at12:29
jwbit's useful, but it keeps losing traction12:30
Rathannthere have already been what, two attempts to get it included?12:30
jds2001that being said, it's still a relatively new process.  It's still evolving. I think test plan/spec wrnagling would have saved us at least one last minute change for F10.12:30
Rathannyes indeed12:30
RathannI do like the contingency plans12:30
jds2001so that's the quantifiable thing that I'm doing for F11.12:31
Rathannthey're definitely essential12:31
ctylerQ6 from notting: What is your feeling on the length of the review queue? Good? Bad? Indifferent? How would you tackle it?12:31
jwbpackage review, or Feature?12:31
ctylerlet's assume package12:32
j-rodassuming package, BAD12:32
j-rodbut not really sure how to tackle it12:32
jwbthis comes up with every new FESCo12:32
jwbthere are no magic bullets12:32
jwbthe typical answers are to encourage more reviewers, solicit more sponsors, etc12:33
j-rodI've done some fairly thorough reviews, then not got any reply from the submitter for ages.12:33
Rathannthere's been some talk about streamlining the review guidelines12:33
jwbwe have some uber, er proven, reviewers out there like tibbs12:33
Rathannautomating what's possible12:33
j-rodand now, like many of our more competent reviewers (not saying I am one), I just don't have that much time for review12:34
jwbRathann, that might help12:34
jds2001that would likely help, automation would help.12:34
jwbit also comes up every time12:34
Rathannas it is, the package review checklist is scary and long12:34
Rathannparts of it duplicate the packaging guidelines12:34
jwbit only grows longer as new guidelines are added12:34
sharkczsome per week statistics (new, done) can help too12:34
jwbsharkcz, we have those12:34
jwbthey didn't make a huge difference12:35
j-rodyeah. package review is mostly thankless, unless its a package you personally want in12:35
jwbour answers might sound doomy-and-gloomy12:36
Rathannexample: I wanted to package scilab but found out someone else was willing to do that so I started reviewing the necessary dependencies to help them12:36
jwbi think they just reflect reality12:36
jwbRathann, yeah.  personal interest in a package has a huge impact on reviews12:36
RathannI think the completed reviews stats posted weekly do make the job more attractive12:37
Rathannpeople like to be recognized12:37
Rathanneven if they don't admit it12:37
* jds2001 might try to whip something up for that.12:37
j-rodcash bounties for x number of reviews completed might help... :)12:37
ctylerQ7 from mdomsch: F10 shipped with virtually no blocker bugs.  How can we improve QA to make this a trend?12:38
jwbisn't the definition of 'blocker' "you cannot ship with this bug"?12:38
jwbor did he generally mean 'no major bugs'?12:38
jds2001do exactly what we did for F10.  Have manic focus on blockers throughout the cycle.12:39
j-rodyeah, I thought we didn't ship if a bug was a blocker... some bugs just get demoted...12:39
sharkczmaybe some blockers were not discovered :-)12:39
jds2001jwb: there's always stuff left on FXBlocker :P12:39
jwbthen it wasn't really a blocker ;)12:39
jwbanyway, semantics aside12:39
jds2001yeah, both of those things too12:39
Rathannlike I said earlier: one thing is to get more testing12:39
j-rodgeneral answer is more QA though, indeed12:39
jwbit comes from an increase focus on QA.  the updated Feature process is going to hopefully help with that12:39
Rathannthe other: get upstream developers involved12:39
Rathann(if they aren't already)12:40
sharkcznot only more testing, but precise testing12:40
j-rodwhich (tying back into my desire for a fedora community kernel qa crew) starts with the kernel12:40
jwbgetting people to do test installs in kvm guests, etc can help too12:40
Rathannjwb: not everyone can run kvm though, so I'm planning to (help) package virtualbox in rpmfusion12:40
j-rodI don't actually know for sure, but I imagine the bulk of critical release blockers are typically in the kernel12:40
jwbRathann, sure.  virtualization technology should hopefully help us in general12:41
jwbj-rod, or anaconda12:41
jds2001likely, and a lot of it is probably testing with specific hardware12:41
j-rodjwb: that was my #212:41
Rathannand mine :)12:41
sharkczj-rod: and in anaconda & co12:41
Rathannsome bugs are only fixable if you have access to the real hardware12:41
ctyler(mdomsch stirs the pot a bit: feature testing is good, what about regression testing?)12:41
jds2001a lot of what we get is hardware X doesn't work, and we can't possibly acquire all this hardwaqre12:41
Rathannso maybe try and arrange for specific hardware donations if some developer needs it?12:42
jwbregression testing how?12:42
sharkczbut we can make a database where the hw is available12:42
Rathanna wiki page with donation requests/wishlist might be worth considering12:42
j-rodI've got a few things on the regression testing front for the kernel...12:42
jwbyou could have someone looking for the magic REGRESSION keyword in bugzilla, tracking that, ettc12:43
j-rodthere's a whole battery of stuff we run internally on rhel kernels12:43
jwbi don't see how that is much different from overall QA though12:43
sharkczor just say - my machine with this hw can be available for testing12:43
jds2001j-rod: is that stuff public?12:43
jds2001we'd love to do teh same on fedora kernels :)12:43
jwbj-rod, some of the problem with that is the duration of time it takes to run it12:43
j-rodI'd rather like to get as much of it running w/a stable of fedora machines12:44
j-rodjds2001: at least a good portion of it could be12:44
jwbif you're doing a new kernel every day and the tests take more than a day to complete, you're not really going to catch up12:44
jwbbut it's probably something that should be done after Beta Freeze or at some other semi-quiescent point12:44
RathannI know of at least one project whose developers have a donations wishlist page and the donations do happen12:44
j-rodnightly quicker sanity checks, weekly longer tests, maybe12:45
ctylerQ8 based on an observation by bpepple: the questions in this townhall have come from people who are/have been FESCo members. How can FESCo more completely engage with the larger Fedora community?12:46
jwbthe SIG interaction is part of that12:46
jds2001and also maybe something on FWN12:46
*** jaroslav has joined #fedora-townhall12:46
RathannFESCo meeting minutes are useful to read too12:47
jds2001right, they go to f-devel, which a lot of folks might not subscribe to12:47
jds2001in the larger fedora community12:47
sharkczand they should into f-d-announce12:47
jwbit depends12:48
jds2001sharkcz: that's against the charter of f-d-a12:48
jwbwhy does the art team care what FESCo does?12:48
jwb(that was an example)12:48
jds2001they don't, but they might like to be informed from time to time.12:48
jwbFESCo is packager and developer oriented12:48
sharkczjds2001: ?12:49
jwbso engaging with the broader community isn't bad, but the topics we discuss likely might not be of too much interest12:49
* jds2001 sees no harm in having a fesco beat on FWN12:49
Rathannit might be worth reminding people that FESCo is there from time to time12:50
* jds2001 might even write it, depending on deadlines.12:50
j-rodI like the FWN idea12:50
Rathannwell, semi-regularly, provided anything interesting did happen12:50
ctylerQ9 from quaid: earlier this year FESCo and the Board had discussions to (re)set FESCo's charter/area of influence; do candidates feel FESCo is covering what it is supposed to be covering?  Should it have oversight over more of Fedora production or less?12:52
*** ubertibbs has joined #fedora-townhall12:52
jwbi think the reset has helped a bit12:52
jwbthere are a few things we should do better at that we've discussed already (Features, communication, interactivity)12:53
* ctyler notes that the observers in #fedora-townhall-meeting aren't noticing much difference between the candidate's answers on these questions12:53
jwbbut i don't think we need entirely more oversight12:53
RathannI believe the technical matters are well covered12:53
* ctyler meant #fedora-townhall-public12:53
jwbour SIGs and developers do a really good job at getting releases out the door12:53
Rathannyes, our job should be just letting people work efficiently12:53
jwbi only have one 'release time' thing i'd like to change12:54
j-rodhm... just to give observers something different... I think we should have more control over everything. All your base are belong to us.12:54
j-rod(ok, not really)12:54
jwband that is to get the glibc people to do their upstream release _before_ final freeze, not after12:54
Rathannwell, I for one am not a RH employee12:54
RathannI'm told that matters to some people12:54
jwbit shouldn't12:54
* jds2001 is not either, but don't see how it matters.12:54
RathannI don't think it should either12:55
j-rodthere's a faction that thinks RH employees have a secret hidden agenda12:55
jds2001world domination? :)12:55
j-rodare you thinking what I'm thinking, pinky?12:55
ctyler(quaid wonders if the answers to the last question boil down to "not too hot, not too cold, but just right?")12:56
jds2001i think so.12:56
j-rodI'm not aware of anything we're not covering that we should be12:56
j-rodmain thing is improving the interactivity12:56
Rathannand communication12:57
sharkczimprove communication12:57
jwbwhich doesn't really have anything to do with oversight of the release :)12:57
jwbctyler, in response to the 'similar answers' comment12:57
jwbthese questions aren't controversial enough :)12:58
jds2001are we covering things that we shouldn't be? I don't think so.12:58
jwbgive us something polarizing!12:58
j-rodlets talk about flash and libcurl or something!12:58
Rathannwe shall not bow to corporate overlords! freedom for all!12:59
sharkczj-rod: you mean e.g. NM?12:59
j-rodsharkcz: new mexico, or network manager? (and what about it?) :)12:59
ctylerQ10 -- final question: provenpackager group is significantly larger than was originally proposed.  Is this a problem?13:00
* j-rod was referring to the whole dust up over flash needing some old curl so13:00
jwbi think it needs a revisit13:00
jwbthere has been enough outcry about it and there are valid concerns13:01
* jds2001 too, but how to do it?13:01
RathannI think we don't really know if it is a problem or not, but I believe it might be: a lot of unproven people have been given commit access to most packages13:01
ctyler(full disclosure: jwb had input into question in -public)13:01
jwbin practice, it hasn't bitten us at all.  but perception of it, and the usefulness of it, has been called into question13:01
j-rodnot a particularly big fan of how the group was initially populated either13:02
jwbjds2001, there are a couple ways.  we could have a voluntary culling to start with13:02
j-rodI mean, sure, it needed a seed13:02
*** glezos has left #fedora-townhall13:02
Rathannjwb: it hasn't bitten us yet13:02
RathannI hope it never will but you never know13:02
jwbif that doesn't prove useful, we can reseed to what was originally proposed which was sponsors13:02
jwbRathann, true13:02
jwbthat's my opinion anyway :)13:03
jds2001we do need more communication about "why would I approve someone here?"13:03
Rathannmaybe I'm a bit paranoid13:03
* jds2001 has a hard stop in just a few13:05
ctylerWe've been talking for about an hour, and I thank you all. Let's finish up with a 1-2 line comment about something unique about your POV, or what you will bring to the position.13:05
jds2001gotta go to physical therapy for my bad shoulder so I can beat jwb in paintball at fudcon :)13:05
jwbi will bring more email13:06
*** EvilBob has joined #fedora-townhall13:06
jwblots more email13:06
RathannI like both desktop and server parts of Fedora and I want to make them grow without disturbing each other.13:06
j-rodI'm a kernel guy in the RHEL group, trying to help make Fedora better13:07
jwbmore seriously, i think that's a hard thing to answer.  i'll do my best to make sure FESCo is an enabler and not a hurdle is about all i can say13:07
sharkczI want to add more focus for the non-desktop usage13:07
RathannAlso my pet-peeve are the multimedia, which is a difficult topic. I want to improve things here as much as I can.13:08
* jds2001 is the token QA guy :)13:08
ctylerThank you all for joining the townhall!13:08
RathannThanks for all the questions13:08
jwbthanks for moderating ctyler13:08
j-rodand thank you to ctyler for moderating13:08
jwband thanks for the questions13:09
jds2001thx for moderating ctyler :)13:09
j-rodthanks to everyone for listening to us babble. :)13:09
*** ChanServ sets mode: +v mdomsch13:09
Rathannand thanks for moderating13:09
*** notting has left #fedora-townhall13:09
* ctyler reminds everyone to vote13:09
mdomschTo everyone - I'm soliciting feedback on this Town Hall process - what worked, what didn't work - was it helpful to you in determining how you would vote?13:09
* sharkcz thanks too :-)13:09
* j-rod can feel the love13:09
mdomschfeedback to fedora-advisory-board@redhat.com (public), or to me in private13:09
Rathannwill do13:09

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!