[01:27] <TheJeff> hey #juju - what's 'Insufficient peer units' mean?  am I falling short here
[01:28] <jrwren> charmers: kindly requeue http://review.juju.solutions/review/2357 please?
[01:38] <marcoceppi> TheJeff: what service is saying that?
[03:10] <jrwren> charmers: i see the requeue happened, but I am unsure about the results. Can someone help me interpret? http://juju-ci.vapour.ws:8080/job/charm-bundle-test-aws/1694/console
[08:11] <marcoceppi> jrwren: we recently migrated to a new server for all of Juju QA, including charm-testing. This seems to be an issue related to that
[12:47] <Icey> can you have multiple base layers with install hooks?
[14:00] <stub> Icey: Yes. You don't know the order they will run though (only that all of the @hook('install') handlers will run before the general handlers)
[14:18] <marcoceppi> Icey: yes
[14:25] <Icey> so, new layer charm, any idea why I'd be getting http://pastebin.ubuntu.com/13824394/
[14:26] <Icey> for some reason it isn't setting up charmhelpers
[14:34] <marcoceppi> Icey: can you give the entire trace?
[14:34] <Icey> that's really it: http://pastebin.ubuntu.com/13824600/
[14:34] <Icey> just pulled out paths
[14:34] <Icey> I was following the debug log since deploy, that's the first line
[14:36] <Icey> do I need to actually include charmhelpers in the lib to make sure it's available? most of the other layers aren't doing that so I didn't think it was necessary but it seems to not have charmhelpers available even though I'm suing the basic layer
[14:36] <Icey> (in another layer)
[14:37] <Icey> http://pastebin.ubuntu.com/13824666/ is the file with the referenced line failure
[14:45] <marcoceppi> Icey: can you pastebin hooks/install
[14:57] <Icey> http://pastebin.ubuntu.com/13824666/ is hooks/install
[14:57] <Icey> marcoceppi ^^
[14:58] <lazypower> hmm.. whats this line do? @translate_exc(from_exc=OSError, to_exc=NotImplementedError)
[14:59] <lazypower> i see it decorating a method in charmhelpers, and its not obvious to me what its doing
[14:59] <lazypower> i'm guessing if the subprocess.check call is invoked and errors, it acts as a try/catch block and instead throws a different error?
[15:00] <marcoceppi> Icey: do you have a wheelhouse directory?
[15:01] <Icey> marcoceppi negative, just wheelhouse.txt
[15:01] <marcoceppi> Icey: well, that's the problem
[15:01] <Icey> with requests and shell
[15:01] <marcoceppi> Icey: what version of charm-tools do you have ?
[15:01] <Icey> shouldn't the charm build make that?
[15:01] <marcoceppi> Icey: oh, what does your layer.yaml looks like?
[15:01] <Icey> charm version charm-tools 1.8.0
[15:01] <marcoceppi> Icey: do you have a link to your layer?
[15:02] <marcoceppi> Icey: you need to upgrade
[15:02] <marcoceppi> charm-tools is currently 1.10.1
[15:06] <Icey> upgrading now
[15:06] <Icey> odd thing is I've made a different layered charm that worked
[15:08] <marcoceppi> Icey: it depends on if it wheelhouses or not
[15:09] <Icey> ah, I suppose one of the layers I'm using does then?
[15:10] <Icey> what is wheelhouse?
[15:10] <Icey> and is it architecture dependant?
[15:14] <marcoceppi> Icey: yes
[15:14] <marcoceppi> Icey: no
[15:15] <Icey> apparently it requires python3 to build now, installing...
[15:15] <marcoceppi> Icey: the point is a wheelhouse is a better way to bundle dependencies for hooks without having to install the dep in the tree like before but also makes sure that the dep is installed on the system architecture
[15:16] <Icey> sure, just annoying to have to change build environment so often :-P
[15:16] <marcoceppi> Icey: well we're still stabilizing the tool
[15:16] <marcoceppi> Icey: this is us ramping up to a 2.0
[15:23] <Icey> marcoceppi my client just went wonky, back now (I think)
[17:33] <lazypower> cory_fu: have we written any interface layers using peer relation(s) that you are aware of?
[17:33] <lazypower> i scanned the list and didnt see any. i think mbruzek and I are at a point where we're going to pilot this if it hasn't been piloted before
[17:34] <cory_fu> No, I haven't.
[17:34] <cory_fu> Sorry, I mean, no I'm not aware of any
[17:35] <lazypower> ack, *fingers crossed* lets traverse into the dark without a torch
[17:52] <lazypower> tvansteenburgh marcoceppi : could use a quick review on this one when you get time - https://code.launchpad.net/~lazypower/charm-helpers/payload-tracking/+merge/279927  - it adds the payload tracking bits from the 1.25 release to charmhelpers.core.hookenv
[17:53] <lazypower> I'm not sure i did this right, i tried to model it off of the leader bits which i know have requirements on which version of juju to use
[17:53] <lazypower> *min version required
[18:40] <cmars> i'm noticing a lot of extra packages getting installed in the install hook, in charms built with 1.10.1. can we apt-get install python3-pip --no-install-recommends, would that still work?
[18:40] <cmars> really don't want g++ on all my charms...
[20:33] <marcoceppi> tvansteenburgh: still seeing charm testing failures http://juju-ci.vapour.ws:8080/job/charm-bundle-test-aws/1702/console
[20:33] <tvansteenburgh> marcoceppi: ugh, back to the drawing board