[00:00] woah what? [00:00] yakkety_unstable_kdeplasma-addons build successfully [00:00] yofel: can you see what's going on? [00:01] tsimonq2: well, I have shell access and a monitoring attached to the CI, but I can't see anything obviously wrong [00:01] yofel: so then rebuild? :) [00:02] I triggered what you said [00:02] :) [00:03] well, that's a big build queue... [00:03] yofel: did you do a global rebuild? just curious [00:03] oh, the daily job for mgmt_merger just triggered [00:03] oh okay [00:03] yofel: so this isn't all to Kubuntu packaging right? [00:04] from what you have seen, I can do this for hours on end [00:04] but what else is there to do? [00:04] tsimonq2: no, that's the CI part. Which does help with the release packaging, but that is a seperate workflow [00:04] yofel: so what else can I help with? :) [00:06] good question right now, clivejo would probably have a better answer for that. Lets see [00:08] hm, the ci isn't really triggering any builds. Now I wonder whether I either fixed it earlier, or broke it... [00:09] well, it did trigger breeze-gtk because of an SCM change, so I guess it's working.. [00:11] yofel: what happens to these CI builds next? [00:11] they go on, until a new upstream release happens. At which the kubuntu-automation tooling that generates the release packages goes and merges kubuntu_unstable into kubuntu_yakkety_archive. [00:12] where's that? [00:12] that way getting the new release done is very little work, as long as the CI packages are in a good shape [00:12] tsimonq2: https://git.launchpad.net/~kubuntu-packagers/+git/kubuntu-automation [00:12] currently being reworked, the README is for the old and still used scripts [00:12] yeah [00:13] yofel: what about my two outstanding MPs? :D [00:13] sec [00:13] k [00:17] done [00:17] and the merger is done too [00:17] aaaaaaaaaaand it triggered 440 builds [00:18] sitter: Is it really supposed to do that? ^ [00:22] yofel: thanks :) [00:25] # docker --version [00:25] Docker version 1.6.2, build 7c8fca2 [00:25] hm, I wonder if we can update that... [00:34] yofel: if there is a file listed in the install file that is no longer present, do you remove it? [00:35] right? [00:35] yes, otherwise the package won't build [00:36] alright [00:49] well I'm off to bed, thanks for all of your help yofel [00:49] o/ [00:50] tsimonq2: nini, thanks to you for fixing so much! [01:09] Hi All, please see error generated by Steam https://goo.gl/FxAncA PLease help me resolve this. [01:22] leumas: #kubuntu for support please [01:28] wheee [01:28] figured out how to run a vanilla kernel on linode. I have a working docker in lxd \o/ [01:35] nice! [01:35] will you use that for builds, or what? [01:41] I just got the email about that boot [01:46] valorie: trying to put another jenkins executor on linode [01:46] nice! [01:46] now I have to figure out how to set *that* up :P [01:47] I'm assuming you never sleep [01:47] well, soon I will. I am getting sleepy after all [01:48] news from Munich is so bad [01:48] and sad [01:48] indeed [02:02] ok, jenkins can now talk to linode... [02:24] jenkins talks to the lxc? [02:24] it talks to the container over ssh [02:24] I already got it to do *something* http://kci.pangea.pub/job/xenial_unstable_modemmanager-qt/190/console [02:25] now the problem is how to deploy the ci-tooling to the slave [02:25] neato [02:26] and that not all jobs that should run on all slaves are matrix jobs :/ [02:28] although, the docker cleanup can just be a cronjob.. [02:35] and there goes LP again.. [02:39] * mamarley kicks LP. [06:24] tsimonq2: if you look at the failed plasma-desktop build timestamp in it's name, it predates the packaging change I made to fix the appstream data. so the sources were generated pre my change, but it just got queue for a stupidly long time [06:25] and it's successor build with the change has never quite got to build yet. was in the queue when the queue was cleared yesterday I thing [09:44] good morning everyone [09:49] moin [09:49] morning :) [10:31] offtopic: just got into lxc containers and this seems awesome [10:36] * yofel spent half of the night making docker work in priviledged lxd containers [10:36] lxc is great though, and I recommend reading https://www.stgraber.org/2016/03/11/lxd-2-0-blog-post-series-012/ for lxd [10:37] are you planning to build stuff on the new Linode server? [10:38] I want a kci slave on linode [10:38] the slave works, docker too now. But now I don't get how to deploy the docker containers for the ci [10:38] rake can't find some ruby modules for some reason [10:40] I am investigating lxc to migrate my ppa simulation invention from libvirt on top of kvm [10:40] right now I'm feeling better working with it directly, without lxd [10:41] docker seems it wouldn't be appropiate as I need persistent storage [10:41] right, if you want system containers you want lxc. And lxc does give you a bit more flexibility than lxd [10:41] KCI seems to be stalled waiting for that linode docker container, which is offline? [10:42] o.O [10:43] lets see if I can abort that [10:43] s/docker/lxc or whatever [10:43] * yofel didn't think it would wait on the offline slave [10:44] that got it moving again :) [10:49] clivejo: Is apps staging for xenial meant to have Qt 5.6.1 in it? [10:50] no :( [10:50] I forgot that was there when I copied frameworks [10:50] it want to remove my plasma, as I'm still on 5.6.5 on this main machine [10:51] ah.. right [10:51] eakk [10:51] * acheronuk disables apps staging for now [10:54] what is it trying to remove? [10:55] * clivejo shouldnt have sat up chatting to Maro so late last night [10:55] I feel so hungover today [10:56] clivejo: http://paste.ubuntu.com/20584915/ [10:56] wow [10:57] with all the changes with apps (missing and disabled packages) maybe I should klear the staging-apps PPA [10:58] and redo them [10:58] how does this work Yofel? [10:59] On YY Im on Qt5.6.1, Plasma 5.7.2 and having problem will apps not being Qt5.6.1 [11:00] lol. back to front then [11:00] plasma-discover seems to be broken [11:00] I posted about discover yesterday [11:00] you might as well [11:00] really I need YY Apps to be built on Qt5.6.1 [11:00] as long as all fixes are in git, that should be easy [11:00] are the package lists fixed now? [11:01] yes [11:01] ie a git-do-all should now see them [11:01] git-clone-all [11:02] but it doesn't use the static package names lists [11:02] clivejo: one prob is plasma-discover 5.7.2 depends packagekit >= 1.0 [11:02] and some obsolete plasma-discover 5.6.5 packages maybe [11:03] http://paste.ubuntu.com/20467363/ [11:03] and http://paste.ubuntu.com/20468770/ [11:03] what I posted yesterday ^^^ [11:06] why does muon have an epoch [11:08] whats building that package [11:08] muon is from muon [11:08] and it had one for a long time [11:09] is muon being released again? [11:09] huh? weren't you one of the people that fixed it last cycle? [11:09] yes, but it wasnt released [11:09] muon-updater (4:5.7.2-0ubuntu1~ubuntu16.10~ppa1) [11:10] that package came from staging [11:11] I worked on the Muon Package Manager [11:12] that is not from muon, but a transitional package from discover [11:14] ah yes [11:15] muon-discover depends on plasma-discover-updater [11:16] but dont see that being created by that control file [11:18] my head is just not in the zone today [11:19] Neon has a plasma-discover-updater.install while Debian doesnt [11:19] That YY VM had been updated multiple times, so I can't guarantee there wasn't some left over oddities on it [11:19] can try a clean one leter maybe [11:20] my recent merge with debian removed those packages [11:20] https://git.launchpad.net/~kubuntu-packagers/kubuntu-packaging/+git/discover/commit/?h=kubuntu_yakkety_archive&id=29d2a111eb591d7bb5c5bc7954223b13f8eb7d70 [11:24] urgh, now jenkins has a hardcoded apt-cache host [11:26] 11:24:50 gpg: keyblock resource `/var/lib/jenkins/tooling/gnupg/secring.gpg': file open error [11:26] oops [11:52] hiho [11:57] 11:54:42 D: Writing logfile: /var/lib/jenkins/workspaces/yakkety/unstable/gwenview/build/gwenview_16.04.3+p16.10+git20160723.1156-0_source.ppa.upload [11:57] 11:54:42 Successfully uploaded packages. [11:57] FINALLY [11:58] to archive ? [11:59] soee: I'm setting up another build executor for KCI, but the tooling is really not intended for that...... [12:03] on the linode vps ? [12:04] yes [12:05] do we have already some running ? [12:05] soee: "some" ? [12:08] the linode is running, and has an LXD setup mostly meant for packaging setup on it if people need a server to work on [12:08] I now put a jenkins slave in an lxd container http://kci.pangea.pub/computer/linode-01/ [12:08] which... took a while... [12:10] ah i thought we were talking about building instances not for users but kci [12:10] so the one that grabs stuff from git tries to build it etc. [12:10] ah no, it was primarily meant for users, but the server has plenty of resources so it could be used for that [12:11] unused resources are lost resources ;D [12:11] that was what harald wanted to do back when he first set the ci up [12:11] we shudl use all we can [12:11] yes, esp. as that linode is actually somewhat expensive if we had to pay for it ^^ [12:11] and hopefully won't run out of memory? ;) [12:11] kci has 4G, linode 24G [12:11] if users want to build something and need more resources some kce instancs could be pasued than [12:11] a "bit" more :P [12:12] but the memory issues is a docker bug I believe [12:12] you find verious bug reports if you look for it [12:12] hmm [12:12] yofel: yes, I was jesting [12:12] but I don't want to upgrade docker 1.6 to 1.11 without having a fallback slave [12:12] tried: http://kubernetes.io ? [12:13] the linode slave runs 1.11, so we'll see [12:14] no, I know that harald or rohan had played with docker swarm, but the tooling requires far too much handwork in the setup currently [12:14] so fixing that comes first [12:14] someone also needs to look at the changes he did for NCI [12:15] NCI even has sound effects :P [12:16] * acheronuk runs [12:16] sounds like harald alright [12:24] howdy all [12:24] heyho [12:25] YES, the slave built gwenview [12:25] lets add some more executors [12:27] hm, might as well overdo it while I'm at it [12:30] Cannot add PPA: 'ppa:~kubuntu-ci/ubuntu/unstable'. [12:30] back to front? [12:31] clivejo: where did you see that? [12:31] in the console log [12:31] ppa:kubuntu-ci/unstable [12:31] clivejo: of *what*? [12:31] oh. right [12:32] http://kci.pangea.pub/job/yakkety_unstable_kshisen/53/console [12:32] 12:25:37 [12:32] 12:27:04 make: *** Cannot allocate memory. Stop. [12:32] well, master [12:32] * yofel set 32 executors on linode-01 [12:33] lets see how that works out:D [12:33] you trying to make it go boom?!? [12:33] 32 :D [12:33] stress test... [12:34] I thought the executors had to be set to stop Jenkins crashing [12:34] you might kill launchpad [12:34] uhm, according to dstat and zabbix, that uses barely any resources, a bunch of disk IO and +1.2G memory [12:34] wow [12:36] uhm, ok. I did kill it [12:37] btrfs that is -.- [12:37] * yofel does desaster recovery [12:37] LOL [12:37] * clivejo thinks yofel is having fun! [12:38] :P [12:38] uh oh [12:38] [39693.784549] BUG: unable to handle kernel paging request at fffffffffffffe58 [12:38] [39693.784557] IP: [] __btrfs_unlink_inode+0x1c/0x470 [btrfs] [12:38] grr [12:38] have to ping BS? [12:39] lets see if I can trigger a hard reset [12:44] what did you kill the linode or KCI machine? [12:45] linode [12:45] oh [12:45] I guess btrfs couldn't handle the kernel 4.5 -> 4.4 downgrade [12:46] because I'm running it in other places under heavy stress in 4.4 and it runs fine [12:47] acheronuk: sorry for cancelling you project yesterday [12:48] ok, linode back up [12:48] lets see [12:48] clivejo: no problem at all. I'm in no hurry. I was more pleased to see Simon getting stuck in [12:49] well that was the intention behind it [12:49] its nice to see things happen! [12:50] If he's back later, perhaps he can look at https://launchpadlibrarian.net/274580284/buildlog_ubuntu-yakkety-amd64.kwidgetsaddons_5.24.0+p16.10+git20160723.0238-0_BUILDING.txt.gz [12:50] hm, btrfsck doesn't see anything, but I don't trust that fs structure anymore :S [12:50] did all his merge reqests get pushed? [12:51] or hm, lets try again [12:51] clivejo: think so. yofel was doing them v late [12:51] how is he so perky today! [12:52] I feel like I could sleep for a week [12:52] I felt like that earlier in the week. Now it's cooler I'm better [12:53] when I'm angry I don't need sleep :P [12:53] and jenkins makes you angry [12:53] do you drink coffee? [12:54] me or yfl? [12:55] I think I've now become mostly immune to caffeine. Doesn't help like it used to.... [12:56] yofel [12:56] I dont drink it [12:56] but thinking I should start! [12:56] at work I do [12:56] Im in a strange mood today [13:06] ^.^ [13:40] thats one way to clear a queue [13:41] yofel: the container you setup for packaging, is it shared for all the packagers? [13:42] it's shared for whoever has his key in the ubuntu account [13:42] so for access to depot who ssh key would it use? [13:42] you can either use one for all, or make a container for each person [13:43] none, it doesn't have access right now [13:43] as for that, the container needs its own ssh key [13:43] I dont fancy putting my keys on it :/ [13:43] just make a password less key in the container and add that to depot [13:43] just make sure to remind people to not take that key out of there [13:44] what about git access to LP? [13:44] there is no push access unless someone decides whose user will be the proxy [13:45] that's why I personally recommend one container per person [13:45] and you can invite people to use yours as needed [13:46] can we use one of the kubuntu accounts as a proxy? [13:46] ie create a key for kubuntu-packagers [13:46] no, please don't do that [13:46] rather create a new user and add the key to that [13:47] well, packagers is a team, so you couldn't add a key to that anyway [13:47] how would you do it? [13:48] the kubuntu-ci seems to be an actual account [13:48] use my account, add a new key to it, give people access to my container while i'm watching them, then kick them out when I have to leave [13:50] yellow belts is a team too? [13:50] yes [13:50] yellow belts should not be member of packagers anyway [13:51] would setting up abrand new account, something like kubuntu-training or something work? [13:52] well, feel free to, as long as it doesn't give people unreviewed committ access to kubuntu-packagers [13:53] hummm [14:36] yofel: is my key still on dev.kubuntu.co.uk ? [14:37] cant log in [14:37] how are you trying to? [14:37] ssh dev.kubuntu.co.uk [14:38] hm, it seems to be there [14:38] Host dev.kubuntu.co.uk [14:38] User ubuntu [14:38] IdentityFile ~/.ssh/id_clivejo [14:38] what's the error? [14:38] oh, wrong user [14:39] whats the username? [14:39] 'clivejo' on port 22 will give you the host. 'ubuntu' on a different port will give you a container [14:39] although, it seems one of the port mappings got lost in the reboot :/ [14:40] just want a container [14:40] clivejo: try port 2201 [14:41] in [14:49] clivejo: sorry, jenkins killed the server again [14:49] linode KCI node just fell over? [14:49] oh [14:49] I figured! [14:49] I think I need to recreate that FS :( [14:49] Jul 23 14:47:14 ubuntu kernel: BTRFS: error (device sdc1) in btrfs_rename:9374: errno=-2 No such entry [14:49] Jul 23 14:47:14 ubuntu kernel: BTRFS info (device sdc1): forced readonly [14:51] I'll try to do a quick reboot to get your container back up and leave jenkins down [14:51] then I'll be gone for a while [14:51] * clivejo cries [14:51] my poor container [14:51] you killed it! [14:53] how do I get my terminal back :/ [14:53] packet_write_wait: Connection to 139.162.164.207 port 2201: Broken pipe [14:53] ah [14:53] clivejo: should work again [14:54] as long as you don't do jenkins-level crazy things nothing bad should happen [14:54] I'll recreate the setup on a clean filesystem in the evening [14:54] bbl [14:56] ok [14:56] have fun [15:00] :( [15:01] git-clone-all uses depot :( === pavlushka is now known as Guest29209 [16:47] lots of action today yofel lol === Guest29209 is now known as pavlushka [16:58] this saturday is sponsored by letter R like Red :-) [17:05] plasma-desktop is about to go GREEN [17:19] \o/ [17:28] o/ [17:28] acheronuk: ahh I see [17:29] yofel: no problem, it's nice to learn this stuff :) [17:31] I'll fix something before I go [17:36] grrrr [17:38] yofel: the failure on xenial_unstable_dolphin is an internal error, not related to the package [17:38] o/ for now [17:39] lot of internal errors at the moment [17:41] yofel: is it posible to run the staging script again? [18:28] re [18:29] clivejo: staging script is unaffected by what I do. Regarding the deleted packages: try it [18:30] if you don't see them as Deleted on the delete page anymore you're fine [18:30] clivejo: do you need linode? I would like to take that offline [18:30] nope, go ahead [18:30] and I think I should be reading up on lvm... [18:33] hmmm... wiki won't let me create a page [18:34] Yea the wiki is kinda broken for login [18:34] Well it works at times [18:34] wiki modification is also restricted to specific teams [18:34] ahoneybun: I'm logged in from the ubunutu side [18:35] Well pagea [18:35] Not making one [18:35] the wiki sucks! [18:35] Well [18:35] ahoneybun: "You are not allowed to copy this page!" [18:35] Copy mine [18:35] last I heard was: The wikis are locked down to only two teams: ~ubuntu-members and a [18:35] Canonical team. [18:35] yofel: great! [18:36] just as well I'm not in a huge hurry to do it! [18:36] ah, there's also ~ubuntu-wiki-editors for non-members [18:37] yofel: so I could join that with the justification of creating a profile page? [18:39] OSM have a thriving mediawiki install ( http://wiki.openstreetmap.org/wiki/Main_Page ) why on earth cant Ubuntu ? [18:39] Oh [18:39] probably...? popey is the team owner, so maybe ask him in #ubuntu-devel? [18:41] yofel: may do that then, they should at least know how to sort it, even if I can't join the team [18:41] right [18:42] but after the weekend. no rush on this. I just had a urge to start doing it, and was thwarted [18:43] oh, I know that feeling combined with the wiki... [18:43] did valorie bribe you with cookies? [18:43] I didn't get cookies when I applied :( [18:44] I want ginger biscuits [18:44] I still havent got them! [18:44] I feel cheated! [18:50] yofel: should apps 16.04.3 be built on Qt5.6.1? [18:51] for yakkety yes, for xenial... probably too as we're almost done with plasma 5.7 [19:02] why wont git-clone-all grab libkgeomap ? [19:04] dunno [19:05] did you map it? [19:06] git-clone-all doesn't use our maps, but creates a package list directly from depot [19:06] see kubuntu-automation/.cache/ [19:07] * clivejo beats head off the table [19:18] I give up! [19:33] :) [19:48] ok, so lets try to kill the server again... [19:53] kill it, kill it [20:03] :O [20:04] better, I was reconfiguring the docker bridge [20:04] because our ci builders for some reason expect apt-cacher-ng to be available at 10.0.3.1:3142 [20:05] ok, that built, so let hell break loose [20:06] hm..... [20:06] well that didn't go too well [20:07] .... [20:08] :S [20:08] that must have hurt xD [20:08] I never set up RVM [20:08] now how do I do that... [20:13] apt install rvm [20:17] that's not a thing [20:17] but I got that installed [20:17] but now it's failing to actually install the correct rvm ruby *-.- [20:17] is it not ?!? *whistles* [20:19] I'm somewhat tempted to just tar up the freakin' rvm env from river and copy it to linode [20:20] or let me try something else [20:20] whats river? [20:21] kci hostname [20:21] sry [20:21] ok, using root seems to work [20:22] oh great, now it's compiling ruby...... [20:25] Error loading RubyGems plugin "/home/jenkins-slave/.rvm/gems/ruby-2.2.0@merger/gems/gem-wrappers-1.2.7/lib/rubygems_plugin.rb": cannot load such file -- gem-wrappers (LoadError) [20:25] what?!? [20:54] finally... [21:03]