IRC logs for #baserock for Monday, 2015-05-11

*** zoli__ has joined #baserock01:05
*** zoli__ has quit IRC01:10
*** zoli__ has joined #baserock01:21
*** zoli__ has quit IRC02:49
*** zoli__ has joined #baserock04:46
*** zoli__ has quit IRC05:03
*** zoli__ has joined #baserock05:06
*** zoli__ has quit IRC06:06
*** zoli__ has joined #baserock06:23
*** sherm_ has joined #baserock06:23
*** petefoth has joined #baserock06:27
*** a1exhughe5 has joined #baserock07:18
*** Albert_ has joined #baserock07:45
*** Albert has joined #baserock07:46
*** sambishop has joined #baserock07:55
*** bashrc has joined #baserock08:04
*** jonathanmaw has joined #baserock08:26
*** sherm_ has quit IRC08:28
*** sherm_ has joined #baserock08:41
*** gary_perkins has joined #baserock08:47
*** CTtpollard has joined #baserock09:03
*** ssam2 has joined #baserock09:05
*** ChanServ sets mode: +v ssam209:05
*** tiagogomes has joined #baserock09:05
*** franred has joined #baserock09:07
*** edcragg has joined #baserock09:08
*** CTtpollard has quit IRC09:22
*** lachlanmackenzie has joined #baserock09:24
*** CTtpollard has joined #baserock09:25
*** Krin has joined #baserock09:25
radiofreelorry request http://paste.fedoraproject.org/220575/3651614309:28
* richard_maw prefers git:// URLs for projects that are publically available, but understands that others prefer https09:29
ssam2I think we discussed "Pelagicore-LaTeX-Class" before with jonathanmaw09:30
ssam2the conclusion then was to try and avoid it, did that turn out not to be possible?09:30
jonathanmawssam2: given we'll never use it, that was the decision09:30
radiofreeit's a submodule http://git.baserock.org/cgi-bin/cgit.cgi/delta/genivi/browser-poc.git/tree/.gitmodules?h=baserock/jonathanmaw/sensible-build09:31
radiofreeif we're happy to pull from random github in the sky, then we'll need to at least change that submodule to use https09:31
jonathanmawand I think we'd have to carry a patch regardless of whether we disable it, or integrate the submodule into baserock09:31
pedroalvarezjonathanmaw: good point09:33
ssam2seems that distbuild in the version of Morph used in Baserock 15.19 is broken09:37
ssam2my fault09:37
radiofreeso what's the decision then? we *are* happy to pull from github?09:37
ssam2radiofree: no, if it's not possible to build without the submodule then +1 to the lorry request09:38
richard_mawradiofree: AIUI we patch out the submodule09:38
ssam2but if it's just as easy to patch out the submodule then it seems better to do that, than lorry a tiny, single-purpose repo09:38
radiofreejonathanmaw: is that easy?09:39
jonathanmawradiofree: `git rm .gitmodules && git commit -a`?09:40
radiofreejonathanmaw: oh right, you have to build the documentation manually anyway?09:40
jonathanmawradiofree: yep09:40
jonathanmawand it's pdf, so it's not really suitable for installing into a baserock system09:40
radiofreeok, do that then09:40
franredwhich Juno release does include the latest baserock release? Openstack Juno 2014.2.1 or 2014.2.3?09:48
pedroalvarezfranred: that is something you can check :)09:48
ssam2oh, I wrote 2014.2.1 in the release notes but I may have got that wrong09:48
*** zoli__ has quit IRC10:08
Zaraoh, so if you want to reply to comments in gerrit, you don't just have to click 'reply' and 'done' inline, you then have to click 'reply...' and 'post' on the main page for your patch... gah.10:11
franredssam2, it is right, baserock-15.19 includes Juno 2014.2.1. Juno 2014.2.3 was merged on friday so it didn't get into this release10:45
KinnisonZara: yeah the first step is draft replies10:45
ssam2franred: cool, thanks for checking!10:45
pedroalvarezZara: if you click in "Done" I believe gerrit just writes "Done" in the reply10:51
ZaraKinnison: I wish the UI made it clearer. I thought I'd posted replies on Friday, realised this morning that that was not the case.10:51
Zarapedroalvarez: heh, you're right :)10:52
KinnisonZara: :(10:54
pedroalvarezperryl: thanks for fixing distbuild :)10:56
perryl:)10:57
ssam2sadly SotK has discovered another bug10:59
ssam2I'm going to send this to baserock-dev and baserock-announce, does that seem like the right think to say? http://paste.baserock.org/inefinoqis10:59
ssam2I guess I can merge https://gerrit.baserock.org/#/c/616 first10:59
franredssam2, sounds good to me both11:02
ssam2I should have used a better subject than 'Re: Baserock 15.19 is released!' though, I now realise! oh well11:09
ssam2it would be good to have a Mason that tracks 'master' of morph.git after every commit, to stop this sort of silliness in future...11:10
ssam2it'd be easy to implement that in a really hacky way (have a cron job that runs 'git pull' every minute), I'll hopefully have a look at that this afternoon11:11
*** Albert has quit IRC11:12
*** Albert has joined #baserock11:12
pedroalvarezssam2: that would be great11:13
pedroalvarezalthough there are some implications to think about11:13
ssam2such as?11:16
ssam2I agree in fact, it will be far from perfect11:18
*** zoli__ has joined #baserock11:19
jjardonHi straycat , If you have some free time, I commented in https://gerrit.baserock.org/#/c/580/11:31
perrylSotK: i'm looking at https://gerrit.baserock.org/#/c/605/2, is the connection_id for the Initiator class only set after the build has started? that is, connection_id has no value during the build graphing stage?11:50
*** zoli__ has quit IRC11:52
*** zoli__ has joined #baserock11:52
*** zoli___ has joined #baserock11:53
*** zoli__ has quit IRC11:53
SotKperryl: good point, I didn't think about cancelling before the build graph was done12:00
perrylSotK: i've just sent a review via gerrit12:01
pedroalvarezI'm currently thinking about OpenStack on baserock and upgrades, and it seems to me that it should just work as long as we mantain a set of migration scripts for whatever changes we do in the setup services of openstack.12:09
pedroalvarezBut, that sounds too easy, so I wonder if I'm missing something.. :)12:10
ssam2I don't know if migration scripts are that easy to maintain12:12
ssam2so that's why it sounds too easy12:12
petefothTesting migration scripts could be tedious / hard12:12
petefothAnd if they're *not* tested...12:13
pedroalvarezssam2 reminded me abuut rollbacks, that is not going to be easy if we want to rollback to the version before the upgrade12:23
SotKOK, I'll base my patch on the GraphProgress patch that straycat sent, since prior to that the start of build-graph calculation is signalled by a build-progress message12:23
*** zoli___ has quit IRC12:24
*** zoli__ has joined #baserock12:24
*** inara has quit IRC12:56
*** inara has joined #baserock12:56
richard_mawCan I persuade anyone to review https://gerrit.baserock.org/#/c/597/5 ? I'd like to have another pair of eyes on it12:59
* SotK wonders how he missed reviewing that patch13:03
richard_mawmany things are explainable as a consequnce of the Gerrit UI13:05
SotKI think I just forgot to vote, I've looked through the code a few times already13:07
richard_maw:¬)13:07
richard_mawstraycat: care to vote on https://gerrit.baserock.org/#/c/599/ or do you require changes to cmd_update_stratum_chunk_ref?13:10
straycatOh sorry13:12
richard_mawstraycat: no worries13:12
straycatjjardon, biff13:20
*** fay_ has quit IRC13:48
*** fay_ has joined #baserock13:48
jonathanmawhrm, the repos in open-source-lorries/genivi.lorry get confusingly-named. for some reason it translates git.projects.genivi.org/lbs/navigation to genivi/poi-service, and git.projects.genivi.org/lbs/navigation-application to genivi/navigation13:55
jonathanmawunfortunately, I don't see any way of changing this that won't confuse things that existed previously13:56
*** ssam2 has quit IRC14:06
*** petefoth has quit IRC14:15
*** petefoth has joined #baserock14:16
*** ssam2 has joined #baserock14:20
*** ChanServ sets mode: +v ssam214:20
*** zoli__ has quit IRC15:01
* richard_maw giggles with glee as he writes a /run/initramfs/shutdown which allows him to `exec /init` and start the system back up again15:18
pedroalvarezrestarting the system without restarting?15:22
richard_mawthat's one way to put it15:28
pedroalvarez:)15:29
pedroalvarezI'm curious now about what are we going to be able to do with that15:30
richard_mawre-execing /init was just me having a bit of fun15:30
richard_mawthe proper use for this is to be able to unmount /15:31
richard_mawcleanly15:31
richard_mawas it's potentially a source of filesystem corruption15:31
rjekCan you not mount -o remount,ro ?15:33
richard_mawwe can, but worryingly I see no evidence that systemd attempts this15:34
*** zoli__ has joined #baserock15:35
richard_mawit normally goes on an unmounting spree, but it can't unmount /, as that's what's being used15:35
*** ssam2 has quit IRC15:35
rjekoops15:36
*** ssam2 has joined #baserock15:48
*** ChanServ sets mode: +v ssam215:48
richard_mawyep, no sign of a remount,ro in the main reboot code, but it might be in .mount unit handling15:54
*** a1exhughe5 has quit IRC16:02
richard_mawI can't eliminate the possibility that systemd changes the mount options of the .mount unit and reloads its configuration (this triggers a remount with the new options, which may include "ro"), but I saw no evidence of it in the shutdown code path16:07
*** gary_perkins has quit IRC16:12
richard_mawrjek: at this point I can only guess that it's expected that if you need to remount ro you put an executable in /lib/systemd/system-shutdown to do it16:17
rjekNod16:19
richard_mawthough I'm perplexed by http://freedesktop.org/wiki/Software/systemd/Debugging/#index2h1 which suggests / is mounted ro at some point16:20
richard_mawoh, and evidently it *does* remount / as ro somewhere16:21
richard_mawthe shutdownramfs has / mounted ro16:21
richard_mawah, I found it: http://cgit.freedesktop.org/systemd/systemd/tree/src/core/umount.c#n38816:28
richard_mawcrisis averted16:29
*** wdutch has quit IRC16:29
*** wdutch has joined #baserock16:30
*** jonathanmaw has quit IRC16:35
*** ssam2 has quit IRC16:44
*** Albert has quit IRC17:03
*** Albert has joined #baserock17:03
* richard_maw wrote https://gerrit.baserock.org/#/c/624 and https://gerrit.baserock.org/#/c/62517:04
*** bashrc has quit IRC17:04
richard_mawthey are the beginning of shutdownramfs support, after the initramfs-scripts change is merged I can produce a definitions.git change to include them in deployments17:04
*** nowster has quit IRC17:29
*** sherm_ has quit IRC17:33
*** tiagogomes has quit IRC17:36
*** lachlanmackenzie has quit IRC17:37
*** nowster has joined #baserock17:39
*** zoli__ has quit IRC17:44
*** zoli__ has joined #baserock17:45
*** zoli__ has quit IRC17:48
Kinnisonnice17:48
*** edcragg has quit IRC17:51
*** zoli__ has joined #baserock18:06
*** Stanto has quit IRC18:25
*** SotK has quit IRC18:25
*** rjek has quit IRC18:26
*** Stanto has joined #baserock18:27
*** SotK has joined #baserock18:27
*** rjek has joined #baserock18:27
*** rjek has joined #baserock18:27
*** zoli__ has quit IRC18:35
*** zoli__ has joined #baserock18:38
*** zoli__ has quit IRC20:02
KinnisonHmm, mercurial security update in Debian -- I wonder if it applies to troves20:07
*** gary_perkins has joined #baserock20:47

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!