IRC logs for #baserock for Thursday, 2016-03-17

*** gtristan has quit IRC01:25
*** gtristan has joined #baserock01:46
*** lachlanmackenzie has quit IRC02:19
*** thrace_ has quit IRC02:19
*** lachlanmackenzie has joined #baserock02:25
*** thrace_ has joined #baserock02:25
*** gtristan has quit IRC06:09
*** gtristan has joined #baserock06:44
*** lc_ has joined #baserock07:07
*** locallycompact has joined #baserock07:07
*** edcragg has joined #baserock07:42
*** toscalix has joined #baserock07:58
*** edcragg has quit IRC08:25
*** fay_ has joined #baserock08:48
gtristanrobtaylor, decent news... virtio-net, when _actually_ compiled into the kernel and used... brings the glibc compile time down from 2hours and 40mins... to just 40min08:59
pedroalvarezgood news then! :)09:03
*** bashrc has joined #baserock09:04
*** edcragg has joined #baserock09:09
*** CTtpollard has quit IRC09:11
*** CTtpollard has joined #baserock09:16
*** jonathanmaw has joined #baserock09:18
*** locallycompact has quit IRC09:19
*** lc_ has quit IRC09:38
robtaylorgtristan: awesome!10:00
gtristansystemd, I havent re-tried yet, it's horribly slow10:01
gtristanI dont understand why either, they arent writing C++ are they ?10:01
gtristanI mean, I havent retried the *whole build*, but started a build of systemd and observed that it is still horribly slow10:02
paulsherwoodgtristan: iirc at one point systemd builds got slower because of building tests... are you doing that by anychance? i may be wrong10:03
robtaylorgtristan: do you mean that each individual cc is slow?10:05
*** tiagogomes has joined #baserock10:07
*** ssam2 has joined #baserock10:10
*** ChanServ sets mode: +v ssam210:10
gtristanrobtaylor, I do, I mean each individual CC takes a loooong time :)10:29
gtristancrazy, irrationally long10:29
gtristandont think I'm building tests no10:30
gtristanbut fwiw, before having virtio-net properly compiled in the kernel and working, the armv5l build took 5 and a half hours10:30
gtristanwhich is longer than it takes to build gcc10:30
robtaylorgtristan: maybe see if pump.mode helps (distcc)10:31
* robtaylor off driving now10:31
*** locallycompact has joined #baserock10:31
gtristanargh, embarrassing10:37
gtristannot that great news at all, it did look faster10:38
gtristanbut it's not fast by that much10:38
* gtristan will have to tweak further10:38
* gtristan tab completed glib.<sha1>.build-log.110:39
gtristanwhen it should have been glibc.<sha1>.build-log.110:39
gtristansorry to get your hopes up... distcc only brought the build time down by 20min, not 2hours10:40
gtristans/distcc/virtio-net I mean10:40
gtristanwithout distcc I suppose it would be much slower10:40
* gtristan off for feeding10:40
*** franred has joined #baserock10:42
*** franred has quit IRC11:48
*** franred has joined #baserock12:02
*** paulw has joined #baserock12:13
*** paulw has quit IRC12:21
*** paulw has joined #baserock12:21
pedroalvarezhrmf... I've just upgraded a trove I have here around, and systemd-networkd fails..12:22
paulsherwoodsoftware is hard12:23
*** paulw has quit IRC12:23
*** paulw has joined #baserock12:23
pedroalvarezbut I've deployed the same version I'm trying to upgrade to, and it works fine12:23
paulsherwoodupgraded using morph, or ybd, or other magic?12:24
pedroalvarezand i have no idea what's failing :/12:24
paulsherwoodybd upgrade is known to fail12:24
pedroalvarezusing morph12:24
paulsherwoodbecause of overlaps... the ordering comes out wrong in ybd12:24
paulsherwoodmaybe morph has hit the same problem12:24
pedroalvarezit's weird.. everything looks fine in the system.. but it doesn't work12:25
paulsherwoodthe overlap causes the wrong thing to be linked last iirc12:25
pedroalvarezand my systemd and dbus foo is not enough12:25
paulsherwoodcan we easily grep all of the logs of this channel? it was discussed here i think12:25
paulsherwoodpedroalvarez: ^^12:26
paulsherwood(i logged the issue for ybd at
paulsherwoodtraceability FTW :)12:26
pedroalvarezthat was because the system was booting busybox init12:27
pedroalvarezthis one is booting with systemd, but fails in the process12:27
*** franred has quit IRC12:28
pedroalvarezthe only thing I can think of, is that this snowflake I'm trying to upgrade has some special configuration that breaks when upgrading..12:28
*** paulw has quit IRC12:28
paulsherwoodok, so not the same problem, then12:29
pedroalvarezwell, not exactly the same problem, but who knows :)12:29
pedroalvarezi'll discard it whenever I have more information about it12:29
*** paulw has joined #baserock12:34
*** franred has joined #baserock12:54
*** paulw has quit IRC12:56
*** paulw has joined #baserock12:56
*** paulw has joined #baserock12:57
*** paulw has quit IRC13:01
*** paulw has joined #baserock13:01
*** fay_ has quit IRC13:02
*** fay_ has joined #baserock13:04
ssam2I made a Git repo that contains all the sources for a minimal-system as submodules13:12
ssam2not because i think that's a good idea, necessarily, just to see how well it works13:13
pedroalvareznext step move definitions to a makefile13:14
paulsherwoodssam2: did you see that thing robtaylor found as a 'better' submodules soln?13:16
* paulsherwood can't find the log13:16
ssam2i'm investigating a bunch of those13:16
robtaylorits pretty winning13:16
ssam2oh, didn't see subrepo though13:16
robtaylori'm using it to vendor in for my go projects13:16
robtaylorgives you the entire repo inside your repo, so you always have histroy13:17
robtaylorbut you just see commits from when you update the subrepos13:17
robtayloryou get remotes for the origins of your subrepos13:18
robtaylorand git subrepo fetch will give you a ref for your subrepo content13:19
ssam2cool, i'll try it out13:20
ssam2any idea if it's been tried with 100s or 1000s of repos?13:22
*** franred has quit IRC13:37
*** wdutch has joined #baserock13:55
*** tiagogomes has quit IRC14:02
*** paulw has joined #baserock14:09
robtaylorssam2: no idea :)14:13
paulsherwoodssam2: are you seriously considering pulling in all of the sources for (say) a system?14:15
robtaylorssam2: as everything ends up in your git repo i doubt it'll work well14:17
*** tiagogomes has joined #baserock14:17
ssam2paulsherwood: yes14:19
ssam2you have to do that anyway in order to build it14:19
paulsherwoodssam2: not if you have a cache server14:21
paulsherwoodbut yes14:21
*** fay_ has quit IRC14:36
*** franred has joined #baserock14:41
*** tiagogomes has quit IRC15:01
pedroalvarezright, I think I'm going to go ahead and merge 'gall' fix, we have not lorring anything new since yesterday15:02
franredpedroalvarez, sound sensible15:04
pedroalvarezfranred: is that a +1? :P15:07
*** fay_ has joined #baserock15:08
franredpedroalvarez, yep15:08
franredI can send the confirmation to the ml15:08
*** tiagogomes has joined #baserock15:14
pedroalvarezhere we go
*** paulw has joined #baserock15:26
franredpedroalvarez, merged15:33
pedroalvarezwoop woop! thanks franred!15:33
ssam2here's a reimplementation of GNU Make which converts Makefiles to Ninja rules!
*** CTtpollard has quit IRC16:27
*** CTtpollard has joined #baserock16:29
pedroalvarezANNOUCEMENT: and have been upgraded, and as a result, we have enabled lorry-controller  in them again17:16
pedroalvarezhappy lorrying17:16
ssam2nice work!17:16
CTtpollard'keep on trucking'17:16
pedroalvarezthanks for investigating the right version17:16
pedroalvarezupgrading g.b.o is really polished at the moment, that's nice17:16
* richard_maw notices he hasn't seen an update to the git in his debian system yet17:18
*** jonathanmaw has quit IRC17:20
ssam2they have quite a bit of backporting to do I think17:21
jmacsMost people are on 1.7.10 and the security update doesn't come in until 2.7.1 (at least)17:22
pedroalvarezheh, there is already a 2.8.0-rc317:25
*** ssam2 has quit IRC17:39
*** bashrc has quit IRC18:01
*** franred has quit IRC18:08
*** tiagogomes has quit IRC18:32
pedroalvarezsome info about configuring fail2ban on the wiki:
*** locallycompact has quit IRC18:41
*** toscalix has quit IRC18:43
*** fay_ has quit IRC19:07
*** edcragg has quit IRC20:03
*** edcragg has joined #baserock20:28
*** locallycompact has joined #baserock20:29
*** locallycompact has quit IRC22:53

Generated by 2.15.3 by Marius Gedminas - find it at!