From Fedora Project Wiki
Fedora Release Engineering Meeting - 2009-05-04
Preview Release
Known issues:
- PPC had a variety of issues
- oversized
- installed the wrong kernel
- failed to install a bootloader
- assorted anaconda partitioning issues
Discussed possibly using a separate kickstart config for PPC to keep it under size constraints, but it was decided to stay with one config.
Deltarpm for F11
Work needs done to either compose updates in a chroot (which has the F11 deltarpm support) or to backport it to the OS release used to generate updates. Seth Vidal is going to investigate which of these makes more sense. Given the timeframe, this is tight for F11 final. rawhide will continue to have deltas, as that's a separate compose process.
F12 schedule
The schedule proposed by John Poelstra for Fedora 12 in https://fedorahosted.org/rel-eng/ticket/1271 was reviewed. The following changes were approved:
- the alpha milestone was removed entirely
- due to conferences such as the Red Hat Summit, LinuxCon, and Linux Plumber's Conference, each milestone from 'Final freeze: development' (2009-09-15) should be shifted out one week.
This pushes GA from 2009-10-27 to 2009-11-03. The schedule will be presented for FESCo discussion at the 2009-05-08 meeting.
Full IRC log
| ---f13 has changed the topic to: Fedora Release Engineering - roll call | May 04 14:00 | |
| f13 | ping: notting jwb rdieter lmacken spot dgilmore-oz warren | May 04 14:01 |
|---|---|---|
| notting | hey | May 04 14:01 |
| rdieter | here | May 04 14:01 |
| <--jwb has quit (Read error: 104 (Connection reset by peer)) | May 04 14:03 | |
| f13 | wow, nobody else. | May 04 14:04 |
| *wwoods is here, though uncalled | May 04 14:04 | |
| f13 | shit, I knew I forgot somebody | May 04 14:04 |
| f13 | sorry wwoods | May 04 14:04 |
| <--danielsmw has quit (Read error: 110 (Connection timed out)) | May 04 14:04 | |
| -->GeroldKa (n=GeroldKa@fedora/geroldka) has joined #fedora-meeting | May 04 14:04 | |
| <--Sparks_Too has quit ("Leaving") | May 04 14:04 | |
| f13 | I'm still not fully in the swing of things again, last week took a lot out of me | May 04 14:04 |
| ---f13 has changed the topic to: fedora releng - preview release | May 04 14:05 | |
| f13 | so preview went out, not as great as we hoped it would be. | May 04 14:05 |
| f13 | big issues, anaconda partitioning, and ppc | May 04 14:05 |
| f13 | ppc was oversized, even though I thought I fixed that. Also it installs ppc32 kernel on ppc64 systems | May 04 14:06 |
| -->jwb (n=jwboyer@fedora/jwb) has joined #fedora-meeting | May 04 14:06 | |
| f13 | and there are some seemingly ppc specific partitioning issues | May 04 14:06 |
| *jwb pops in late | May 04 14:06 | |
| davej | I managed to get through a network ppc install last week, and it had some kind of bootloader catastrophe | May 04 14:06 |
| f13 | and then anaconda partitioning in general is still a big rough. | May 04 14:06 |
| f13 | jwb: we're talking about preview. | May 04 14:06 |
| f13 | davej: catastrophe in the form of not actually installing a boot loader? | May 04 14:07 |
| davej | never had chance to diagnose it, but I just got a sad mac face icon when I rebooted | May 04 14:07 |
| jwb | whoa... | May 04 14:07 |
| jwb | what kind of ppc machine? | May 04 14:07 |
| f13 | yeah, I put in a patch back on the 17th to fix this | May 04 14:07 |
| davej | olde imac g3 | May 04 14:07 |
| jwb | odd | May 04 14:07 |
| f13 | could have sworn it's in | May 04 14:08 |
| davej | that would have been on teh 26th, so unless it wasn't in rawhide at the time... | May 04 14:08 |
| f13 | so probably not hte same bug. | May 04 14:08 |
| <--izaac has quit () | May 04 14:08 | |
| rdieter | I was asked, didn't know the answer, are there going to be ppc live images ? | May 04 14:08 |
| <--mep has quit (Read error: 104 (Connection reset by peer)) | May 04 14:08 | |
| jwb | probably not | May 04 14:08 |
| davej | I filed an anaconda bug on it. should be blocker material maybe? | May 04 14:08 |
| *davej hunts for it | May 04 14:09 | |
| rdieter | jwb: is stuff broke or just not much interest to work on it? | May 04 14:09 |
| jwb | rdieter, not much demand. livecd-tools was broken for a long time until last week | May 04 14:09 |
| davej | https://bugzilla.redhat.com/show_bug.cgi?id=497745 | May 04 14:09 |
| buggbot | Bug 497745: medium, low, ---, anaconda-maint-list, NEW, fails to install boot loader on ppc | May 04 14:09 |
| rdieter | jwb: ok | May 04 14:09 |
| f13 | davej: yeah, should be a blocker | May 04 14:09 |
| davej | ok | May 04 14:09 |
| f13 | I'll try a ppc install today | May 04 14:10 |
| jwb | f13, there hasn't been an updated anaconda in a few days. maybe worth pinging the anaconda team about first? | May 04 14:11 |
| f13 | I dont' know if there have been that many commits in the last few days | May 04 14:11 |
| f13 | given it was the weekend | May 04 14:11 |
| rdieter | are the anaconda partitioning problems well understood and/or being fixed... or is more testing/feedback needed? any ideas? | May 04 14:11 |
| f13 | rdieter: all of the above | May 04 14:11 |
| jwb | f13, there is at least one upstream fix that should allow upgrades to work | May 04 14:12 |
| rdieter | heh, gotcha. I'll test some too, bugs, etc... | May 04 14:12 |
| f13 | jwb: the last build was from the first | May 04 14:12 |
| f13 | jwb: so there are likely only commits from today | May 04 14:12 |
| wwoods | there have been 35 commits since 11.5.0.48 | May 04 14:13 |
| f13 | anyway we know new anaconda stuff is coming | May 04 14:13 |
| f13 | the real question is will there be enough time to fix enough of the problems. | May 04 14:13 |
| f13 | We're supposed to start RC composes on the 12th, that's a week from tomorrow | May 04 14:14 |
| notting | are composes scriptable? | May 04 14:15 |
| wwoods | answering that requires a review of the blockers. | May 04 14:15 |
| notting | or are we just stuck with testing rawhide for now? | May 04 14:15 |
| f13 | not really | May 04 14:15 |
| f13 | well I mean you can easily run pungi on your own | May 04 14:15 |
| f13 | but we're not in a place where we can do automated composes and sync to somewhere that we can get to it | May 04 14:15 |
| notting | sure, but that's time && space && bandwidth | May 04 14:15 |
| f13 | yep | May 04 14:16 |
| f13 | the bandwidth cost is very little after the first run though | May 04 14:16 |
| f13 | since you use cached data for subsequent runs | May 04 14:16 |
| f13 | Testing rawhide is the best thing though | May 04 14:17 |
| f13 | composes are useful for testing things like split media installs, but those have been working well | May 04 14:17 |
| f13 | and size issues, which is going to be a big concern, given ppc | May 04 14:17 |
| f13 | we're probably going to have to find something big to shave off, a whole group or two | May 04 14:17 |
| jwb | in the past we have used common ks files across all the arches | May 04 14:17 |
| jwb | why can't we have a separate ppc ks? | May 04 14:18 |
| f13 | we could, it just makes things more ... complicated | May 04 14:18 |
| jwb | does it really? | May 04 14:18 |
| -->danielsmw (n=danielsm@user-24-214-179-165.knology.net) has joined #fedora-meeting | May 04 14:18 | |
| f13 | in theory yes. In practice it's not a huge difference. | May 04 14:19 |
| jwb | i'm ignoring theory and politics | May 04 14:19 |
| ---stickster is now known as stickster_afk | May 04 14:19 | |
| f13 | if we got to a point where things were scriptable it would be more of a concern. | May 04 14:19 |
| f13 | but since I'm doing everything by hand anyway, I'd just have to remember to point at a different config file for ppc (and maintain a different config file for ppc) | May 04 14:19 |
| jwb | i can help with the latter part | May 04 14:20 |
| jwb | if we really want to stick with one, that's fine | May 04 14:21 |
| jwb | ppc just hits space issues first | May 04 14:21 |
| jwb | so arguably by fixing it for that, we prevent x86_64 from going over at the same time | May 04 14:21 |
| f13 | yeah, if we have to fix it for ppc, we're not far off from having to fix it for x86_64 | May 04 14:21 |
| f13 | it's just that ppc eats up considerably more space in isofs overhead | May 04 14:21 |
| f13 | er, bootable isofs overhead | May 04 14:22 |
| f13 | so anyway, we know we need to fix that. | May 04 14:27 |
| f13 | and review blocker list | May 04 14:27 |
| f13 | and continue tagging/signing things | May 04 14:27 |
| jwb | continue? | May 04 14:27 |
| <--N3LRX has quit (Remote closed the connection) | May 04 14:27 | |
| jwb | in the signing case i mean | May 04 14:27 |
| f13 | and work out the kinks in delta generation | May 04 14:27 |
| f13 | and figure out how to create deltas for updates | May 04 14:27 |
| f13 | I've been signing everything yes. | May 04 14:27 |
| -->tk009 (n=tk009@adsl-074-166-251-205.sip.bct.bellsouth.net) has joined #fedora-meeting | May 04 14:28 | |
| jwb | ok. so skvidal is going to look at the required packages for deltas | May 04 14:28 |
| jwb | i think he can use app1 for mucking with things first, and then i guess we'll take a leap of faith and throw it on releng2? | May 04 14:28 |
| f13 | I haven't seen the proposal but... sure! | May 04 14:29 |
| jwb | what proposal? | May 04 14:29 |
| jwb | it requires updates to yum, createrepo, and deltarpm | May 04 14:30 |
| f13 | of what you're trying to do | May 04 14:30 |
| jwb | then changes to mash configs in bodhi | May 04 14:30 |
| f13 | honestly I'd rather see bodhi use mock to mash | May 04 14:30 |
| jwb | hell, so would i. you think that is going to happen before f11? | May 04 14:30 |
| -->josedamiangarrid (n=damian@200.49.17.134) has joined #fedora-meeting | May 04 14:31 | |
| f13 | I honestly don't know. I think that depends on how much lmacken we get. | May 04 14:31 |
| f13 | spot: ping; how much lmacken can we get for this? | May 04 14:31 |
| jwb | we lack a sandbox test environment as well | May 04 14:31 |
| f13 | alright so that's a concern. | May 04 14:34 |
| <--Pikachu_2014 has quit ("WeeChat 0.2.6.1") | May 04 14:35 | |
| jwb | yes | May 04 14:35 |
| -->Pikachu_2014 (n=Pikachu_@81-66-11-215.rev.numericable.fr) has joined #fedora-meeting | May 04 14:35 | |
| f13 | so lots of work to do and little time to do it. | May 04 14:35 |
| f13 | anything else anybody would like to add/talk about? | May 04 14:35 |
| jwb | f13, we're going with one key, right? | May 04 14:36 |
| f13 | yes | May 04 14:36 |
| notting | in lieu of poelstra showing up, i'll bang his drum - f12 schedule approval discussion? | May 04 14:36 |
| jwb | cool | May 04 14:36 |
| jwb | notting, where is it? | May 04 14:36 |
| <--mustu has quit (Read error: 60 (Operation timed out)) | May 04 14:37 | |
| f13 | .rel 1271 | May 04 14:37 |
| zodbot | f13: #1271 (Fedora 12 Schedule) - Fedora Release Engineering - Trac - https://fedorahosted.org/rel-eng/ticket/1271 | May 04 14:37 |
| notting | so, that's an alpha freeze 7 weeks from now | May 04 14:38 |
| f13 | hrm. | May 04 14:38 |
| f13 | something wild and off the cuff. | May 04 14:38 |
| rdieter | no rest for the wicked | May 04 14:38 |
| f13 | what if we cut out alpha as a freeze/release point for F12, and go straight through to beta? | May 04 14:39 |
| jwb | please | May 04 14:39 |
| f13 | maybe pause along the way to do a torrent only live image snapshot | May 04 14:39 |
| jwb | 'devel snapshots' | May 04 14:39 |
| f13 | (and if it works well, keep doing that for future releases) | May 04 14:40 |
| notting | f13: hm, that has the virtue of never being tried | May 04 14:40 |
| f13 | notting: I've been trying to reduce the number of freezes/releases for a while now | May 04 14:40 |
| f13 | unsuccessfully | May 04 14:40 |
| jwb | sort of. alpha has been a non-blocking freeze for a couple releases now, right? | May 04 14:40 |
| notting | f13: might have to tweak the beta dates a bit | May 04 14:40 |
| f13 | alpha being a non-blocking freeze has helped somewhat, but still draws a lot of attention away during critical development time | May 04 14:41 |
| jwb | attention from whome? | May 04 14:41 |
| jwb | er, whom | May 04 14:41 |
| notting | installer | May 04 14:41 |
| -->danielbruno (n=danielbr@thor.argo.com.br) has joined #fedora-meeting | May 04 14:41 | |
| -->josedamiangarri1 (n=damian@200.49.17.134) has joined #fedora-meeting | May 04 14:41 | |
| f13 | installer, qa, releng | May 04 14:41 |
| f13 | we end up focusing on week+ old bits for far too long | May 04 14:41 |
| <--josedamiangarrid has quit ("Leaving.") | May 04 14:41 | |
| jwb | ok | May 04 14:41 |
| notting | it may help them to get more uniterrupted devel time without having to break to make a working milestone in the middle | May 04 14:41 |
| <--danielbruno has quit (Client Quit) | May 04 14:42 | |
| jwb | notting, at the slight risk of less testing? | May 04 14:42 |
| f13 | yeah, in combination with more focused day to day testing | May 04 14:42 |
| f13 | instead of "wait for alpha" testing | May 04 14:42 |
| pjones | fewer freezes/releases would certainly be nice. | May 04 14:42 |
| <--josedamiangarri1 has quit (Client Quit) | May 04 14:43 | |
| *skvidal reads scrollback | May 04 14:43 | |
| jwb | skvidal, you popped up in the deltarpm stuff | May 04 14:43 |
| f13 | hehe, now we have a subject that gets people interested | May 04 14:43 |
| skvidal | ? | May 04 14:43 |
| skvidal | I'm working on the pkgs needed now | May 04 14:43 |
| jwb | f13, i'm good with no alpha. need to run it past fesco | May 04 14:44 |
| skvidal | I'll try them out on app1.stg | May 04 14:44 |
| notting | f13: so, if we did that, and just removed that section of the schedule, what's your opinion on the rest of the milestones as posted? | May 04 14:44 |
| jwb | skvidal, f13 would rather see bodhi mash in a chroot if possible | May 04 14:44 |
| skvidal | okay | May 04 14:44 |
| skvidal | so then that's over to lmacken? | May 04 14:44 |
| jwb | skvidal, which... right | May 04 14:44 |
| jwb | skvidal, also would arguably require more time | May 04 14:44 |
| skvidal | so what's the call? | May 04 14:45 |
| jwb | skvidal, i think we're waiting on spot to tell us how much lmacken time we get? | May 04 14:45 |
| skvidal | no deltas on f11 or should I pursue making the pkgs? | May 04 14:45 |
| skvidal | ok, sounds fair to me | May 04 14:45 |
| <--kulll has quit (Read error: 110 (Connection timed out)) | May 04 14:46 | |
| -->danielbruno (n=danielbr@thor.argo.com.br) has joined #fedora-meeting | May 04 14:46 | |
| jwb | f13, fallback is update releng2 or punt for f11? | May 04 14:46 |
| skvidal | OR | May 04 14:46 |
| f13 | notting: I have some concern around final freeze and preview | May 04 14:46 |
| skvidal | we don't have it for F11-GA + zero-day updates but we do have it during the F11 cycle | May 04 14:47 |
| -->tkjacobsen_ (n=tkjacobs@pppoe2-ves.broadcom.dk) has joined #fedora-meeting | May 04 14:47 | |
| jwb | skvidal, also possible. i'm assuming yum would cope just fine | May 04 14:47 |
| <--tkjacobsen has quit (Remote closed the connection) | May 04 14:47 | |
| ---tkjacobsen_ is now known as tkjacobsen | May 04 14:47 | |
| f13 | notting: RH summit is about 2 weeks before the final freeze, and LinuxCon + Linux Plumbers Conference is the week after the freeze | May 04 14:48 |
| skvidal | jwb: yes, if the presto md is not there it just doesn't use it | May 04 14:48 |
| notting | f13: there's always something | May 04 14:48 |
| notting | but yeah | May 04 14:48 |
| notting | f13: not sure how we avoid that without slipping somewhere | May 04 14:49 |
| f13 | notting: ship Nov 3 instead of Oct 27 | May 04 14:49 |
| jwb | 9 month cycle | May 04 14:49 |
| *jwb ducks | May 04 14:49 | |
| notting | jwb: 40 weeks? | May 04 14:49 |
| f13 | basically move the finalfreeze + preview dates out a week | May 04 14:49 |
| notting | then we can talk about f13 being in its second trimester | May 04 14:49 |
| jwb | i like it | May 04 14:49 |
| f13 | *cough* | May 04 14:50 |
| -->cassmodiah (n=cass@fedora/cassmodiah) has joined #fedora-meeting | May 04 14:50 | |
| notting | f13: wouldn't that make the final freeze the week *of* lpc/lc? | May 04 14:50 |
| -->kulll (n=kulll@203.82.79.102) has joined #fedora-meeting | May 04 14:51 | |
| f13 | notting: yes, the freeze would happen while at the event, which is easy to do | May 04 14:51 |
| notting | that's the pre-preview final freeze, right? | May 04 14:52 |
| f13 | LinuxCon is over on the 23rd, the day after the freeze | May 04 14:52 |
| f13 | yeah | May 04 14:52 |
| ---knurd is now known as knurd_afk | May 04 14:52 | |
| notting | wwoods: as an interested observer, you have not commented | May 04 14:53 |
| f13 | We'd have the entire week after LinuxCon/LinuxPlumbers to compose/stage/sync | May 04 14:53 |
| f13 | well "entire week" == Mon/Tues/Wed/Thur | May 04 14:53 |
| *wwoods catching up on scrollback | May 04 14:53 | |
| wwoods | so are we discussing Alpha or Presto at the moment? | May 04 14:54 |
| -->kital (n=Joerg_Si@fedora/kital) has joined #fedora-meeting | May 04 14:54 | |
| notting | wwoods: f12 schedule | May 04 14:54 |
| jwb | wwoods, back to f12 schedule. sorry | May 04 14:54 |
| wwoods | gotcha | May 04 14:54 |
| notting | and the possibility of killing alpha (before it comes back and slices us all up) | May 04 14:54 |
| wwoods | I'm fine with killing Alpha | May 04 14:55 |
| f13 | we pretty much did our own alpha testing anyway | May 04 14:55 |
| notting | ok, so proposal: poelstra's schedule @ http://poelstra.fedorapeople.org/schedules/f-12/f-12-summary-tasks.html, with the following changes: | May 04 14:55 |
| wwoods | I'd feel less evil if we called the first post-feature-freeze milestone "Alpha" and the subsequent milestone "Beta" | May 04 14:55 |
| notting | 1) alpha section removed | May 04 14:55 |
| wwoods | having an Alpha where it's not required to even have any of the planned features *present* is ludicrous | May 04 14:56 |
| notting | 2) all dates from "final freeze: development" onwards shifted one week later | May 04 14:56 |
| f13 | notting: that seems pretty reasonable to me, particularly for the short release. | May 04 14:57 |
| <--danielbruno has quit ("Saindo") | May 04 14:57 | |
| f13 | good enough to pitch at the public for more date tweaking if there are events we don't know about | May 04 14:57 |
| notting | ok. i'll take it to fesco this week | May 04 14:57 |
| notting | wwoods: not sure beta -> alpha, preview -> beta makes sense, given the freeze state at preview | May 04 14:58 |
| f13 | ok, we're near the hour mark. Any other topics for discussion? | May 04 14:59 |
| -->danielbruno (n=danielbr@thor.argo.com.br) has joined #fedora-meeting | May 04 14:59 | |
| wwoods | notting: true. No worries - removing the Alpha release also fixes the "Alpha is a big fat lie" problem | May 04 15:00 |
| ---stickster_afk is now known as stickster | May 04 15:01 | |
| f13 | Looks like not. Time to call it a meeting. Thanks all. | May 04 15:02 |
| <--mitr (n=nnnmitr@popelka.ms.mff.cuni.cz) has left #fedora-meeting ("Leaving") | May 04 15:02 | |
| ---f13 has changed the topic to: Channel is used by various Fedora groups and committees for their regular meetings | Note that meetings often get logged | For questions about using Fedora please ask in #fedora | See http://fedoraproject.org/wiki/Meeting_channel for meeting schedule | May 04 15:02 | |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!