14:03:37 <pknirsch> #startmeeting Power management for Fedora
14:03:37 <zodbot> Meeting started Wed Jan 13 14:03:37 2010 UTC.  The chair is pknirsch. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:03:37 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
14:03:42 <pknirsch> Hello everyone :)
14:03:54 <pknirsch> i hope you all had a good holiday?
14:04:33 <plautrba> yea :)
14:04:41 <jvcelak> pknirsch: hi! of course. ;)
14:04:46 <pknirsch> good! i hope no work on the computer! :)
14:06:12 <pknirsch> So, first of i totally forgot that we had a public holiday last Wednesday and my wife kinda dragged me away, so i couldn't make it that day :(
14:07:03 <pknirsch> but as we're all kinda busy this week i'll try to make this a short meeting at least.
14:07:17 <pknirsch> so lets quickly get to the topics:
14:07:33 <pknirsch> #topic tuned/ktune
14:07:52 <pknirsch> First of, thanks Jan for doing quite a lot of work in adding logging and fixing stuff with it recently!
14:08:19 <jvcelak> :-)
14:09:43 <pknirsch> Thomas can't be here today, but i'll try to get you two together to work on more future enhancements to it. Thomas has been working on a bigger picture for the ktune/tuned work we can do and it would be great if you two could coordinate your work on it.
14:10:12 <pknirsch> As soon as he has a more complete draft of his ideas he'll send them out to all of you.
14:10:39 <pknirsch> Jan, do you have any specific things you'd like to see/do in tuned/ktune?
14:12:31 <jvcelak> Sounds good... I don't have any ideas now. Only some "cosmetical changes" - for example the interface for plugins should be the same for every module.
14:14:09 <pknirsch> alright. Let's just get you and Thomas together and see what comes out of it for now. :)
14:14:12 <jvcelak> And the speed of the ethernet cards should be more flexible. Now we just switch between 100 and 1000 Mbips
14:14:17 <jvcelak> OK. :-)
14:14:35 <pknirsch> yea, true
14:15:20 <pknirsch> i've added an updated version of the battery powersave ktune.sh script with the newest changes from Matthew so that it's compatible with the latest .32 kernel.
14:15:36 <pknirsch> (did that prior to the holidays i think actually...)
14:16:56 <pknirsch> anything else we need to cover for ktune/tuned? i don't have any new features for F13 at the moment, and the ideas Thomas is working on are more longterm anyway, so more F14++ material.
14:18:09 <pknirsch> alright, seems not, so lets move on to the next topic
14:18:22 <pknirsch> #topic Tasks
14:19:19 <pknirsch> Good news is that we've covered basically all the technical implementation what we wanted to reach by now, so great work everyone!
14:20:11 <pknirsch> I've also verified and got confirmation from Dan that NM now has a 4s grace period when we switch network speeds. That typically will prevent the connection from being dropped when doing this, which is really good.
14:21:20 <pknirsch> One thing that was mentioned on a NM discussion last month was to maybe add a setting capability for it as well so that the timeout could be changed. For old HW that might be necessary (using switches with old Spawning Tree)
14:21:44 <pknirsch> But that has a lower prio atm.
14:22:11 <pknirsch> Regarding programming i've sent my notes about the memory allocator to you today, Jan.
14:22:27 <jvcelak> Perfect, thanks!
14:23:03 <pknirsch> This is also rather longterm and i'm not sure if the notes include some preliminary results of my tests with malloc_trim. If not i expect we'll chat quite a bit in the weeks to come about it or i'll just send them to you.
14:24:35 <pknirsch> Switching to documentation: I still need to finish the 2 documents i've promised to do, hope to get them done next week.
14:25:06 <pknirsch> marcela, have you heard anything from Ruediger or Don lately? If not we might want to ping them again and see where they stand and if they need anything more from us.
14:25:18 <marcela> pknirsch: not for a long time
14:25:38 <marcela> pknirsch: only Don wrote me that he doesn't need more about programming techniques
14:26:22 <pknirsch> marcela: alright, good. do you want to ping Ruediger once more about the PM guide and check if he's fine with the info he has?
14:26:32 <marcela> pknirsch: yes, I'll do it
14:26:40 <pknirsch> marcela: awesome, thanks!
14:28:54 <pknirsch> which brings me to the next topic, QE :)
14:30:27 <pknirsch> have you had any time to work on this with jscotka yet, marcela?
14:30:35 <marcela> I talk with QE about use of scomes in their automatic tests, but I hadn't time to prepare anything yet
14:31:23 <pknirsch> alright, thats fine. do they plan to use the test rpm or something similar? or might we want to suggest to them to use it?
14:32:20 <marcela> test rpm of what?
14:32:29 <plautrba> f12 test rpm
14:33:07 <pknirsch> ye
14:33:18 <plautrba> there is a plan to use scomes in some performance tests
14:33:34 <pknirsch> jlaska really liked it, and maybe it would be something to consider it for our internal tests as well
14:34:00 <plautrba> we should rewrite our tests to rhts/bakker framework
14:34:20 <pknirsch> hm, ye, probably
14:34:24 <plautrba> to be more compatible with qa group
14:34:27 <pknirsch> yea
14:35:14 <plautrba> also, QE will need more configuration on scomes
14:35:28 <plautrba> it's can be done with some shell/python wrapper around
14:36:02 <pknirsch> good idea
14:36:28 <pknirsch> would it be possible to somehow make the rhts scripts and the ones in the test rpm base on the same code?
14:36:37 <pknirsch> that way we could reuse them for fedora and rhts
14:39:36 <jvcelak> its beaker ;) and yes, its possible
14:40:37 <pknirsch> that would really save us a lot of trouble and duplicate work imo
14:41:55 <plautrba> yes
14:43:11 <pknirsch> alright, cool. could you folks work with QE then to get those in rhts/beaker?
14:43:20 <marcela> yes
14:43:26 <pknirsch> perfect :)
14:45:02 <pknirsch> and working with QE on a wider range of tests would be great as well. I've gotten one new rhts review regarding powertop recently by Kexin from QE, but i'd really love to see more, especially in areas where no automation can do things (suspend/hibernate  & resume e.g.)
14:48:06 <pknirsch> i'll send out the tasks list once more today or tomorrow with some updates.
14:48:29 <pknirsch> and last topic (i know this will be hilarious for you folks :P):
14:48:46 <pknirsch> #topic Again, request to change meeting time!
14:48:52 <pknirsch> I know, i know.
14:49:30 <pknirsch> Thing is, the 15:00 slot on Wednesday i have to join another meeting from now on. :/
14:50:04 <plautrba> the only time I
14:50:06 <pknirsch> So any time you folks would prefer to move this now? I'm running out of afternoon timeslots tbh :)
14:50:24 <plautrba> the only time I can't is Wednesday 11:00 - 12:30 CET
14:50:51 <pknirsch> would a morning meeting be better/prefered for you folks?
14:51:04 <pknirsch> as i still have all the time in the world in the morning :)
14:51:08 <pknirsch> (sorta)
14:51:56 <plautrba> what is morning? :)
14:52:06 <plautrba> 10 - 11 CET?
14:52:17 <pknirsch> ye, please not earlier! ;)
14:54:18 <pknirsch> plautrba: would you prefer it then on another day to not take your whole Wednesday morning away?
14:54:31 <jvcelak> pknirsch: I don't mind actually. I don't have a time-schedule for new semester yet... so I will have to adjust.
14:54:55 * plautrba too
14:55:14 <pknirsch> jvcelak, plautrba : we can always adjust if we have to if we move it to the mornings
14:55:32 <pknirsch> as i'm a lot more flexible at that time.
14:55:49 <pknirsch> i'll just send out a quite email to all of you and then we'll find something for now.
14:56:31 <pknirsch> but i gotta run to another meeting already again.
14:56:48 <pknirsch> So happy new year again everyone! and talk to you then next week sometime in the morning. :)
14:57:08 <jvcelak> ok. :-) bye
14:57:13 <marcela> bye
14:57:13 <plautrba> bye
14:57:18 <pknirsch> #endmeeting