20:10:15 <cwickert> #startmeeting Spins SIG meeting 2011-03-28
20:10:15 <zodbot> Meeting started Mon Mar 28 20:10:15 2011 UTC.  The chair is cwickert. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:10:15 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
20:10:23 <cwickert> #chairs brunowolff nirik
20:10:30 <cwickert> ok, let's make this a quick one
20:10:46 <cwickert> we have nothing really on the agenda for today
20:10:57 <cwickert> or no news on the tings we have on the agenda
20:11:07 <cwickert> however there is one thing I'd like to discuss
20:11:18 <cwickert> it is the usage of %include
20:11:22 <cwickert> in kickstart files
20:11:38 <cwickert> we have a quite a lot of broken spins atm
20:11:59 <cwickert> basically everything that is based on the desktop and has not been updated for GNOME 3 is broken
20:12:00 <cwickert> more or less
20:12:01 <brunowolff> Also the hwclock issue that was just reassigned to us.
20:12:18 <cwickert> ok, lets come to that later, ok?
20:12:21 <nirik> also, did we every decide the prelink thing? ;)
20:12:44 <cwickert> we agree that we want it but not on how to implement it
20:12:53 <cwickert> anyway includes first, ok?
20:13:21 <cwickert> #topic Usage of %include in kickstarts
20:14:04 <cwickert> we have quite a lot of snipetts in the snippets folder
20:14:19 <cwickert> for example gdm config, gnome-screensaver, packagekit and all that
20:14:35 <brunowolff> I am hoping we can drop some of the unmaintained spins before going to the trouble to fix them.
20:15:00 <cwickert> I think we should use the snippets as includes instead of duplicationg the code
20:15:23 <brunowolff> I think that is a good idea.
20:15:27 <cwickert> an %include statement is treated as if the text appeared at that very same place
20:15:35 <cwickert> so it is easy to do it
20:15:49 <cwickert> if everybody is fine with that I'll go ahead and make the changes
20:15:59 <cwickert> ok for everybody?
20:16:14 <cwickert> +1 (obviously)
20:16:28 <nirik> sure. +1
20:16:36 <cwickert> ok, passed then :)
20:16:36 <brunowolff> Please also have design suite base off of livecd-desktop (or live-desktop).
20:16:48 <cwickert> brunowolff: yes, I need to poke them
20:17:28 <cwickert> #agreed: we will use snippets as %includes rather than duplicating the code in every kickstart again
20:17:38 <cwickert> #action cwickert to announce this change
20:17:42 <cwickert> #undo
20:17:42 <zodbot> Removing item from minutes: <MeetBot.items.Action object at 0x2ed127d0>
20:17:53 <cwickert> #action cwickert to announce the %include change
20:18:04 <cwickert> #topic hwclock
20:18:20 <cwickert> brunowolff: what is the problem here, I honestly missed it
20:18:52 <brunowolff> There was a very old ticket complaining about the hwclock function messing with the hardware clock
20:18:57 <brunowolff> on live images.
20:19:22 <brunowolff> It just got spin-kickstarts people cc'd on it after being bounced around.
20:20:08 <brunowolff> I think the live images should not save back to the hardware clock when shutting down.
20:20:35 <cwickert> what script is responsible for that?
20:20:38 <brunowolff> Some suggestions were being bandied about in the ticket on how best to do this.
20:20:49 <cwickert> you have the link at hand?
20:21:28 <brunowolff> https://bugzilla.redhat.com/show_bug.cgi?id=297421
20:21:34 * cwickert looks
20:22:16 <cwickert> dud, this is old! :O
20:22:21 <cwickert> s/dud/dude
20:23:10 <brunowolff> The issue is still current, and I think I saw the time shift happen to one of my machines about a month ago.
20:23:47 <brunowolff> It sounded like it might be assuming the hardware clock is locatime when starting up and utc when shutting
20:23:54 <brunowolff> down or vise versa.
20:24:40 <brunowolff> I am not sure what really makes sense for timezone and hardware clock defaults. But I feel pretty
20:25:00 <cwickert> I am not sure if I fully understand the problem
20:25:07 <brunowolff> strongly that we don't want to save a time back to the hardware clock when shutting down.
20:25:10 <cwickert> well, I do understand the problem itself
20:25:24 <cwickert> but not the proposed solutions and their downsides
20:25:53 <brunowolff> I wasn't sure either and didn't have a specific proposal at this point.
20:27:13 <cwickert> we only need to stop the livemedia from syncing to hwclock on shutdown, right?
20:27:34 <cwickert> so how about we make the change that Lennart proposed in the livecd script?
20:27:43 <cwickert> ln -sf /dev/null /etc/systemd/system/hwclock-save.service that was
20:28:21 <brunowolff> Functionally I think that is what we want to achieve.
20:28:31 <cwickert> s/livecd script/etc/init.d//etc/rc.d/init.d/livesys
20:28:48 <cwickert> then it should be in live mode but not on the installed system
20:29:32 <cwickert> ok, who can investigate this issue?
20:29:44 <cwickert> #link https://bugzilla.redhat.com/show_bug.cgi?id=297421
20:29:49 <brunowolff> Assuming that works as advertised, I think that is a reasonable fix for preventing messing with the hardware clock.
20:30:07 <brunowolff> And I think that is the more important issue.
20:30:21 <cwickert> do you have the time to work on that next week?
20:30:36 * cwickert is pretty busy atm
20:30:42 <brunowolff> I think I could try it and do a test.
20:30:59 <cwickert> ok, please do
20:31:06 <brunowolff> If things don't work as expected, I may not have enough time to fully figure things out.
20:31:46 <cwickert> #action brunowolff will try if the fix proposed by Lennart (ln -sf /dev/null /etc/systemd/system/hwclock-save.service) can go into /etc/rc.d/init.d/livesys
20:31:49 <brunowolff> #action brunowolff will implement Lennart's proposed solution for bug 29421
20:31:55 <cwickert> #undo
20:31:55 <zodbot> Removing item from minutes: <MeetBot.items.Action object at 0x2ed27910>
20:32:00 <cwickert> I think one is enough ;)
20:32:26 <cwickert> #agreed if lennart's fix works, well use this one
20:32:38 <cwickert> anything more on that?
20:32:49 <brunowolff> I think that covers it.
20:33:04 <cwickert> #topic prelink
20:33:14 <cwickert> nirik: anything from your side?
20:33:46 <nirik> well, I was thinking we should just add it in at compose.
20:33:48 <nirik> time.
20:34:20 <cwickert> in %post that is?
20:34:29 <brunowolff> By putting in livebase?
20:36:19 <cwickert> I think so. what about mini?
20:37:08 <nirik> yeah, in post...
20:37:11 <nirik> in livebase...
20:37:20 <brunowolff> I don't know what makes sense for the tiny spins.
20:37:32 <cwickert> +1 brunowolff
20:37:41 <brunowolff> The way things are going they might end up being dropped.
20:37:58 <cwickert> I think livebase should be fine. for mini they might not want to have prelink installed
20:38:07 <nirik> I guess in mini too?
20:38:26 <cwickert> well, if mini is really mini... ;)
20:38:47 <cwickert> optinions?
20:38:57 <cwickert> erm, opinions
20:38:58 <nirik> sure. whatever they want.
20:39:00 <nirik> we could ask peter
20:39:03 <brunowolff> I think just livebase for this go around.
20:39:21 <cwickert> ok, we all agree on live-base and for mini we ask peter
20:39:24 <brunowolff> If Peter wants to weigh in, than go with what he suggests.
20:39:54 <cwickert> #agreed to add prelink to %post of fedora-live-base.ks
20:40:01 <brunowolff> I don't think Dave Huff has worked on AOS for a long time, so probably only the Meego spin is relevant.
20:40:37 <cwickert> #action cwickert to ask probinson if he wants prelink in fedora-live-mini.ks
20:41:15 <cwickert> nirik: can you make test it and the change?
20:42:35 <nirik> I can try...
20:44:25 <cwickert> ok
20:44:41 <cwickert> anything more? otherwise I'll close the meeting in 3 minutes from now
20:45:13 * nirik has nothing.
20:45:14 <cwickert> #action nirik to add prelink to %post of fedora-live-base.ks
20:45:36 <brunowolff> Nothing here eitehr.
20:45:56 <cwickert> #endmeeting