IRC logs for #baserock for Tuesday, 2016-11-29

*** gtristan has quit IRC01:52
*** rdale has quit IRC03:34
*** rdale has joined #baserock03:34
*** radiofree has quit IRC03:35
*** radiofree has joined #baserock06:46
*** jude_ has joined #baserock08:17
*** ctbruce has joined #baserock08:22
*** cosm has quit IRC08:25
*** franred has joined #baserock08:26
*** cosm has joined #baserock08:27
*** toscalix has joined #baserock09:02
*** anahuelamo has joined #baserock09:45
*** anahuelamo has quit IRC09:50
*** anahuelamo has joined #baserock10:02
jjardonHi! Any idea what of the generic docker images support btrfs? Or that depends on the host entirely? I'm trying to think on a solution to be able to deploy in the runners; currently is not possible because they do not support btrfs10:05
rjekI assume the kernel will need to have btrfs support enabled10:07
rjek(On the host)10:07
*** anahuelamo has quit IRC10:19
*** anahuelamo- is now known as anahuelamo10:19
*** gtristan has joined #baserock10:34
*** locallycompact has joined #baserock10:39
*** leeming has joined #baserock10:40
locallycompactThis bug...10:51
locallycompactshutil.make_archive is behaving the same way11:19
locallycompactif I go into a ca-certificates debris folder and run11:20
locallycompactshutil.make_archive('foo', format='gztar', root_dir='.')11:20
locallycompactit truncates the filename11:20
locallycompactnothing to do with ybd11:20
locallycompactseems to be also doing it in python2 though11:22
* SotK wonders why its suddenly begun breaking11:22
locallycompactnot broken with py2 sorry11:25
*** jonathanmaw has joined #baserock11:29
locallycompactsomeone to confirm
jmacsConfirmed, file is truncated in python 311:39
paulsherwoodlocallycompact: thanks for tracking this down. i assume we need to rollback ybd to  py2 for the moment?12:13
locallycompactMaybe not strictly, I'm told this is fixed in 3.5.2
locallycompactfrom the duplicate link12:15
locallycompactworks in 3.5.212:21
locallycompactand nothing lower12:21
locallycompactpaulsherwood, what do you want to do?12:32
paulsherwoodrevert, i think12:33
locallycompactwho do I have to kill to move debian stable to 3.5.212:34
paulsherwoodno-one here, afaik12:35
radiofree+1 to revert as well (sadly)12:38
locallycompactCan we scrub the tarballs from the definitions build yesterday? They will be wrong.12:44
locallycompactor force over?12:45
radiofreelocallycompact: is there an upstream bug for this issue?12:45
radiofreewith python12:45
* radiofree wonders when the next debian stable is released12:46
rjekstretch?  First half of next year12:49
* locallycompact hibernates12:49
rjekIt's frozen now I think anyway12:49
rjekSo only bug fixes12:49
paulsherwoodlocallycompact: *all* artifacts from yesterday?12:51
locallycompactprobably a good idea12:51
paulsherwoodlocallycompact: will you submit the revert mr?12:52
franredlocallycompact, why you don't install the 3.5.2 package?13:03
franredif it doesn't have to many dependencies, you could always installing the deb package using "dpkg -i debian_package"13:03
locallycompactThe problem if anybody doesn't have 3.5.2 they will be uploading brkoen artifacts13:03
locallycompactif you run this with 3.4 it won't error until splitting13:04
* franred hides again13:04
radiofreeanother reason to move splitting until deployment13:26
radiofreewhat's the reason for splitting during artifact creation again?13:27
paulsherwoodsmaller artifacts13:29
paulsherwoodbut note even though 'it won't error until splitting' it breaks silently before that13:30
paulsherwood(ie it creates broken artifacts)13:30
* paulsherwood has removed all of the artifacts from yesterday13:31
paulsherwoodpls can someone else approve
locallycompactradiofree, sorry, I meant it fails at split installation time13:57
locallycompactnot at chunk splitting13:57
locallycompacthow is it smaller artifacts when misc pulls in .*13:57
locallycompactThe builds seem to have completed but it didn't merge14:33
locallycompactfine now14:38
paulsherwoodack, thanks14:38
locallycompactradiofree, this one might work now
radiofreeif i approve it can't merge until it builds right?14:45
jjardonradiofree: yes, builds have to succeed15:20
jjardonlocallycompact: maybe is a good idea to enforce that in ybd when the strict mode is selected , only warning if not?15:25
locallycompactI'll just leave this here for the rapture
locallycompactjjardon, enforce precise dependencies?15:25
jjardonYes, so those patches to "fix" definitions would be not needed in the future15:27
locallycompactI'd rather just have that way and not the other way15:27
jjardonSure, but you are braking older definitions if you only support precise deps15:28
locallycompactI say they are already broken15:28
*** brlogger` has joined #baserock15:37
*** ctbruce has quit IRC15:37
*** ctbruce has joined #baserock15:40
pedroalvarezNB All: network connection to baserock services hosted at DC (such as irclogs, or will be affected by some network issues similar to the ones we had yesterday15:41
benbrown_pedroalvarez: any chance you could let me know the disk space taken up by delta/cpython (not including lorry w-a)15:42
benbrown_on gbo, that is15:45
pedroalvarez~ # du -sh /home/git/repos/delta/cpython.git/15:45
benbrown_something is very wrong here15:45
benbrown_have a downstream trove where the mirror is taking up almost 4x as much space15:46
pedroalvarezthat's odd15:47
pedroalvarezbenbrown_: not sure if a copy of the contents may help you identify what extra data you have15:49
benbrown_an `ls -l` of objects/pack would be nice15:51
benbrown_actually, probably not15:52
benbrown_there's a 1.9G pack file here15:52
benbrown_no idea15:52
pedroalvarezmaybe a git gc helps there?15:53
* benbrown_ checks lc logs15:55
pedroalvarezis this in /home/lorry or /home/git ?15:56
pedroalvarezcan you also check the size in /home/lorry?15:57
benbrown_2.6G git dir o.O15:58
benbrown_also, hg?15:59
pedroalvarezis this a downstream trove?15:59
pedroalvarezor a trove that also has the same lorries?15:59
benbrown_lorry-controller.conf says downstream16:00
pedroalvarezwhy hg dir?16:00
benbrown_is cpython mercurial?16:01
benbrown_so it is16:01
SotKif its a downstream trove it should be lorrying from g.b.o's git mirror though iirc?16:02
benbrown_I think at some point it may have had the same lorries, for whatever bizarre reason, and somehow they've amalgamated into this horrifically sized repo16:02
* benbrown_ checks conf git log16:02
pedroalvarezI wonder what the state of the current repo is16:03
pedroalvarezI wonder if the 2 different lorries (from git, and hg) coexisted for some time, and messed everything up16:03
*** jude_ has quit IRC16:21
*** jude_ has joined #baserock16:23
*** ctbruce has quit IRC16:54
*** toscalix has quit IRC17:53
*** jude_ has quit IRC18:02
*** jonathanmaw has quit IRC18:11
*** locallycompact has quit IRC18:14
*** faybrocklebank has quit IRC18:20
*** gtristan has quit IRC18:34
*** jude_ has joined #baserock19:07
*** gtristan has joined #baserock22:13
*** gtristan has quit IRC23:12

Generated by 2.14.0 by Marius Gedminas - find it at!