fedora-meeting
LOGS
15:01:55 <pknirsch> #startmeeting Power management for Fedora
15:01:55 <zodbot> Meeting started Thu Oct  8 15:01:55 2009 UTC.  The chair is pknirsch. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:55 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:02:12 <pknirsch> First of all, as usual, hello and welcome everyone :)
15:02:43 <skalnik1> hi 8-)
15:02:49 <pknirsch> hey :)
15:03:06 <pknirsch> everyone else here i suspect? i see marcela and jvcelak and plautrba
15:03:22 <jvcelak> hi everyone :)
15:03:39 <plautrba> hello
15:03:59 <pknirsch> alright, lets get rolling then. :) typical last week recap
15:04:07 <pknirsch> #topic Last week recap
15:04:59 <pknirsch> We talked a lot about the upcoming test day and the issues that we still saw. Main thing was how to help people test with a live-cd, and we agreed to do that via a rpm.
15:05:27 <pknirsch> Also updates to the wiki in regard to tests and so forth were discussed
15:05:55 <pknirsch> And marcela suggested that everyone runs the tests on their machines to see if they work properly.
15:06:57 <pknirsch> Next we talked about documentation. We figured that doing a wiki to collect our ideas would be a good idea and marcela took care of that.
15:07:35 <pknirsch> Last but not least we quickly went over the remaining tasks and distributed who was working on what.
15:08:03 <pknirsch> thats about it for last week i think.
15:08:07 <pknirsch> anything i forgot?
15:08:55 <pknirsch> alright, then lets move on to the next topic, test day again :)
15:09:04 <pknirsch> #topic Test Day Oct-22
15:09:37 <pknirsch> Lets start with the wiki. skalnik1, any updates there?
15:09:55 <skalnik1> yes, I made changes
15:10:28 <pknirsch> great!
15:11:40 <pknirsch> How about the test results template? shouldn't we change that to reflect the tests we're doing?
15:11:44 <skalnik1> and the other helped me
15:13:06 <skalnik1> yes
15:14:21 <pknirsch> alright, but i really like the page by now, so we're getting there :)
15:14:39 <pknirsch> i see the rpm is there already too, so thats great!
15:15:08 <skalnik1> i suppose thank to marcela
15:15:17 <pknirsch> thanks marcela ! :)
15:15:26 <pknirsch> and plautrba for his scomes tests
15:16:09 <marcela> ehm that's first attempt, it will need more
15:16:20 <pknirsch> :)
15:16:48 <marcela> we'd like to automatically tar files and sent them somehow
15:17:33 <pknirsch> to make it easier for testers, yea, sounds like a great idea if it's possible to do that automatically somehow.
15:18:17 <pknirsch> btw, has everyone ran the tests or found any issues with them? i've just done so today and the worked fine on my machine here with reasonable results.
15:19:16 <jvcelak> i run them too (except bltk tests) and had no problem. except some missing package (but i think this will be solved with new rpm)
15:19:21 <plautrba> we have some strange results on scomes tests with network
15:19:33 <plautrba> jhutar is focused on that
15:19:50 <pknirsch> jvcelak: great!
15:20:01 <marcela> I suppose we should make some easy output from bltk tests. It's too much numbers, nothing is clear.
15:20:10 <pknirsch> plautrba: ok, good. haven't seen that in my tests, hopefully it's something fixable
15:20:50 <pknirsch> you mean like a summary for the results for bltk, marcela?
15:20:57 <skalnik1> marcela, i'm just working on bltk results filtering
15:21:03 <marcela> skalnik1: great
15:21:07 <pknirsch> perfect :)
15:21:12 <skalnik1> i have data from rhel5 and f12
15:21:32 <skalnik1> now i have to make user friendly output
15:21:32 <pknirsch> oh, good!
15:22:29 <skalnik1> then i'd like fulfill it with results from f11
15:23:38 <pknirsch> mhm
15:25:23 <pknirsch> alright, so we'd need to get an updated rpm done till next week, investigate if automatic uploading of results is possible somehow and do it if is, fix the scomes test issues that plautrba sees and bltk result filtering.
15:25:47 <pknirsch> oh, and update the test result template on the wiki
15:25:51 <jvcelak> actually i have some remarks to the tests: i think the bltk tests on non-laptops are useless for us. it would be good to tell the people without laptops, that they don't need to run bltk test on them. now there are some misleading information (for example in the tune-adm test: changing profile to "server-powersave" and running bltk)
15:26:25 <pknirsch> good point jvcelak.
15:27:22 <marcela> hm I wrote use your laptop...
15:27:38 <marcela> but maybe I make it prerequisity of measurement
15:29:13 <jvcelak> marcela: i didn't notice anywhere...
15:30:05 <pknirsch> maybe we should add that to the overall testing section on the wiki. that measurements with laptops are typically more useful than with desktop machines.
15:31:04 <skalnik1> i think both are useful but laptops doesn't need additional hw
15:32:06 <marcela> I'm little worried about number of tests. It takes really long time to measure it all. I hope testers will be willing to test it.
15:32:09 <jvcelak> we can just "mark" the tune-adm and bltk test not to be performed on desktops
15:34:41 <marcela> we have to do more changes, moved the easy test on top of list and so on
15:35:11 <pknirsch> maybe separate the test lists in desktop and laptop tests?
15:35:24 <pknirsch> (just an idea)
15:35:54 <skalnik1> pknirsch:  good idea
15:38:01 <pknirsch> alright, so lets rework that list a bit till next week.
15:38:07 <jvcelak> another idea: we could put the "tuned" and "init 1" together ... because both of this tests needs booting into single mode. and changing runlevels or rebooting is not very pleasant. or we can just write it there, that these tests can be performed in sequence.
15:38:24 <pknirsch> sounds good, jvcelak
15:39:13 <skalnik1> then we will rework it to test groups
15:39:27 <jvcelak> +1
15:39:34 <pknirsch> +1
15:39:57 <marcela> +1
15:40:48 <plautrba> +1
15:41:07 <jvcelak> 1) all systems - init 5  2) all systems - signle 3) laptops-only  ?
15:42:16 <skalnik1> yes, i know this +1 - you'll liberate her and I'll get her married
15:42:30 <pknirsch> haha :)
15:42:31 <jvcelak> :D
15:42:50 <skalnik1> pknirsch: great, you know Jara Cimrman
15:42:54 <skalnik1> :-D
15:43:43 <pknirsch> hehe
15:44:39 <skalnik1> well, jvcelak the groups seems to be good
15:44:48 <pknirsch> so we need volunteers for rpm update, automatic uploading, scome test fixes and blkt filtering ;)
15:44:49 <skalnik1> +5 :-D
15:45:03 <pknirsch> s/blkt/bltk/
15:45:37 <jvcelak> but we should name the groups some better way
15:45:52 <skalnik1> pknirsch: bltk filtering is not intended to put it on wiki
15:46:06 <skalnik1> or in the first step for sure
15:46:08 <pknirsch> skalnik1: oh ye, ofc, silly me :)
15:46:29 <skalnik1> i will use it for final execution of results
15:46:35 <pknirsch> mhm
15:46:57 <skalnik1> perhaps we should have something user friendly in the future
15:47:31 <skalnik1> I need to have usable results to report it
15:47:44 <skalnik1> next step should be to automate it
15:49:28 <pknirsch> yea
15:50:49 <pknirsch> alright, anything else regarding test day? otherwise i'd like to continue to documentation
15:51:25 <jvcelak> it would be perfect if all the tests save their results in one directory. which will be compressed and send by some "collecting" script we can provide. to have all the results in the same format.
15:51:59 <plautrba> jvcelak: that's what marcela working on
15:52:07 <jvcelak> i'm not sure wheather plautrba didn't mentioned it
15:52:12 <jvcelak> oh... i see :-)
15:52:19 <jvcelak> sorry ;)
15:52:25 <pknirsch> :)
15:52:54 <skalnik1> marcela, plautrba, then we should adjust bltk.conf file
15:53:12 <skalnik1> or should? it's one possibilty
15:53:12 <plautrba> skalnik1: or just pack files after test
15:53:19 <skalnik1> yes
15:54:03 <skalnik1> bltk...f12 can have bltk.conf in the home directory that has higher prio
15:54:20 <marcela> I suppose we can install some script with rpm, tar all tests and rm them ;-)
15:56:22 <pknirsch> think you can get that working till next week, marcela?
15:56:26 <pknirsch> :)
15:56:29 <marcela> yeah
15:56:42 <pknirsch> cool, lets aim for that then.
15:58:13 <pknirsch> but lets move on to the next topic, otherwise we'll be running way late again in our meeting ;)
15:58:32 <pknirsch> #topic Documentation
15:58:59 <marcela> I sent an email with real draft of documentation to Rudi but no answer yet
15:59:11 <pknirsch> And thanks marcela for putting up that documentation wiki!
15:59:20 <marcela> it's on internal brno wiki for the meantime
16:00:06 <marcela> question is do we need scomes and desktop thinigies like powerdevil in Deployment Guide for users?
16:01:17 <pknirsch> it's a good idea imo. people will be using powerdevil and g-p-m and we should include them at they are related to the power management side from a desktop view.
16:01:52 <pknirsch> we might want to get some desktop folks on board though to provide some more details about them, e.g. richard hughes for g-p-m and one of the kde folks for powerdevil
16:02:14 <pknirsch> or at least try to get the on board ;)
16:02:20 * nirik notes as a side note that there is also xfce4-power-manager too. ;)
16:02:27 <pknirsch> ahh!
16:02:31 <plautrba> i have few lines about using tuned-adm with powerdevil (in czech), i could translate it and add there too
16:02:40 <marcela> nirik: thanks
16:02:41 <pknirsch> i see a volunteer for documentation there, nirik ;)
16:02:56 <marcela> nirik: I'm using only kpowerdevil, so...
16:03:07 * nirik isn't sure he has time, but can try and help if you let me know exactly what you need. ;)
16:04:08 <pknirsch> nirik: we're trying to put together more consistent documentation about power management from a user perspective, so any quick 1-2 paragraph description on what xfce4-power-manager does would be already very helpful
16:04:10 <nirik> I'd be happy to note things about xfce4-power-manager in a wiki if you let me know where and when. ;)
16:04:16 <pknirsch> hm
16:04:39 <pknirsch> can we move the wiki to fedora then, marcela?
16:04:48 <marcela> yeah, tommorow
16:04:53 <pknirsch> perfect, thanks :)
16:05:28 <nirik> cool.
16:07:17 <pknirsch> let's poke rudi once more this week and i hope i'll find some time to add some more content to the documentation wiki as well.
16:09:11 <pknirsch> i'll be probably adding some more detailed sections about the single tunings in the ktune scripts and tuned itself, explain each a bit and what it does, how to enable/disable it and how much you can save with it.
16:10:29 <pknirsch> have you heard anything yet from Don, marcela ?
16:10:34 <marcela> not yet
16:10:48 <marcela> I believe it will be 2 weeks he mentioned the next week
16:10:59 <marcela> but I can bother them both next week
16:11:08 <pknirsch> alright, lets to that then :)
16:12:13 <pknirsch> anything else we need to cover regarding docu?
16:12:21 <marcela> no
16:12:43 <pknirsch> oki, then lets finish with how we're standing regarding the longer term tasks.
16:12:46 <pknirsch> #topic Tasks
16:13:03 <pknirsch> plautrba, did you get a chance to look at NM yet?
16:13:27 <plautrba> pknirsch: no, the guy from brno is ooo this week
16:13:50 <pknirsch> roger, then we'll see where we stand with it next week.
16:14:39 <plautrba> regarding to tuned
16:14:58 <plautrba> kobo dependence wes remote with last commit
16:15:04 <plautrba> *was
16:15:05 <pknirsch> perfect, thanks :)
16:15:19 <plautrba> and i'd like to mention that jvcelak works on unit tests for tuned
16:15:29 <pknirsch> ohh, great!
16:15:35 <plautrba> to make tuned more enterprise :)
16:15:38 <plautrba> http://git.fedorahosted.org/git/?p=tuned.git;a=shortlog;h=refs/heads/with-tests
16:16:24 <jvcelak> yes. but it's a bit outdated in git. i have rewritten some parts. the tests for plugins are complete. and now i am finishing the tests for tuned-adm
16:16:37 <plautrba> there have been found and filled some bugs based on these tests already
16:17:47 <pknirsch> awesome :)
16:17:47 <plautrba> and I'm going to appply for post-receive hooks on git repo to send commits/pushes via email
16:19:15 <plautrba> that's all from me for now
16:19:58 <pknirsch> i haven't heard anything yet from the virt or perf guys, but will keep poking them
16:21:56 <pknirsch> and regarding the scripts i've done some changes in my local repo and will commit and push them early next week i think
16:24:13 <pknirsch> i think thats it for this week.
16:24:29 <pknirsch> anything else or any last comments/concerns/questions?
16:24:44 <skalnik1> no
16:25:58 <pknirsch> Thanks everyone for joining today then and talk to you next week again!
16:26:01 <pknirsch> #endmeeting