IRC logs for #baserock for Thursday, 2015-10-01

*** jwadas1 has joined #baserock02:21
jwadas1Hey all. Anyone home?02:21
*** Walkerdine has joined #baserock02:53
WalkerdineMy morph seems to be broken now03:30
*** gtristan has quit IRC03:31
*** gtristan has joined #baserock04:06
*** gtristan has quit IRC04:41
*** gtristan has joined #baserock04:42
*** zoli__ has joined #baserock05:24
*** zoli__ has joined #baserock05:26
*** zoli__ has quit IRC05:30
*** gtristan has quit IRC05:48
*** gtristan has joined #baserock06:00
*** zoli__ has joined #baserock06:26
*** zoli__ has quit IRC06:30
*** paulwaters_ has joined #baserock06:35
*** jwadas1 has quit IRC06:50
gtristanOk so... ubuntu lts version of btrfs-tools; segfaults when mkfs.btrfs is given the --features options to unflag some features07:04
gtristanit segfaults, it does not print the unrecognized option message checked for in extensions/writeexts.py07:05
* gtristan is checking if it's one of the particular 3 options passed, or any of them07:06
gtristanif all three are omitted... image is almost created... fails on smth else (/me collects log for that one)07:06
*** ssam2 has joined #baserock07:07
*** ChanServ sets mode: +v ssam207:07
gtristanah, the other error is lack of extlinux tool07:25
gtristanlooks like definitions/extensions uses various system tools... are those dependencies declared somewhere ? so that, e.g., tools like ybd can check for them before blindly running the scripts ?07:26
gtristanaha, now I have a gnome-system-x86_64.img which I dont expect to boot :)07:33
* gtristan tries07:33
gtristanboots into a shell, that's a start07:40
*** zoli__ has joined #baserock07:42
*** zoli__ has quit IRC07:46
ssam2gtristan: they aren't declared in any machine-readable way at present07:52
ssam2the original idea was that everything would run in a Baserock system/chroot, which would have all the deps available07:53
ssam2hi by the way :-)07:53
gtristanNice to meet here :)07:53
*** toscalix__ has joined #baserock07:54
wdutchhello all concerned with CIAT, the history has vanished from the frontend but I did back it up07:54
gtristanI'm taking down notes of all the brute force hacks I needed (not SO many btw) to get where I did with ybd on ubuntu 14.0407:54
wdutchI'm experimenting with the new pipeline methods, I tried to get it all done last night while nobody was looking but it took a long time due to a change in an upstream for foundation07:55
wdutchpdar: deploy seems to be breaking at the lines08:00
wdutch# Remove the artifacts08:00
wdutchsudo rm -rf "$ybd_base"/artifacts/"$s_n".*08:00
wdutchhere's a log
wdutchI have no idea why :(08:01
wdutchpossibly it's not that but the command after it when control is returned to the python wrapper08:01
gtristanjjardon, around ?08:03
wdutchpdar: ignore me08:05
pedroalvarezwdutch: we don't have arm pipeline anymore?08:05
wdutchpedroalvarez: not currently! sorry08:05
wdutchit's a bit chaotic08:05
wdutchbut should be super easy to spin up new pipelines once I'm done debugging08:06
* pedroalvarez looks a the new names of the columns08:06
wdutchthe names are generated from the name of the pipeline + the 'step', which is one of Build, Deploy, Test, Publish08:07
wdutchIntegration is special for now08:07
*** zoli__ has joined #baserock08:08
pedroalvarezi liked the word "generated"08:08
* wdutch needs a test pipeline that isn't subject to the whims of upstream08:17
*** mariaderidder has joined #baserock08:21
wdutch"TypeError: coercing to Unicode: need string or buffer, bool found"08:21
wdutchI feel like we've seen that before08:21
pedroalvarezcould we keep triggering builds O;)08:26
wdutchwe could08:26
pedroalvarezoh, json gives you a "eta"08:30
rjekSoon, then.08:31
pedroalvarezI was using the time of the previous build08:31
*** jonathanmaw has joined #baserock08:36
wdutchright lets get an arm pipeline08:44
*** gtristan has quit IRC08:46
* pedroalvarez has progress bars for the builds :)08:47
pedroalvarezif the build is taking more than the previous build, it will stay at 90%08:49
toscalix__pedroalvarez: what is the criteria used to define "progress"?08:53
toscalix__in the progress bar?08:54
*** toscalix__ is now known as toscalix08:54
pedroalvarez("time of actual run" *100) / "time of last build"08:55
toscalixis it possible to change the "time of last build" by "average time of a build" in the context of the demo?08:59
toscalixI mean, the demo will be a controlled environment, so is it possible to define an average time that we can be confident on?08:59
wdutchthe average will be a vast over estimate09:00
toscalixI think your equation is good for uncontrolled enviromnments, that is, when we cannot stablish a good reference09:00
toscalixI wonder if we can in the context of the demo09:00
wdutchthe data is skewed by the rare builds which can take >3hours09:01
toscalixso what you are saying is that we cannot define a reference time in the context of the demo?09:01
wdutchwe could define one09:01
wdutchthe average would be a bad way to define it09:01
toscalixdo you have a better suggestion then?09:01
pedroalvarezit should be possible to hardcode a "time of last build" yes09:02
Zaramedian vs mean?09:02
toscalixwe could use the median, yes09:02
toscalixof some values, if the distribution of values is a tall gauss chamber or similar function09:03
pedroalvarezgetting more than one value of previous build times, will be a pain09:03
toscalixpedroalvarez: understood09:03
pedroalvarez(from my non-js experience POV)09:04
toscalixso we will use the "time of last build" because getting data for a better reference is too painful09:04
pedroalvarezit, of course, can be done09:05
* pedroalvarez triggers a build, and looks at the status bar09:05
toscalixWe have other tasks to do, so good enough for me now09:05
toscalixpedroalvarez pdar let me know when you consider the progress bar done so we talk about the dialogs. I am taking a look at them09:07
pedroalvarezIf we only want progress bars in the builds column, then I think it's done09:08
pedroalvareztoscalix: pdar is not around yet. I think we can start talking about the dialogs09:09
pdari had imagined them in the builder box09:09
pedroalvarezoh, he is!09:09
pedroalvarezpdar: what exactly?09:10
pdari am!09:10
pdaroh the progree bar itself09:10
toscalixpdar: that is a good idea09:11
toscalixinstead of having a different dialog, we include the info in the box itself09:11
toscalixpedroalvarez: how do you see that?09:12
pedroalvarezdone :)09:12
wdutchmay I quickly reboot orchestration?09:13
toscalixah, you mean the profgrees bar09:13
toscalixI also like it better, yes09:13
wdutchgoing once09:13
wdutchgoing twice09:13
wdutchgoing down09:13
toscalixWe only need to be carefull with colours09:13
pedroalvarezwhen built, green, when building orange09:14
toscalixIf it will be green, the contrast need to be high. Remember than people will see the screen from 1 or 2 meters of distance09:14
rjekHave bloody great tick or cross09:14
toscalixpedroalvarez can you use the same color that the logo, and instead of white, the same color than the background?09:15
toscalixin the progress bar?09:15
pedroalvarezfor the bars, I think I only have 4 choices09:16
pedroalvarez(of colors)09:16
pedroalvarez(using the twitter bootstrap library)09:16
toscalixwhich are?09:17
SotKpedroalvarez: you can override the css in your local stylesheet09:17
pedroalvarezyes, I can try to do that :)09:18
pedroalvareztoscalix: so, different colors when building and finished building is ok?09:18
wdutchwe have an arm pipeline but I can't push to the candidate ref :(09:19
wdutch"error: src refspec cu010-trove/arm-base does not match any."09:19
SotKI think something like `.pass > .progress { background-color: #77ee77; }` would work to change the background to green when in a green box09:19
* pedroalvarez is still trying to understand "use the same color that the logo, and instead of white, the same color than the background"09:19
*** gtristan has joined #baserock09:19
toscalixpedroalvarez: in my opinion, there are two options: 1.- Going for green, ensuring a high contrast between the box color and the bar 2.- going for the calors we are already using, that is, dark (background) and red/white09:19
rjekcolours ;-)09:20
toscalixanalize/analise ... and my favorite one: virtualization - virtualisation09:21
rjektoscalix: analize does not mean what I think you think it means.09:21
rjekartefact only became the more popular spelling in British English in the past ~20 years or so09:21
toscalixpedroalvarez: we don't have such a problem in the Canary Islands, neither in South/central america :-)09:21
pedroalvareztoscalix: ooh, yes we have it09:22
toscalixartifat (english) artefacto (spanish) from the latin arte facto - done with art09:23
pedroalvareztoscalix: I'll be having a meeting soon, I'll ping you when I'm done09:24
toscalixso the british are becoming more influenced by latin lately..... strange09:24
toscalixand cool at the same time09:24
rjekartifact and artefact have been used interchangably in English for hundreds of years09:25
rjekBut it appears that the British have recently decided that the e-form is more beautiful.09:25
toscalixnowadays everything is e-09:26
*** ssam2 has quit IRC09:28
wdutchtaking orchestration down briefly09:31
pedroalvarezwdutch: publish is not being triggered09:32
pedroalvarezwdutch: I guess it's my fault09:33
wdutchit's my fault09:33
wdutchwill deal with it hopefully before lunch09:33
wdutchyay arm build09:35
toscalixthis is why we always need to set up a dev and production instances. When the deadline is close..... we block each other09:35
wdutchI tried to set one up but buildbot and nebula didn't get on09:36
wdutchalso I have no idea how to set up a trove09:36
pedroalvarezwdutch: I could have helped, 10 minutes of my time09:38
wdutchpedroalvarez: feel free to give it a go if you have time, the problem was when running orchestration buildbot timed out09:38
* pedroalvarez is here again09:49
* wdutch wishes chinese addresses would stop trying "GET /manager/html"09:49
pedroalvarezyou only realise how dangerous is internet when you put something on production09:51
pedroalvarezan public09:51
wdutchlike when we were DoS'd by a russian the other day ...09:51
rjekI wish russian IP addresses would listen to my mail server when it says "AUTH used when not advertised."09:51
rjekBut, wooyay fail2ban09:51
pedroalvarezI need to learn how to configure it09:52
rjekIt's hateful and surprisingly complex09:53
rjekBut it does work09:53
pedroalvareztoscalix: I'm back09:53
pedroalvarezI don't really like how the status bars are right now09:53
pdarIt would be good if they only appeared when the builder was active09:55
* rjek can't see any progress bars, and feels he is missing out on all the excitement.09:56
rjekMOAR EIGHTS09:57
wdutchrjek: if you want to see the progress bars don't use a terminal09:57
toscalixpdar: agree09:58
pedroalvarezonly if active, ok09:58
SotK+1, also can they lose the 20px margin they have underneath them? :)09:59
pedroalvarezSotK: I can try :)09:59
rjekOr put the text "This space intentionally left blank" there09:59
pedroalvarezSotK: I'd like to solve what happens when you make the screen small09:59
SotKpedroalvarez: what do you mean?10:00
SotKhow everything goes vertical?10:00
pedroalvarezgoes vertical, but only using the 20% of the screen10:00
pedroalvarezbecause I'm crap at css10:01
SotKuse col-xs-2 rather than col-md-2 on the lane divs10:01
pedroalvarezyay, an expert in tha house10:01
wdutchanybody around who knows ybd?10:06
wdutchI need help with
SotKcache.get_cache() appears to return False if cachedir doesn't exist10:08
rjekThat's a surprising API10:08
gtristanwdutch, I wont claim to 'know' it... but I've just spent a few days arm wrestling with it ;-) ... you have linux-user-chroot I presume ?10:09
wdutchgtristan: there is no linux-user-chroot on the host but it's worked in the past without it10:10
gtristanyeah I see, it should pass without it... I insisted to install it myself10:11
gtristanwdutch, latest stable tag ?10:11
gtristanhmmm thats what I'm using too... HEAD had problems10:12
SotKwdutch: is "genivi-demo-platform-x86_64-generic.190a24b09d38856cc5cf128b55c25b2d4ae72fa08dcfea85171ad2e9f0c6af1a" in your cache?10:12
* wdutch looks10:13
* wdutch twiddles thumbs10:14
wdutchthere is quite a lot of cache10:14
wdutchSotK: it looks like it's not there10:16
SotKhm, are you using a remote cache too at all?10:17
wdutchI have no idea, how would I know?>10:17
* SotK has no idea :)10:17
wdutchpdar: Idon't suppose you have any insight into this problem?10:18
pdarno insight10:21
pdarcan look into it though10:21
pedroalvarezwdutch: maybe the artefact is in there, but ybd can't open it? (permissions)10:21
* SotK doesn't see how deployment could be attempted before assembly, so assumes that either something deleted it between assembly finding it and deployment not, or its in a remote cache somewhere so assembly finished fine10:21
pdarits not in a remote cache i dont think10:22
wdutchpedroalvarez: ybd is being run as root afaik10:22
wdutchyup it's run as rot10:22
pedroalvarezheh, OK10:23
SotKthere is a kbas-url set in your config, is kbas actually running?10:24
pdarit is, I am looking at why it aint working now10:24
SotKis the artifact I mentioned earlier in kbas's cache?10:25
pdarit isnt10:25
pdarybd doesnt seem to be talking to kbas10:26
*** CTtpollard has joined #baserock10:26
wdutchpedroalvarez: where is the code for ' curl -X POST -F ...'10:27
pedroalvarezwdutch: in the testing_trigger.sh10:28
pedroalvarezor whatever is under the triggers/ folder10:28
*** toscalix has quit IRC10:31
pdarho hum, when trying to use the same kbas server from my baserock dev machine I get the following:10:32
pdar15-10-01 00:00:06 [libvorbis] WARNING: remote artifact server is not working10:32
*** gtristan has quit IRC10:47
*** gtristan has joined #baserock10:52
* pdar fixes the remote caching problem10:57
wdutchthat's good because this build is 90% done according to the bar!10:57
pedroalvarezwdutch: hahahah10:59
pedroalvarezwdutch: shhh!10:59
* wdutch isn't sure what to do while waiting for this build11:00
pedroalvareztrigger a test, and make it trigger publish11:01
wdutchhuh testing finished an no publish :(11:12
pedroalvarezas I said, probably my fault11:13
wdutchI thought I'd fixed it11:13
wdutch127.0.0.1 - - [01/Oct/2015 11:11:35] "POST /testing_complete HTTP/1.1" 200 011:13
*** bwh_ is now known as bwh11:13
wdutchit's pinging bottlerock c'rectly11:13
wdutch'%s Pulbish' ...11:14
* wdutch finds the problem11:14
CTtpollardheh bottlerock, nice name11:14
wdutchCTtpollard: Thanks! :D although I regret it ...11:14
*** ssam2 has joined #baserock11:26
*** ChanServ sets mode: +v ssam211:26
wdutchpublish is now being triggered11:28
pedroalvarezwdutch: thanks!!!11:34
pedroalvareznicer progress bars now11:34
wdutchred? :(11:34
* wdutch finds the in progress boxes being red quite distressing11:35
wdutchthen it's the same as the bar11:35
SotKwdutch: red means last build failed11:35
pedroalvarezwell, red now means "not success"11:35
SotKpulsing yellow outline means in-progress11:36
wdutchpersonally I think that's misleading but okay11:36
pedroalvarezI can go with purple for interrupted builds ;)11:36
jmacsI think the pulsing yellow should be accompanied by a low frequency throbbing noise11:36
SotKwdutch: why is it misleading?11:37
SotKpedroalvarez: that would be nice11:37
pedroalvarezI think it may be confusing11:37
wdutchSotK: I interpret the box as being the current build11:37
tiagogomeswhat's provisioning? deploying?11:38
pedroalvarezprogress bar is the current build, color of box, previous build11:38
SotKpedroalvarez: yeah, I guess purple doesn't have an obvious meaning like red/green11:38
pedroalvareztiagogomes: yes, deploying images for the tests11:38
pdarcould make a key?11:39
pedroalvarezwe could make a key, yes11:39
jmacsMouseover text, perhaps?11:40
pedroalvarezmouseover is an antipattern :)11:40
pedroalvarez(smartphones  can't do that)11:40
nowsterI wonder how xkcd looks on a phone...11:41
tlsapedroalvarez: I'd s/Automatic test/Test/11:41
tiagogomespedroalvarez so why not deploying? :) Also why some verbs are on the gerund and others on the infinitive?11:42
nowstersort of works with a long press in Chrome... not at all well in Firefox.11:42
nowster(That's the ALT text on the image)11:43
tlsaAnd s/C.I.A.T.:/CIAT:/11:44
tiagogomesAlso, the website is making my computer sweat a bit (on firefox)11:45
pedroalvareztlsa: I'll ponder that change11:46
pedroalvareztiagogomes: deploying may mean another different thing11:47
pedroalvarezwe are not doing "deployments" to anywhere, just generating disk images11:47
tiagogomesyou mean you are 'deploying' a tarball or raw image11:48
tiagogomesI don't think 'provision' would be better if that's the case. Deployment is the vocabulary that we use in Baserock, so I think we should stick to it11:49
tlsapedroalvarez: the trouble with "Automated test" is it wraps in narrow windows, which makes the blocks get missaligned11:55
*** zoli__ has quit IRC12:04
*** zoli___ has joined #baserock12:04
* Kinnison notes that if the tests aren't automated, the CIAT system can't run them anyway12:25
Kinnisonso I'm all for 'Test'12:25
*** gtristan has quit IRC12:32
*** gtristan has joined #baserock12:46
pedroalvarezhehe, I guess that "Automated" makes it sound better12:47
*** zoli___ has quit IRC13:12
pedroalvarezoops, I made the status bar flash too13:21
*** zoli__ has joined #baserock13:22
pedroalvarezit is of course an error, but it doesn't look bad13:23
*** toscalix__ has joined #baserock13:26
* SotK wonders what the intended effect was13:27
rjekhypnotic nested pulsation13:28
*** zoli__ has quit IRC14:32
*** gtristan has quit IRC15:02
pedroalvarezrjek: I think in 24h the world will be mine15:03
wdutchboo failed build15:18
wdutchlooks like my terminal running bottle got eaten15:20
*** paulwaters_ has quit IRC15:23
*** edcragg has quit IRC15:31
*** gtristan has joined #baserock15:32
Kinnisonscreen is your friend15:33
wdutchI forgot15:35
myselfset it to your login shell :)15:35
wdutchit was messy anyway because I'd restarted bottlerock midflight so as not to interupt that epic build15:37
wdutchironic in the end15:37
*** persia has quit IRC15:38
*** persia_ has quit IRC15:38
*** Lachlan1975 has joined #baserock15:47
*** zoli__ has joined #baserock15:48
*** ssam2 has quit IRC15:49
*** zoli__ has quit IRC15:50
wdutchthere's a red thing flashing at me15:51
wdutchit makes me think there's an emergency15:51
wdutchwell more pulsing than flashing15:52
nowsterabandon shop16:04
rjekthis is not a daffodil16:05
*** zoli__ has joined #baserock16:11
*** zoli__ has quit IRC16:12
*** jonathanmaw has quit IRC16:41
*** mariaderidder has quit IRC16:49
*** persia has joined #baserock16:57
*** zoli__ has joined #baserock17:18
*** zoli__ has quit IRC17:23
*** toscalix__ has quit IRC17:30
*** zoli__ has joined #baserock17:35
*** zoli__ has joined #baserock17:44
*** tiagogomes has quit IRC18:05
*** Lachlan1975 has quit IRC18:43
*** zoli__ has joined #baserock19:43
*** zoli__ has quit IRC20:09
*** paulwaters_ has joined #baserock20:23
*** paulwaters_ has quit IRC20:52
WalkerdineDoes the baserock system on the jetson have access to GPIO?21:31
*** jwadas1 has joined #baserock21:45
*** jwadas2 has joined #baserock21:48
*** jwadas1 has quit IRC21:52
*** zoli__ has joined #baserock21:56
myselfgah, of course you'd ask right after I leave where my board is sitting22:10
myselfdoes /sys/class/gpio exist?22:11
*** jwadas2 has quit IRC22:21
WalkerdineIt does not22:24
Walkerdinemyself: I'm guessing thats a problem22:30
myselfYeah, though I don't have my head around enough of that to confidently give further advice, it looks like that's probably because the kernel needs to be built with support, and then there may be some interaction with a devicetree, but I'm not sure how those pieces fit together22:39
WalkerdineWould you happen to know where I could find out more about how these peice together22:42
myselfnot off the top of my head, but some better-documented platforms like beaglebone might be instructive, as I think the OMAP folks were pretty heavily involved in devicetree's early days.23:47
WalkerdineYou dont think there would be a difference between the dev image and the genivi demo platform do you23:50
WalkerdineIn terms of having the GPIO show up23:50
Walkerdinein /sys/class/gpio23:50

Generated by 2.15.3 by Marius Gedminas - find it at!