IRC logs for #baserock for Friday, 2014-10-24

*** vmeson [~quassel@] has quit [Ping timeout: 245 seconds]06:04
*** vmeson [~quassel@] has joined #baserock06:05
Mode #baserock +ov richard_maw richard_maw by asimov.freenode.net07:02
*** richard_maw [] has joined #baserock07:02
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock07:16
Mode #baserock +o ChanServ by asimov.freenode.net07:16
Mode #baserock +oov ChanServ richard_maw richard_maw by asimov.freenode.net07:37
*** Kinnison [~dsilvers@gateway/shell/pepperfish/x-hhzujylmdusfoekc] has joined #baserock07:39
*** jmacs [] has joined #baserock07:39
*** ssam2 [] has joined #baserock07:56
Mode #baserock +v ssam2 by ChanServ07:56
ssam2pedroalvarez: are you merging the 14.40-fixes branch or waiting for another +1 in order to do so?07:59
ssam2if so, +1 to merge it ...07:59
ssam2or I can do the merge if you want07:59
pedroalvarezssam2: thanks!07:59
pedroalvarezI rememberd this morning that it never had a +207:59
radiofreeis there any chance i can get a fix into 14.40?08:00
radiofreeit's for nouveau-drm, has been tested08:00
pedroalvarezhmm.. what's broken?08:00
radiofreeit's loading the in-tree kernel module and not the out-of-tree kernel module08:00
ssam2getting it in to 14.40 would be difficult :) we could do a 14.40.2 jetson-only release if you think it necessary08:01
pedroalvarezactually it would be only the genivi system, no?08:02
radiofreeit's necessary for the genivi baseline08:02
ssam2right. So it's merging back something that you had to do to get the baseline demo based on 14.40 to actually work, I guess ?08:03
radiofreewell i don't actually know, violeta's branch did work when we tested it, but when i was building a branch (not based on master) based off the genivi platform, it didn't work08:04
radiofreethis patch will guarantee that it works though08:04
pedroalvarezthis is an important fix actually. I gess we could do a 14.40.2. To do that this patch should be sent for review and merged in the baserock/release/baserock-14.40  branch. Then tag the HEAD of the branch with "baserock-14.40.2" and rebuild and redeploy the genivi-jetson system. 08:05
ssam2pedroavalrez: do you want me to merge 14.40-fixes or are you doing it?08:08
radiofreepedroalvarez: ok i'll submit it for review08:09
pedroalvarezssam2: please merge them into master 08:11
pedroalvarezssam2: and thanks :)08:11
*** tiagogomes [] has joined #baserock08:16
petefothratmice_______: thanks for the link to Writing great documentation. 08:17
radiofree"Your message to baserock-dev awaits moderator approval"08:19
pedroalvarezradiofree: uh?08:19
radiofreesent from my personal laptop, different e-mail08:20
pedroalvarezI think I can moderate,but don't know how08:23
Kinnisonpedroalvarez: You can if you got a mail telling you how :-)08:23
pedroalvarezKinnison: yeah, but, I can't remember the password08:24
rjekI can moderate it, I think :)08:25
KinnisonLet Pedro do it08:25
*** jonathanmaw [] has joined #baserock08:27
pedroalvarezheheh i can add him to a sender filter, and ban him at the same time :)08:27
Kinnisonbest do that, radiofree is nothing but a troublemaker08:28
* Kinnison ruffles radiofree 08:28
pedroalvarezso, there are 2 options that I don't know what they do: Preserve messages for the site administrator and Forward messages (individually) to:08:29
pedroalvarezwell, what are they for08:29
Kinnisonthe first means that you're preserving the message so that the site admin (me/rjek) can do something about it -- this is something you could do with spam08:29
pedroalvarezaha, I understand08:29
Kinnisonthe second allows you to extract the mails from the mailman queue to a target address without sending them (or in addition to sending them) to the list08:29
pedroalvarezok, makes sense :)08:30
* pedroalvarez moderates08:30
KinnisonMailman's UI is fairly well documented in other places :-)08:30
pedroalvareznot too many buttons to document :)08:30
pedroalvarezradiofree: ?08:33
*** dutch [] has joined #baserock08:34
pedroalvarezradiofree: I'd appreciate some instructions about how to test it08:34
pedroalvarezis it simple as running weston with the right backend?08:35
radiofreepedroalvarez: boot up the genivi image08:41
radiofreeexport XDG_RUNTIME_DIR=/tmp08:42
radiofreeexport EGL_DRIVER=egl_dri2.so08:42
radiofreeif you see weston, it worked!08:42
pedroalvarezradiofree: thanks! :)08:44
*** Krin [] has joined #baserock09:17
ssam2my branch seems to have broken Mason :(09:19
ssam2or rather, Mason has broken. it seems to be a cache server related issue09:20
ssam2'2014-10-24 09:18:52 DEBUG <urlopen error [Errno 104] Connection reset by peer>' when tries to fetch the built libyaml artifact from
pedroalvarezssam2: :) we are debugging the same09:22
Kinnisonconnection reset by peer usually indicates the webserver in front of the cache has hiccoughed09:23
KinnisonI'd poke that09:23
pedroalvarezthis doesn't work even from the mason system itself09:24
pedroalvarezso I'd restart the morph-cache-server09:24
pedroalvarezssam2: btw, do you think that the morph-cache-server-write logs that you are looking at are useful? I added them in that trove for debugging.09:26
ssam2they are definitely useful09:30
ssam2i did 'systemctl stop mason.timer' on the mason to stop it going hugely red09:30
ssam2but yeah, the artifact server can't even ping the mason09:31
pedroalvarezssam2: ping is disabled in our datacentred tenant09:33
ssam2oh, right. :)09:33
pedroalvareztelnet against the 8080 works09:33
Kinnisonand making a request?09:33
Kinnisontry curl09:33
pedroalvarezI tried with wget, and fails09:33
ssam2i'll enable cache server logs on the mason09:34
pedroalvarezeven from the same host09:34
KinnisonI'd restart the cache server frontend then09:34
KinnisonI believe it's fronted by lighttpd09:34
pedroalvarezin troves, but not in the distbuild workers, which is the case now09:36
pedroalvarezso the problem here is when the trove tries to fetch the artifact from the worker09:37
Kinnisonwhat does strace on the worker suggest?09:37
Kinnison(stracing the cache server on the worker)09:37
ssam2interestingly, there are two morph-cache-server processes ...09:40
pedroalvarezssam2: is not that normal?09:40
ssam2not on a distbuild node, I think09:40
Kinnisonparent and worker09:40
Kinnisonif the worker dies, the parent restarts it09:40
Kinnisonit's a thing09:40
ssam2well, one process is waiting on a bottle lock file, the other is stuck in a recv()09:41
ssam2recvfrom() rather09:41
pedroalvarez(bottle lock log:
Kinnisonwhich way around are they?09:41
ssam2the locking one is the parent09:42
Kinnisonsounds right09:43
Kinnisonso the child is stuck in a recvfrom09:43
pedroalvarezradiofree: wrt the "Jetson TK1" email from Gavin in  baserock-dev. Am I right assuming that is not possible to upgrade the kernel of a JetsonTK1 without flashing a new u-boot?09:52
Mode #baserock +v ssam2 by asimov.freenode.net09:53
radiofreepedroalvarez: the kernel will boot fine, just that the gpu won't work09:55
radiofreeand maybe hdmi won't either09:55
ssam2Kinnison, pedroalvarez: are you still debugging the morph-cache-server issue, or should I restart the cache server (which now has logging enabled) ?09:56
radiofreepedroalvarez: nor the pcie, so no networking09:56
* Kinnison has only been debugging by suggestion09:57
radiofreepedroalvarez: i'll reply to him if you want09:57
radiofreei actually have branch which could be used for a devel image09:57
radiofreemeans we can get away with using one version of u-boot, and one kernel09:58
radiofreerdale is currently testing that kernel09:58
radiofreeif it's stable enough to use, i'll submit a patch to start using it in genivi/devel images09:58
ssam2Kinnison, pedroalvarez: ok, I'll restart it and see if we get anywhere09:59
ssam2oh, logs are in the journal10:00
ssam2public mason works again after I restarted its cache server10:44
KinnisonI wonder what caused that10:45
*** dutch [] has quit [Quit: Quit]10:45
*** dutch [] has joined #baserock10:47
pedroalvarezssam2: thanks! 11:03
pedroalvarezradiofree: I think you are the right person to answer that email :) 11:03
pedroalvarezAlthough it seems to me he shouldn't upgrade it 11:05
* petefoth approves of public pledges to fix documentation being a: included in patch review emails and b: honoured / fulfilled :)11:12
straycatIf we had something that could render from branches we could submit docs as patches just like we do with code.11:13
petefothstraycat: I did that once: submitted the patch to  the ML, and was tols, just make the chnage directly in master.11:15
petefothstraycat: I think that is the correct approach because ti reduces friction, and friction discourages people from making the effort to work on docs11:16
petefothHaving some people who care enough about the docs to keep an eye on any changes and revert / change anything that it obvioulsy wrong, will probably work11:17
straycatMy point is that we're not really making good use of git with our wiki, because you have to push to master to see your changes.11:17
petefothThough I think it is worht having a thought-out, discussed and sgreed vision of what should be in the  docs and how they should be structured, so that any changes people do make can go in the correct place11:18
petefothstraycat: or you can edit directly and use the 'Preview' feature11:19
petefothbut yes, a 'rensder some branch' feature would be useful11:19
petefothI recall having something like that on the Mustard instances I worked with11:20
ssam2trying to deploy a system with non-Baserock tools, I've somehow ended up with different, incompatible versions of Django in my production deployment to my development deployment11:48
ssam2if anyone has been thinking 'why bother with Baserock when you can just use 'pip' to install things', here's an answer :)11:48
* straycat thinks ssam2 may possibly be preaching to the converted :p11:51
ssam2it's not too hard to fix, I can just say 'pip install django==1.7' in my deployment script, instead of 'pip install django'11:51
ssam2oh, actually I'm being too hasty here in any case11:58
ssam2seems the problem is in fact that my disk image got corrupted during deploymeny11:58
ssam2-y +t11:58
rdaleradiofree: i have a qt 5.3.2 build on the jetson running12:21
*** genii [~quassel@ubuntu/member/genii] has quit [Ping timeout: 244 seconds]13:08
radiofreerdale: in weston?13:29
rdaleunfortunately the samegame demo doesn't run13:29
rdaleso i was a bit optimistic saying that qt 5.3.2 was running13:30
radiofreerdale: can you try13:30
radiofreegit clone git:// && cd wayland && git checkout 1.6.0 && ./ --prefix=/usr --disable-documentation && make -j6 && make install13:31
radiofreecd && git clone git:// && cd weston && git checkout baserock/james/weston-1.6.0 && ./ --prefix=/usr --disable-libinput-backend --disable-weston-launch13:32
rdalei've got meetings now, will try later13:32
radiofreemake -j6 && make install13:32
rdaleok thanks13:32
radiofreemight need to disable some other backends for weston, can't remember13:32
radiofreethen try it in that weston, could be a problem with the weston-ivi-shell on there13:32
*** genii [~quassel@ubuntu/member/genii] has joined #baserock13:39
*** jjardon28 [] has joined #baserock14:29
*** jjardon28 [] has quit []14:31
*** jjardon_ [sid723@gateway/web/] has joined #baserock14:32
jjardon_Hi, seems is not working?14:34
ssam2seems so. thanks for pointing it out, but it's just a test anyway so I don't know when we'll get around to fixing it14:37
richard_mawperhaps we should remove it from the topic then14:37
ssam2oh, actually I know why it's not working14:37
ssam2I stole its IP address and didn't realise ;)14:37
ssam2we have a couple more IPs available so I'll fix that14:38
rdaleradiofree: i can't get weston to start: 'weston' gives 'cant open display'14:42
ssam2 is working again. thanks Javier :)14:46
radiofreerdale: wayland-backend is to launch weston in weston14:46
radiofreeyou don't need a backend, we use the drm14:46
jjardon_ssam2: cheers! :)14:46
*** paulsherwood [] has joined #baserock14:53
*** fay__ [] has quit [Remote host closed the connection]15:05
*** dutch [] has quit [Ping timeout: 246 seconds]15:08
*** dutch [] has joined #baserock15:21
*** zoli__ [] has joined #baserock15:25
*** zoli_ [~zoli_@linaro/zoli] has quit [Read error: Connection reset by peer]15:25
*** ratmice_______ [] has quit [Ping timeout: 246 seconds]15:33
*** ratmice_______ [] has joined #baserock15:33
radiofreerdale: in your /root/weston.log15:38
radiofreeerm.. can you fpaste it?15:38
*** zoli__ [] has quit [Remote host closed the connection]15:39
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock15:40
radiofreeyou're using a different version of mesa15:44
radiofreedo you have a /usr/lib/libGL* ?15:45
radiofreewhat does you strata/mesa-common/mesa.morph look like?15:45
rdalesurely the same as yours?
*** zoli__ [] has joined #baserock15:50
*** zoli_ [~zoli_@linaro/zoli] has quit [Read error: Connection reset by peer]15:50
radiofreeyeah it's the same, can't imagine it's mesa causing the problem15:51
radiofreemy qt stuff works in weston 1.615:51
*** dutch [] has quit [Disconnected by services]15:52
*** dutch [] has joined #baserock15:52
*** dutch [] has quit [Disconnected by services]15:53
*** dutch__ [] has joined #baserock15:53
rdalei've built without the '-qt-xcb' option, maybe that affects mesa etc in some way15:54
radiofreei wouldn't have thought so15:56
radiofreei'll build that version of mesa and see what happens15:56
radiofreehmm pedroalvarez was going to test my drm fixes :\15:58
radiofreedid he go home?15:58
radiofreeyou're using those fixes right?15:59
rdalei'm not sure15:59
radiofreerdale: cat /baserock/nouveau-drm-bins.meta | grep -A6 system-int16:00
radiofreeif you see two lines removing nouveau.ko and nouveau_platform.ko16:00
radiofreeyou're using it16:00
rdaleyes, i have those16:01
radiofreeis that good enough for +216:04
*** jonathanmaw [] has quit [Quit: Leaving]16:05
*** sam_ [] has quit [Quit: Leaving]16:50
*** dutch__ [] has quit [Quit: Quit]16:51
*** tiagogomes [] has quit [Ping timeout: 244 seconds]17:02
paulsherwoodam i right in thinking that i the dc mason is publishing all its artifacts?17:06
paulsherwoodso i should be able to use --artifact-cache-server= ?17:07
ssam2it's a little more complex, but yes17:10
paulsherwoodhow much more complex? :)17:10
ssam2the Mason uploads its artifacts to
ssam2and you need to pass a base URL, so --artifact-cache-server=
ssam2that should work17:11
paulsherwoodah, ok17:12
ssam2we should set up appropriate DNS and default configuration so that this is a bit easier17:12
paulsherwoodso much magic to know :)17:12
persiaI thought Kinnison created DNS for the DC services.17:12
paulsherwoodyay, it's working :)17:12
ssam2sweet !17:13
persiaIs the mason dual-arch, or single-arch?17:14
*** Krin [] has quit [Remote host closed the connection]17:14
paulsherwoodgreat question! :)17:14
paulsherwoodssam2: ^^17:14
paulsherwoodi know there are a bunch of jetsons sitting near it17:14
ssam2single arch17:15
persiaThat should enable it to be tri- or even quad- arch.17:15
persiaI suppose for multiple architectures, we'd actually need multiple masons, all pushing to the same cache server.17:16
paulsherwoodi don't see why, but maybe17:17
ssam2I'm not sure exactly what'd be required, but it's more than just a config change17:18
ssam2I believe SotK is looking at Mason right now so he's the man to make this happen :)17:19
persiaMy memory is that the way Mason works is that it does stuff locally, then reports, and polls the repo.  This would probably need restructuring to trigger multiple distbuilds, and collate the results.17:19
paulsherwoodno collation needed, i think.?17:20
persiaNo?  Do you not want to see the results for the various architectures?17:21
* paulsherwood pushes his luck, by shooting for a jetsoncloud distbuild17:21
paulsherwoodnot collated?17:21
persiaIf it's not collated, then you are running multiple masons.  If you want one mason to do it, that mason needs to be able to understand architectures, and store the results appropriately.17:22
paulsherwoodmaybe. i haven't even begun to understand the current mason. i just look at  the pretty colours and whine when it's red. i could continue to do that, without collation i think?17:23
persiaYou could, but you'd have to look at one page for each enabled architecture to see if anything turned red.17:25
paulsherwoodhmmm. i'm still not convinced, but i've been wrong before.17:26
persiaI think the right answer is to just add "multiple architectures" to the general wishlist.  There's a lot of different things that folk have wanted out of Mason, most of which seem to be causing it to become a more distributed system, which helps make this easier.17:26
paulsherwoodanyways, anyone know why the jetsons might time out? did someone switch them off?17:27
persiaMaybe what I mean when I say "collation" is different than what you understand.17:27
paulsherwoodi just expect mason fires off a bunch of tasks, reports success or fail17:27
paulsherwoodif it can do them in parallel, fabulous. but all i really care about is green, or an error log if red.17:28
*** violeta_ [] has quit [Ping timeout: 245 seconds]17:29
persiaHrm.  Perhaps dispatching to each architecture in series, and limiting the information to one bit would work, but it might be slow.17:30
paulsherwoodparallel would be fab, as i said. and i'm assuming that caching happens as a happy by-product17:32
*** ssam2 [] has quit [Ping timeout: 246 seconds]17:32
paulsherwoodwhat info do you think mason should be collating? 17:33
persiaIn the event that we want multiple bits of data, we'd need to collate distinct data points and present a more complex dashboard.17:36
persiaPersonally, I'd like to see the results on a per-architecture basis, as it helps distinguish the case of something being just plain bad from the case of something being specific to one or more architectures.17:36
paulsherwoodoh, well if we're heading on to dashboards, there's all kinds of things i'd like to see. but for something to tell the project whether things are currently working or not, i still think plain red/green is more likely to get noticed and acted on :)17:40
persiaI suppose, although if we only have one bit, I'd prefer to get the info faster (single-arch), rather than waiting to run everything.17:41
paulsherwoodmaybe we should wait and see what SotK is actually cooking? :)17:45
* paulsherwood is trying to work out what the pros and cons of gearman and celery are, before starting to reply to the thread hijackers 17:45
persiaI like gearman mostly because zuul was using it.17:46
paulsherwoodwas using it? has that changed?17:46
persiaNot to my knowledge.17:48
persiaOr perhaps without the "gating" :)17:49
jjardon_Hi, trying to clone a repo from github Im getting this: (works fine in my laptop) any ideas why is not working?17:52
paulsherwoodin a baserock system?17:52
jjardon_baserock chroot, yes17:52
paulsherwoodca certs were added recently... if your br system predates that, it has not certification....17:53
franredjjardon, you don't have the latest devel system which includes the ca-certificates17:53
paulsherwoodyou could git config http.sslVerify "false" maybe17:53
jjardon_will try that, thanks!17:54
franredor try to change the http[s] by git17:54
franredshould work17:54
paulsherwoodyes, that works too17:54
jjardon_can I ask here to lorry libepoxy?
paulsherwoodyes... why do we need it? :)17:56
jjardon_paulsherwood: gtk3! 17:57
jjardon_and probably more important: xserver17:57
jjardon_(for xwayland)17:58
*** franred [] has quit [Quit: Leaving]17:58
jjardon_gtk3 just added opengl support so it can be used to easily test if 3d acceleration is working correctly18:00
jjardon_rjek: readline is used for nmcli: a command line utility to control NetworkManager18:04
jjardon_Was there any agreegement about the best way to allow the upgrade of readline?18:06
* paulsherwood has no clue18:08
jjardon_paulsherwood: thanks! :) only a little note: its needed by "xwayland", ie, the module needed to run X apps on top of wayland18:10
paulsherwoodjjardon_: i've submitted the lorry. now lc seems stuck. perhaps it's me, again18:11
paulsherwoodang gbo maintainers able to check what lc thinks is going on?18:37
paulsherwoodwell, it figured things out...19:03
jjardon_\o/ thanks!19:05
paulsherwoodany idea how i could extract the rootfs from a baserock img file?19:21
*** zoli__ [] has quit [Remote host closed the connection]19:31
Kinnisonmount it?19:35
paulsherwoodi thought i'd tried that and failed... will repeat19:36
paulsherwoodin any case, i found an alternative hack.19:36
paulsherwoodthanks though :-)19:37
* paulsherwood does wonder, for the nth time, why anyone ever objects to tagging morph at the same time as a release19:41
rjekjjardon_: Patch it so it doesn't need readline, or the latest version of readline?19:45
*** zoli_ [] has joined #baserock20:05
*** zoli_ [] has quit [Changing host]20:05
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock20:05
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]20:07
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock20:24
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]20:48
*** zoli_ [] has joined #baserock21:12
*** zoli_ [] has quit [Changing host]21:12
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock21:12
*** persia_ [quassel@2400:8900::f03c:91ff:feae:3452] has joined #baserock21:20
*** persia_ [quassel@2400:8900::f03c:91ff:feae:3452] has quit [Changing host]21:20
*** persia_ [quassel@ubuntu/member/persia] has joined #baserock21:20
*** persia [quassel@ubuntu/member/persia] has quit [Ping timeout: 272 seconds]21:20
*** zoli__ [] has joined #baserock21:28
*** tiagogomes [] has joined #baserock21:28
*** tiagogomes [] has quit [Client Quit]21:28
*** zoli_ [~zoli_@linaro/zoli] has quit [Ping timeout: 258 seconds]21:32
paulsherwoodalthough i like wikicat, i'm delighted it's not here to see my shambolic editing tonight22:07
pedroalvarezpaulsherwood: the jetsoncloud is only usable now from our datacentred tenant. I can't debug anything this weekend, sorry 22:09
paulsherwoodah. i forgot about that.22:13
*** genii [~quassel@ubuntu/member/genii] has quit [Read error: Connection reset by peer]22:31

Generated by 2.14.0 by Marius Gedminas - find it at!