IRC logs for #baserock for Tuesday, 2015-08-04

*** bashrc_ has joined #baserock08:04
*** jonathanmaw has joined #baserock08:17
*** mariaderidder has joined #baserock08:23
*** mariaderidder has quit IRC08:35
*** pedroalvarez has quit IRC08:35
*** CTtpollard has quit IRC08:35
*** cyndis has quit IRC08:35
*** kejiahu has quit IRC08:35
*** ctgriffiths has quit IRC08:35
*** mariaderidder has joined #baserock08:37
*** pedroalvarez has joined #baserock08:37
*** CTtpollard has joined #baserock08:37
*** cyndis has joined #baserock08:37
*** kejiahu has joined #baserock08:37
*** ctgriffiths has joined #baserock08:37
*** card.freenode.net sets mode: +v pedroalvarez08:37
*** edcragg has joined #baserock09:23
tiagogomesclicking on refs on the mason doesn't work. Is this a known problem?09:41
pedroalvarezyes09:56
nowsterIt looks like newest systemd wants lzo around.09:58
richard_mawnewest newest, or latest we've got on git.baserock.org?09:59
richard_mawwe need to update the lorry since they moved the repository to github09:59
nowsteryep10:01
nowsterLatest in definitions10:01
nowstercompiling now...10:02
pedroalvareztiagogomes: we have configured Mason to use the internal IP of git.baserock.org as Trove so that we could stop using internet on them and detect reproducibility problems like this one: https://mason-x86-64.baserock.org/log/7d3a901a14c1b3d6f636af9e60a5f84f5ad0777c--2015-07-19%2020:20:36.log10:02
nowsterAlso newer autoconf (and autoconf generated ./configure files) need "file" around to work out what binary format MIPS is using.10:03
nowsterHence, to get some of "core" to compile, I needed to move "file" into "core" and add a few build-deps.10:04
tiagogomespedroalvarez instead of doing that, why not adding a rule to drop outbound traffic if the destination IP address is not the *public* trove IP address. It is not good to have links in the public mason instance containing private IP addresses10:26
pedroalvarezthat would be a good idea, although I didn't set this up, so maybe I'm missing something10:29
pedroalvarezAll the rules are here: http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/infrastructure.git/tree/firewall.yaml10:34
pedroalvarezAlso, the configuration of mason is here: http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/infrastructure.git/tree/baserock_mason_x86_64/mason.conf10:34
*** locallycompact has joined #baserock10:39
nowsterHmm... not lzo. It probably needs xz, which is in "core".10:48
nowsterxz is another one that needs 'file' as a build-dep (on MIPS)12:15
* nowster finds a few more chunks in "core" that haven't had shared libraries built12:25
richard_mawouch12:29
nowster./configure tries to compile a test program, uses "file" to work out what binformat is in use, and if that fails disables building shared libraries12:32
*** mdunford has quit IRC13:07
*** mariaderidder has quit IRC13:33
edcragghi, i'm slightly confused by this error 'ERROR: Deployment failed as 43f302e8cd9b088fad87ae7149c54c6197772e321e4b76caa4fabfb5d6a6d3eb.system.devel-system-x86_64-generic-rootfs is not present in the artifact cache.'13:34
edcraggit happens when i rebased some definitions i was working on onto master, forcing me to rebuild13:34
edcraggdo all built systems have to also be present in the cache?13:34
SotKif you weren't trying to deploy an x86_64 devel system that looks weird13:35
edcraggit's referring to the local cache, too i guess13:35
edcraggit was originally built with a slightly older version of morph13:36
edcraggwhy does it look weird, SotK?13:36
SotKbecause the error suggests it is trying to deploy an x86_64 devel system13:37
edcraggi was trying to deploy an x86 devel system13:37
edcraggi guess that's not usually done, i was just using it as a test system to deploy13:38
edcraggi think my initial confusion over the error message was my brain associating 'cache' with 'cache server', in which case, silly question!13:41
*** tiagogomes has quit IRC13:42
pedroalvarezThere has been a change in morph that changes almost all (if not all) the cache keys of definitions.git13:43
edcraggpedroalvarez: yes i suspected that was the case, thanks13:44
*** mariaderidder has joined #baserock13:49
*** mdunford has joined #baserock14:45
*** tiagogomes_ has joined #baserock15:07
*** paulw has quit IRC16:10
*** mariaderidder has quit IRC16:39
*** jonathanmaw has quit IRC16:45
*** bashrc_ has quit IRC17:01
tiagogomes_random thought, I wonder whether the list of configuration extensions to run would fit better in the deployment cluster rather than in the system .morph17:02
*** edcragg has quit IRC17:23
paulsherwoodinteresting idea21:45
*** edcragg has joined #baserock22:45
*** edcragg has quit IRC22:53

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!