[00:11] <magicaltrout> hey marcoceppi how angry were you when looking at code?
[00:12] <magicaltrout> I can't help but think it was something like this
[00:12] <magicaltrout> https://dl.dropboxusercontent.com/u/8503756/8252253547_c234b97423_b.jpg
[01:41] <lazyPower> skay_ xenial right?
[01:41] <skay_> lazyPower: yes
[01:42] <lazyPower> I just ran it on a Xenial host, seems to have worked here with the hello-world snap...
[05:36] <stub> skay_: Are you by any chance trying to do this under LXD? Snaps don't work in LXD containers for reasons I haven't looked into. I was hoping this would be fixed in the short term so I wouldn't need to document the limitation or put checks in the code.
[08:14] <kjackal> Good morning Juju world
[08:20] <anrah> Good morning!
[09:10] <admcleod> hello - what, specifically, do i need to upgrade to resolve this: ERROR unrecognized command: charm publish
[09:33] <evilnickveitch> admcleod, i believe it has been replaced with 'charm release'
[09:54] <andrey-mp> Hi! How to define bindings for the charm? Some charm uses 'network-get public --primary-address' and gets the same as 'unit-get private-address' but in juju 1.25 it was public-address
[09:54] <andrey-mp> and it's a problem
[13:33] <skay_> stub: hey, yes indeed I'm using lxd. maybe I can ask in the snappy channel
[13:37] <skay_> lazyPower: were you using lxd?
[13:44] <rick_h> admcleod: s/publish/release
[13:47] <MotherDuckingNew> I did everything it told me to do, but this is still different:
[13:48] <MotherDuckingNew> I did everything it told me to do, but this is still different:Unit         Workload  Agent       Machine  Public address  Ports  Message
[13:48] <MotherDuckingNew> mediawiki/0  waiting   allocating  0                               waiting for machine
[13:48] <MotherDuckingNew> mysql/0      waiting   allocating  1                               waiting for machine
[13:48] <MotherDuckingNew> Machine  State    DNS  Inst id  Series  AZ
[13:48] <MotherDuckingNew> 0        pending       pending  trusty
[13:48] <MotherDuckingNew> 1        pending       pending  trusty
[13:50] <marcoceppi> MotherDuckingNew: what's the cloud you're using?
[13:51] <MotherDuckingNew> I don't know. I just followed this. https://jujucharms.com/docs/stable/getting-started
[13:53] <rick_h> MotherDuckingNew: can you please pastebin the output of juju status --format=yaml
[13:59] <MotherDuckingNew> http://pastebin.com/tR1KJwAP
[14:00] <rick_h> MotherDuckingNew: so the feedback there is important: "copying image for..."
[14:26] <lazyPower> skay_ - i was not.
[14:30] <skay_> lazyPower: I guess I could switch to using juju1 unless there is a way to use juju2 with lxc
[14:30] <skay_> lazyPower: juju1 doesn't have layers, right?
[14:30] <lazyPower> skay_ - you can use juju2 with lxd
[14:31] <lazyPower> you can also use layered charms with juju 1
[14:31] <skay_> lazyPower: old lxc
[14:31] <skay_> lazyPower: if the layer-snap has trouble with lxd vs. old lxc then I can't use lxd
[14:32] <skay_> lazyPower: ok, to recap. I can still use layers, regardless
[14:33] <lazyPower> yep :)
[14:33] <lazyPower> i dont think juju 1.25 has resource support though
[14:33] <lazyPower> and you were attempting to deliver a snap as a charm resource right?
[14:46] <MotherDuckingNew> (I just got disconnected) What is the lesson I should learn from that pastebin?
[14:47] <rick_h> MotherDuckingNew: sorry, the status there states that lxd was downloading the ubuntu-trusty images and it was at 97% done
[14:48] <rick_h> MotherDuckingNew: so the lesson was that there's more detailed information on "waiting for machine" in the yaml output that pointed to what was going on
[14:58] <petevg> Hiya, cory_fu. How do you use that interface you added to zeppelin to register notebooks? Is it in the current zeppelin charm in the store?
[14:59] <cory_fu> petevg: It is, and it's documented in the interface's README: https://github.com/juju-solutions/interface-zeppelin#charms-registering-a-notebook
[15:03] <petevg> cory_fu: cool. I am jet lagged, but gave it a read. We don't have a way of just dumping a notebook in without it being part of another charm now, right?
[15:03] <petevg> ... but I guess I can expose the register notebook routine as an action.
[15:04] <cory_fu> petevg: Indeed. I had planned to make an action for it, but haven't done so yet
[15:05] <petevg> cory_fu: cool. If I get ambitious, I will add it.
[15:23] <marcoceppi> cory_fu: does this seem too crazy? https://github.com/juju/charm-tools/pull/284
[15:25] <cory_fu> marcoceppi: Seems fine to me
[15:25] <marcoceppi> cory_fu: I have no idea what entity is
[15:25] <marcoceppi> or if it has makedirs_p as a method
[15:27] <cory_fu> marcoceppi: It's a Path object, so yes, it should have that method
[15:27] <cory_fu> marcoceppi: It gets populated by https://github.com/juju/charm-tools/blob/master/charmtools/build/builder.py#L326 via https://github.com/juju/charm-tools/blob/master/charmtools/build/builder.py#L295
[15:30] <ryebot> Do child layers have access to relations from parent layers?
[15:32] <ryebot> Like, if I have layer-foo that depends on layer-bar, and layer-bar includes interface-baz, does layer-foo have access to interface-baz?
[15:33] <marcoceppi> ryebot: yes
[15:33] <ryebot> marcoceppi: excellent, thanks!
[15:34] <ryebot> Also, I assume it's bad form for interface layers to do things beyond communication, like install software. Is that correct?
[15:38] <marcoceppi> ryebot: generally, yes, typically if you have to do things like installation, it's best to make a layer-<interface> which does the common components in a reactive directory for installation and handling of complex  measures then raising it's own state
[15:38] <marcoceppi> ryebot: that's a guide line, and there are always exceptions
[15:38] <marcoceppi> cory_fu: so, it works, but doesn't
[15:38] <marcoceppi> cory_fu: jk, it works
[15:38] <ryebot> marcoceppi: perfect, thanks
[15:43] <skay_> lazyPower: yes, I wanted to use resources
[15:43] <skay_> lazyPower: I like the idea of snapping my application, but I don't want to publish it int he store
[15:44] <marcoceppi> cory_fu: can you TAL again re: https://github.com/juju/charm-tools/pull/284
[15:44] <skay_> lazyPower: I thought maybe if I used a resource I wouldn't need to
[15:45] <cory_fu> marcoceppi: Oh, yeah, that makes sense
[15:46] <marcoceppi> cory_fu: I actually tested it, lol
[15:46] <cory_fu> Madness
[15:51] <deanman> Evening, is there a seperate channel for juju docs and contributing to that?
[15:54] <evilnickveitch> deanman, there isn't a specific docs channel no, but I can help you.
[15:54] <mgz> deanman: I don't think there's specific one for docs, but also happy to have docs talk in #juju-dev
[15:54] <evilnickveitch> oh yeah, or there :)
[15:55] <deanman> evilnickveitch: i'll catchup with you in juju-devs then
[15:59] <marcoceppi> cory_fu: do you have a few mins to chat about the snap problems?
[15:59] <cory_fu> marcoceppi: In a meeting.  Give me a few min
[15:59] <marcoceppi> kk
[16:13] <cory_fu> tvansteenburgh: Can you join us in https://hangouts.google.com/hangouts/_/canonical.com/daily-bigdata?authuser=1 real quick?
[16:17] <cory_fu> Oh, he's at the dentist
[16:17] <cory_fu> marcoceppi: Ok, I'm free.  What HO should I join?
[16:29] <marcoceppi> cory_fu: how about
[16:30] <marcoceppi> cory_fu: https://hangouts.google.com/hangouts/_/canonical.com/snappy-pip
[16:55] <tvansteenburgh> cory_fu: back from dentist, you still need me?
[16:57] <cory_fu> tvansteenburgh: kwmonroe, kjackal, and I wanted your help with getting the RQ updated in production, since we tested it in staging last week
[16:58] <cory_fu> tvansteenburgh: I have another meeting right now, but if you wanted to jump on a HO w/ kwmonroe and kjackal, I can join when I'm done
[16:58] <tvansteenburgh> cory_fu: sure, whatever you guys prefer
[17:00] <kwmonroe> cory_fu: kjackal, let's do it after the CI chat since that might surface more work for tvansteenburgh.  so, like in an hour-ish.
[17:01] <tvansteenburgh> kwmonroe: i have a meeting in an hour
[17:05] <kwmonroe> tvansteenburgh: if you're free now, https://hangouts.google.com/hangouts/_/canonical.com/juju-ci (kjackal cory_fu too if you want)
[17:12] <jrwren> the ceph charm suggests it should work with directories instead of devices: https://jujucharms.com/ceph/  osd-devices says "charm assumes anything not starting with /dev is a directory instead" Does that actually work? If it doesn't were should I file a bug?
[17:54] <jrwren> lazyPower: how/where are we meeting re: elasticstack
[17:54] <lazyPower> jrwren invite deployed
[17:54] <lazyPower> https://hangouts.google.com/hangouts/_/canonical.com/elastic-stack?authuser=0
[17:54] <lazyPower> to anyone else following along at home, you can join here
[17:54] <jrwren> lazyPower: thanks!
[20:11] <bdx> I'm hitting this on charm build across the board http://paste.ubuntu.com/23482201/
[20:14] <bdx> ok nm only in certain charms
[20:19] <lazyPower> bdx - is there an exclude: directive in the layer.yaml?
[20:19] <bdx> lazyPower: no
[20:20] <bdx> I've removed some layers from my layer.yaml ... getting charm build to succeed now ... should find the cause momentarilly
[20:21] <lazyPower> bdx - i know that right now etcd only builds against charm-tools from tip, which is why i asked. and its due to the exclude: directive in the layer.yaml
[20:21] <bdx> grrrrr - https://github.com/jamesbeedy/juju-layer-composer/blob/master/layer.yaml#L6
[20:21] <bdx> srry - my b
[20:22] <lazyPower> tadaaaa
[20:22] <lazyPower> invalid yaml, dun dun dunnnnn
[20:24] <lazyPower> bdx np, i do it all the time too