[08:58] Hello Juju world! [09:50] o/ kjackal (belated) [09:51] ho lazyPower! [09:51] ahoy [09:51] *yo [09:55] i'm too hype to sleep. Google fiber incoming in ~ 3 hours. [10:02] Only slightly envious. You getting the 1000 or 100 Mpbs? [10:22] 1000 [10:22] I wont be able to fully leverage it until i get my POD delivered on the 17'th. I'll be limited to wifi, but right now i'm on some terrible DSL. This is a most welcome upgrade [10:24] i hate my 1mbit dsl [10:24] drawback to being in the country [10:31] I feel your pain [10:31] you can run a DC out of your home now :D [10:31] This particular DSL modem has the fun property of cutting out when someone in the building uses the microwave [10:31] heh, i'll be happy to have my homelab setup again with a nice fat pipeline. I was on 100 Mbit business class w/ comcast when i was in pittsburgh [10:32] and paying 2x as much as the google fiber cost, which blows my mind [10:49] lazyPower: Any news on the ansible layer front by any chance? [10:50] zeestrat - i haven't touched it outside of the initial work (circa charmer summit @ ghent), but i'm happy to mentor/handoff with someone that's interested in it [10:50] https://github.com/chuckbutler/ansible-base for reference [10:52] Cool. I might take you up on that! [10:54] Excellent :) [10:57] Speaking of layers, what is the main focus of the charm-tools project these days? Working on the reactive layer? [11:14] hah, my hubot charm works, there's earlbot1 [11:16] earlbot1: the rules [11:16] ok well not so much, back to config-changed hook :( [11:22] zeestrat - charm-tools is a handy collection of all kinds of goodness. Basically our tooling focus has been on building charms from layers yeah, but its also got a sister component for the charm store now, you'll notice when you issue `charm version` you get two versions back. One is for charm-tools (Developer bits) and the charm store cli (everyone) [11:22] jamon - Nearly congrats then :) [11:23] Gotcha :) [12:06] having a bit of a hard time with this (my first charm) [12:07] stuff like using config-get and knowing what the output looks like to a bash script. i've tried juju debug and checking it that way, but even then config get expects a bunch of environment variables like socket path, context id, things that I can't seem to locate [12:25] juju bootstrap seems to have lost the --config option in the latest beta. How do I specify the apt-proxy and apt-mirror now? [12:35] Huh. juju is juju 1. Which seems to require purgin juju-1, juju1.25 and juju-core to get rid of.... leaving me without any juju [12:37] Ok... purge juju 2.0 and reinstall, seems good. [12:37] jamon - those variables are available via "config-get" [12:37] lazyPower: yes, i've been using config-get in install/start/config-changed [12:38] but for inspecting the variable on the remote machine, it isn't so easy [12:38] jamon - link to your layer and output for the issue? [12:38] heh i'm just tidying things up here now for a commit :) [12:39] i'm happy to take a look and help where i can [12:39] awesome! [12:39] doing the readme is slow going [12:39] i have too many assumptions about how someone other than me might use this [12:59] right here lazyPower: https://gitlab.com/jamonation/layer-hubot/blob/master/hooks/install#L24-42 [12:59] and the corresponding config.yaml: https://gitlab.com/jamonation/layer-hubot/blob/master/config.yaml#L10-13 [13:00] i've resorted to checking if the value is "True" as a string [13:00] also, total aside, what would people here think of having a hubot in the channel if there isn't one already? [13:01] marcoceppi ^ [13:01] i know that marco has strong opinions on the subject, i defer to his judgement there [13:02] jamon - try if [ "${use_default_scripts}" -eq "True" ]; then ... fi [13:03] jamon - also, slight suggestion on line 45, try using -z to test if its empty instead of == "" [13:04] i can't wait to do this all with ansible :D [13:05] those suggestions ought to tidy it though, thanks [13:05] is using default: "" in config.yaml good/bad/acceptable? [13:06] i didn't dive too far into the go code to see how that's handled, "" seems to work [13:08] jamon - we tend to leave no value completly empty in the definition that way it retains the null characteristics but empty string is equally as fine, you just have to detect it in the charm code [13:08] i'm going to drop at any minute, fiber install beginning now. [13:08] will be back in a bit to check in on ya jamon [13:09] great, enjoy [13:09] send me some packets heh [13:15] jamon: personally, I prefer signal over noise. Bots just create noise and unless there's a very well scoped reason for one to exist I think it's best to leave them out of the channel [13:19] yeah i can't stand channels where every other message is a vcs commit, and then a slew of jenkins messages about staging builds [13:20] at the same time, some bot functionality can be really handy, like grabbing and announcing titles for http pages, so you don't have to open them if you aren't interested, etc. [13:20] kind of goes either way [14:01] natefinch-afk: ping for standup if you're around === natefinch-afk is now known as natefinch [14:02] rick_h_: coming [14:08] jacekn - back. hows the refactoring going? [14:24] tvansteenburgh: mbruzek: I'm trying to get juju2/bundletester to play nice. I've got jujudeployer 0.8, jujuclient 0.52, and bundletester 0.7.1 (all latest via pip). I'm getting the jujuclient.EnvironmentNotBoostrapped exception, which looks like it was fixed back in May. What am I missing? [14:25] aisrael: are you using charmbox or is this your own system [14:25] mbruzek: My own system, lxd inside lxd [14:26] aisrael - did you grab from upstream? not everything has landed afaict [14:26] i could be wrong, i think tim was saying it was like done. [14:26] aisrael https://github.com/chuckbutler/charmbox/blob/beta-11/install-review-tools.sh#L22-L30 [14:26] lazyPower: Not if it's in a ppa [14:27] thats what i was doing to get the latest bits in charmbox [14:27] works with beta11/beta12 [14:27] lazyPower: ok, cool, I'll grab those. Thanks! [14:29] aisrael, lazyPower: stuff is "done" but not released yet. if you want bleeding edge you'll need to run everything from source [14:30] aisrael: I have had success using the devel edition of charmbox [14:31] mbruzek - i dont have that in devel yet. I was waiting for them to land in the ppa, i did press a temporary charm box with those tho [14:31] lazypower/charmbox:beta-11 [14:31] oh my mistake [14:39] Amulet is chocking trying to import distro_info. Hmm. [14:42] tvansteenburgh: That seems like a missing dependency w/Amulet. Installing python3-distro-intro brought in the missing module [14:43] I'll file a bug [14:46] aisrael: huh? [14:50] tvansteenburgh: I hit a case where I was missing the `python3-distro-info` package after installing amulet from source. I'm verifying if it was me not following directions. [14:54] aisrael: if installing from source there wouldn't be any apt dependencies [15:05] tvansteenburgh: Right, I get that, but would it attempt to pull in dependencies via pip? [15:06] aisrael: if you install it with pip, sure, but python3-distro-info is not one of them [15:13] tvansteenburgh: Would it make sense to have the pip install depend on distro-info then? [15:15] aisrael: do you have a traceback i can look at? afaik amulet itself does not depend on that - if anything it would only be a packaging dep [15:23] tvansteenburgh: http://pastebin.ubuntu.com/21768436/ [15:23] oh ffs. I see it now. charmhelpers.contrib.amulet [15:23] :) [15:23] * aisrael gets another cup of coffee [15:24] tvansteenburgh: Thanks for putting up with that. ;-) [15:25] aisrael: no worries [18:28] petevg ping [18:29] lazyPower: hi. [18:29] hey there :) I recall some talk prior about you having written unit tests for a layered charm. are those somewhere i can take a look? [18:29] lazyPower: yes. Two places. Let me grab the linkies ... [18:29] brilliant, thanks for this [18:30] np. Here's a layer with unit tests: https://github.com/juju-solutions/layer-apache-bigtop-base (look under tests/unit) [18:30] https://github.com/juju-solutions/layer-apache-bigtop-base/blob/master/tests/unit/test_reactive_handlers.py -- oh my i've found the holy grail [18:31] ... and here's a proposed generic harness that will, among other things, solve your ImportError woes if you don't have a lib/layer/charms directory in your layer: https://github.com/juju-solutions/charms.unit/pull/1 [18:31] lazyPower: Hopefully it helps. :-) (And if it does, please drop a +1 in that PR, so that I can push the generic harness for others to use ...) [18:31] Will do [18:31] I wont be on it today but this is for tomorrow/later this week [18:32] so i can get you early feedback for sure [18:32] Cool. I'll look forward to the feedback (positive or negative) :-) [19:25] hey stub, petevg and i are having trouble installing a subordinate charm that includes layer-apt. it fails pretty early: http://paste.ubuntu.com/21799021/ because of the recent inclusion of clear_removed_package_states (which eventually tries to import apt.apt_pkg): https://git.launchpad.net/layer-apt/commit/?id=625d18edfbba37210adf9e0f198b7be4bbd7e1d8 [19:26] stub: this particular subordinate has use_venv: true, but does *not* have include_system_packages: true as one of it's layer options [19:26] i know we could add "include_sys_pkgs: true" to the layer.yaml, but i was wondering if the apt layer could do something so that this was not required.. for example, include python-apt in it's wheelhouse.txt [19:29] full disclosure stub, i know it's not that easy. pip installing python-apt requires DistUtilsExtra, which i can't seem to find as a pip'able req: http://paste.ubuntu.com/21799623/ -- it seems like it's only available from https://launchpad.net/python-distutils-extra === natefinch is now known as natefinch-afk [21:25] hey whats up guys? [21:25] is there any up-to-date network spaces docs floating around?