15:01:30 <pknirsch> #startmeeting Power management for Fedora 15:01:30 <zodbot> Meeting started Thu Oct 1 15:01:30 2009 UTC. The chair is pknirsch. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:30 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:01:43 <pknirsch> Hello and welcome everyone :) 15:02:12 <pknirsch> first of all thanks to skalnik1 for running the meetings the last 3 weeks while i was away 15:02:31 <pknirsch> it was really greatly appreciated :) 15:03:39 <pknirsch> as usual then lets start with a quick recap of last week: 15:03:42 <skalnik1> thanks 15:03:42 <pknirsch> #topic Last week recap 15:04:00 <pknirsch> skalnik1, could you tell us quickly what you discussed last week please? 15:04:32 <skalnik1> yes, i have to find my memory :-) 15:04:51 <pknirsch> hehe :) 15:05:25 <skalnik1> well, we discussed things about Test Day 15:05:48 <skalnik1> then documentation and tests 15:06:12 <skalnik1> jvcelak propagated nice test of socmes in virtual machines on various archs 15:07:07 <skalnik1> in the doc topic i failed about marcela's activity 15:07:57 <skalnik1> i suppose it would be better if she can clarify her activity and communication with Don Domingo 15:08:43 <skalnik1> the last thing i'd like to remember - test day and live cd 15:09:18 <skalnik1> we would like to provide a rpm that will install all test cases and dependency 15:09:20 <skalnik1> that is 15:11:10 <marcela> Don Domingo told us that he's working content of Developers GUide and he will tell us in two weeks, if he want use the content of the presentation about programming techniques 15:11:24 <marcela> in Developer Guide or somewhere else 15:12:05 <marcela> so let's concentrate on Deployment Guide now in Documentation part 15:12:29 <pknirsch> thanks! 15:13:35 <pknirsch> that sounds like a good choice. and if we don't get the programming techniques into the Developers guide we can always think about writing a small whitepaper about it. 15:14:53 <pknirsch> lets continue on to the test day, which is coming up in 21 days! ;) 15:14:58 <pknirsch> #topic Test Day Oct-22 15:15:51 <pknirsch> looks like we now have already a pretty good start for the test day wiki with tests and testplans for the various things. 15:16:06 <pknirsch> how about the live-cd? i think that was still a hot topic for last week, right? 15:16:10 <pknirsch> any news there? 15:16:55 <marcela> I suppose jvcelak measured it? 15:17:19 <skalnik1> yes, but there were problems 15:17:27 <jvcelak> I tried to run the tests from last year. 15:18:15 <jvcelak> And I was able to run only few of them because of I/O errors probably caused by combination of ext4 + scratchfs on LiveCD. 15:19:18 <jvcelak> Unfortunatelly I didn't manage to figure out how to bypass it. 15:20:24 <pknirsch> hm 15:20:54 <pknirsch> so what are our options then? leave out that test that fails or don't do a live-cd at all i guess? 15:22:12 <marcela> we thought about rpm which require all needed packages 15:22:16 <marcela> or repo... 15:23:13 <plautrba> +1 for rpm with needed dependences 15:23:16 <pknirsch> ah, so that users won't need the live-cd but only need to install one "main" package? 15:23:24 <pknirsch> s/users/testers/ 15:23:36 <jvcelak> +1 for rpm 15:23:43 <plautrba> yes 15:23:44 <skalnik1> +1 rpm 15:23:48 <pknirsch> +1 rpm 15:23:58 <plautrba> live-cd could be used as source for rawhide instalation 15:24:07 <pknirsch> mhm 15:24:27 <plautrba> *would 15:24:48 <pknirsch> lets do that then. 15:25:55 <pknirsch> and i've seen plautrba has prepared some scomes tests as well, right? 15:26:10 <plautrba> yes 15:26:27 <plautrba> it's not published yet 15:27:04 <plautrba> need to unify sending results 15:27:11 <pknirsch> ah, ok. i was looking at the test day page and didn't seem them there yet 15:27:28 <plautrba> i'll put it there tommorow 15:27:36 <pknirsch> thanks :) 15:29:19 <plautrba> scomes tests are focused to find out if it creates right results 15:29:48 <plautrba> and then we can start measuring systems, utilities, ... with it 15:30:16 <pknirsch> awesome! 15:30:45 <pknirsch> are we missing anything else for the test day other than that? the results table seems to be in need of updating i guess :) 15:31:20 <skalnik1> i should update wiki - remove live cd things 15:31:39 <skalnik1> then i have to send an e-mail to previous attendees 15:32:02 <skalnik1> to ask them to participate again 15:33:16 <pknirsch> that would be great, skalnik1 ! 15:33:53 <skalnik1> yes, i promised it but ... :-( 15:35:17 <pknirsch> lets just get the wiki updated till next week, thats fine. we still have 3 weeks, so no problem so far. :) 15:35:33 <pknirsch> anything else we're missing for the test day? 15:38:03 <zodbot> Announcement from my owner (stickster): Fedora Board IRC meeting in ~25 minutes -- refer to https://fedoraproject.org/wiki/Board/IRC for details 15:38:26 <marcela> I suggest all of us can run the tests 15:38:35 <marcela> in advance 15:38:50 <pknirsch> good idea. in case some of them don't work as expected. 15:38:52 <marcela> we can polish tests in case we will have strange numbers 15:39:05 <pknirsch> yes, definitely 15:40:03 <pknirsch> lets all do that till next thrusday so we have some time to fix things in case we need to or discuss what to do about tests that don't work properly 15:40:31 <pknirsch> alright, but only 20 minutes left till next meeting, so lets go to the next topic 15:40:43 <pknirsch> #topic Documentation 15:41:56 <pknirsch> I've sent out the draft that Rudi sent me to all of you. it's "only" the topics so far, but i've already seen a few things i'd like to either add or change already. i'll send him my changes either today or tomorrow and will copy all of you on the email as well. 15:42:55 <marcela> thank you 15:43:06 <pknirsch> i think we'll need to start thinking about detailed content for the single topics as well as i'm not sure how much Rudi can do with what he has so far. 15:44:02 <pknirsch> but i'll ask Rudi as well what/if he needs more input or details 15:44:17 <marcela> I volunteered to write about it, so if we make some notes to topics. I can write text 15:44:18 <pknirsch> and we need to check how much time he actually has to work on this besides his other tasks. 15:44:23 <pknirsch> perfect! 15:44:30 <marcela> and he can make from it English text 15:46:42 <pknirsch> if possible i'd like to get something we can actually show to people at the end of the month. doesn't need to be finished or polished, just a rough draft where people can start reading and commenting on. 15:47:19 <pknirsch> any more comments regarding documentation? 15:47:41 <pknirsch> (sorry for rushing things now :/) 15:47:46 <marcela> I can start some wiki 15:47:53 <marcela> we can paste it there 15:48:24 <pknirsch> mhm. that would make it easier for Rudi to pick things up as well i think and for us to share the updates. 15:48:43 <pknirsch> especially with his 12h timezone difference 15:49:18 <marcela> yes 15:49:32 <marcela> I'll start tommorow, we can summarize there our findings 15:49:39 <pknirsch> alright, thanks marcela :) 15:50:51 <pknirsch> last topic are stuff that we can work on in the next weeks/months 15:51:02 <pknirsch> #topic Tasks 15:51:34 <pknirsch> i've sent out a quick overview of what i saw as things we should work on over the next weeks/months yesterday. 15:51:59 <pknirsch> we've basically already covered the test day and documentation parts, so lets quickly run over the remaining things: 15:52:14 <pknirsch> - NetworkManager needs extension to allow speed switches for ethernet devices 15:52:55 <pknirsch> that'll be quite a bit of C work and upstream involvement needed for that. 15:53:08 <plautrba> that should be quite problem, but i'm going to talk to nm guy here in brno about posibilities how to do it 15:53:20 <pknirsch> great, thanks plautrba 15:53:46 <pknirsch> and yea, i took a brief look at it a while ago and it didn't look that easy, especailly together with the 2nd point: 15:53:54 <pknirsch> - Based on that NM needs to have an option to allow the physical link to drop for a few seconds before actively then shutting down the interface. Otherwise each time we switch speeds NM will 15:53:54 <pknirsch> ifdown and then ifup that interface which will lead to lost connections and VPN being dropped which isn't necessary 15:54:55 <pknirsch> lets just see how far we get with this 15:55:08 <plautrba> maybe we can somehow inform nm via dbus to not accept any events from hal/... for ten seconds when we do change or something 15:55:18 <pknirsch> mhm 15:56:12 <pknirsch> i was thinking about having some sort of timeout for lost connections as well which would allow you to unplug your cable as well 15:56:22 <pknirsch> s/lost connections/lost link/ 15:56:39 <pknirsch> which you could set for any network device 15:56:56 <jvcelak> pknirsch: this will be useful even if not switching the netcard speed 15:56:58 <plautrba> but you have somehow recognize that new cabel is same as old 15:57:09 <pknirsch> jvcelak: exactly 15:57:19 <pknirsch> jvcelak: or for a short loss of wireless signal 15:57:32 <pknirsch> plautrba: yea, thats the tricky part i think :/ 15:57:52 <jvcelak> we have to discuss with somebody experienced with it 15:58:18 <pknirsch> quick note btw, especially to plautrba: I just talked with the systemtap guys today and our systemtap scripts will now be sort of officially tested and included in the upstream systemtap ;) 15:58:33 <plautrba> cool 15:58:37 <pknirsch> yep :) 15:59:09 <pknirsch> ok, next point was: 15:59:11 <pknirsch> - tuned-adm needs to be changed to work without kobo. 15:59:28 <pknirsch> that should be fairly easy to do as we're not using that much functionality from kobo so far. 16:00:09 <pknirsch> i can easily do that, but if you want to do it marcela that would be fine with me as well :) 16:00:18 <pknirsch> (but only if you think you have time) 16:00:28 <plautrba> i'll do it tommorow or at monday 16:00:43 <pknirsch> ok, check :) 16:00:54 <pknirsch> and thanks 16:01:13 <pknirsch> last 2 points then 16:01:15 <pknirsch> - More profiles and possibly improved scripts for tuned-adm 16:01:16 <pknirsch> - Talk with virt & performance teams about possible profiles for them for RHEL-6 16:02:00 <pknirsch> first one is just looking at the profiles we have so far and see if we can improve them. i mainly dislike some of the names, but mye, didn't come up with anything better when i wrote them 16:02:19 <marcela> great 16:02:35 <pknirsch> 2nd points i'll try to get in touch with the perf and virt guys over the next week and see if they have some input for us. 16:03:40 <pknirsch> alright, but i think we should make room for the Fedora Board IRC meeting now :) 16:03:54 <pknirsch> any last comments/concerns/questions? 16:04:51 <pknirsch> alright, then thanks everyone for joining and talk to you next week again! 16:05:00 <pknirsch> #endmeeting