[11:17] <rick_h_> juju gui release love http://jujugui.wordpress.com/2014/11/18/juju-gui-1-3-0-released-now-with-added-services-view-2/ and https://plus.google.com/116120911388966791792/posts/7TAs5UiubKj and https://plus.google.com/116120911388966791792/posts/cbnGW5iw2Hy
[13:31] <stub> Are there any magic filenames that are ignored when juju packages a charm for deployment? I want to have a Python virtualenv in my charm directory containing pip installed package dependencies, which all goes well until deployment fails due to it containing symlinks outside of the charm.
[13:50] <rick_h_> stub: we build them in the charm
[13:50] <rick_h_> stub: you might be interested in our 'make deploy' target stuff http://bazaar.launchpad.net/~juju-gui-charmers/charms/trusty/juju-gui/trunk/view/head:/Makefile#L68
[13:51] <rick_h_> stub: we use that so from our dev checkout we can rerun that and redeploy the charm and have it build/etc.
[13:52] <stub> rick_h_: That seems to be what I'm doing, except when things get as far as amulet invoking juju-deployer invoking juju, it explodes because the charm contains the .venv containing symlinks outside of the charm directory.
[13:52] <stub> rick_h_: Do you repackage your charm without the .venv and deploy that instead?
[13:53] <rick_h_> stub: I'm double checking what the deploy target does. I believe so.
[13:53] <rick_h_> stub: because when we do a real release the charm builds the venv from a local download cache of the python deps
[13:53] <rick_h_> stub: so that way it mirrors a real deploy
[13:54] <stub> My .venv is being used to install dependencies the amulet tests need, so I don't need it pushed as part of the charm.
[13:54] <rick_h_> stub: yea, we split that into test-requirements and actual requirements
[13:56] <rick_h_> stub: so we've got the  server-requirements.pip vs test-requirements.pip
[14:06] <lazyPower> cargill_: cheers on your mail to the list, thanks for the follow up
[14:11] <stub> I'm getting very confused.
[14:12] <stub> On one hand, my amulet tests fail to deploy if they contain symlinks outside of the charm. These are not under revision control.
[14:13] <stub> If I remove the symlink issue, I have a different problem where only the version of my charm that has been committed is being tested, with uncommitted changes ignored.
[14:14] <stub> is this explainable, or do I just need to try again tomorrow with a freshly caffeinated brain?
[14:15] <jcastro> marcoceppi, ping
[14:16] <lazyPower> stub: do you have 2 copies ofthe charm in your juju_repository?
[14:17] <rick_h_> stub: ok, so there's something there in that. I know we hit an issue because amulet bzr'ifies your stuff? We hit issues with our git charm because of amulet trying to do bzr things but we wre using bzr-git and there was no .bzr. So I wonder if amulet is messing with you
[14:17] <stub> lazyPower: I no longer have $JUJU_REPOSITORY set (separate bug - amulet packaging up the charm I'm development and stuffing it in there without permission)
[14:18] <stub> rick_h_: I suspect it is actually juju-deployer hiding in there, but yeah, sounds familiar
[14:20] <lazyPower> stub: just making sure you weren't hitting this: https://bugs.launchpad.net/juju-core/+bug/1260170
[14:20] <mup> Bug #1260170: local charms are not deploy by filename <charms> <deploy> <juju-core:Fix Released> <https://launchpad.net/bugs/1260170>
[14:25] <lazyPower> tvansteenburgh: How much bzr magic are we still doing in amulet?
[14:25]  * tvansteenburgh reads scrollback
[14:26] <lazyPower> starts at 9:11:59
[14:26] <lazyPower> o, no it goes further back, my b
[14:26] <tvansteenburgh> stub: yes this is a limitation of amulet, b/c it's a limitation of deployer
[14:27] <stub> ta
[14:28] <stub> I can probably stuff my charm through tar into a tempdir and have amulet test that instead
[14:28] <tvansteenburgh> i noticed recently that we pass -L to deployer from amulet; removing that may solve the "doesn't deploy uncommitted changes" prob
[14:28] <stub> I'll look at that tomorrow, maybe a patch to amulet is possible.
[14:29]  * stub wanders off in search of late dinner.
[15:27] <mbruzek> Hey lazyPower someone opened a bug for a new charm, but it is not yet ready for review.  What is the correct state for the bug to keep it from going in the review queue?
[15:28] <mbruzek> I was thinking "In Progress" but there is also "Incomplete"
[15:28] <lazyPower> IIRC we just need to set the bug to "in progress" or "incomplete" and leave a follow up
[15:28] <lazyPower> the status + comment combo will do its due dilligence to put it in teh proper bin in the queue
[15:29] <lazyPower> when the bug author makes an update, it will see activity and move it back into the active state in the queue
[15:31] <lazyPower> rick_h_: quick update - worked on the first load, reloading the page brought the breakage to ff
[15:32] <lazyPower> gah wrong channel - sorry :(
[15:32] <rick_h_> lazyPower: it's ok, I'll find you whereever you hide
[15:33] <lazyPower> abandon thread!
[16:08] <hazmat> stub, tvansteenburgh, deployer uses the standard bzr api... it should work with plugins .. but its only going to do it if you specify vcs for the charm
[16:09] <hazmat> tvansteenburgh, oh.. uncommitted changes, yeah -L should do it
[16:10] <hazmat> stub, tvansteenburgh, the alternative is don't use branch: to pull the charm.. not sure if amulet is doing something around that
[16:16] <tvansteenburgh> hazmat: in a deployer file is it possible to specify a local charm w/o using branch: ?
[16:16] <hazmat> tvansteenburgh, definitely .. just set JUJU_REPOSITORY   and use charm: charm_name
[16:17] <tvansteenburgh> hazmat: ok cool, i may be able to make amulet handle this better then
[16:55] <mgz> gnuoy: bug 1392602
[16:55] <mup> Bug #1392602: local provider agent fails to restart on reboot of host - log dir missing <local-provider> <lxc> <regression> <juju-core:In Progress by wallyworld> <juju-core 1.21:In Progress by wallyworld> <https://launchpad.net/bugs/1392602>
[16:59] <jcastro> hey lazyPower
[17:00] <lazyPower> yo yo jcastro
[17:00] <jcastro> any videos from last week worth posting on insights? are they reusable or were they specific?
[17:02] <lazyPower> jcastro: everything from day 1 was instructional
[17:02] <lazyPower> specifically the first 5 minutes charming, and charm testing
[17:02] <jcastro> ok so I can reuse those?
[17:02] <lazyPower> yep
[17:03] <lazyPower> charm helpers didn't do as deep of a dive as I would think useful - we covered a lot of what is in there, and how to get into it + some of the future of charm helpers
[17:03] <lazyPower> maybe give that one a watch and make a yey/ney vote - but i'm leaning towards ney
[17:06] <lazyPower> jcastro: also the open feedback session was stellar
[17:06] <jcastro> might want to post a link to it to the list?
[17:06] <jcastro> I'm looking for the reusable ones for the website.
[17:40] <mgz> gnuoy: you can find the lxc-specific log for start issues in for eg /var/lib/juju/containers/ubuntu-local-machine-1/container.log
[18:07] <gnuoy> mgz, http://paste.ubuntu.com/9075776/
[18:13] <mgz> gnuoy: first ERROR line in that paste seems the relevent part, not sure why you're getting that
[18:14] <mgz> gnuoy: check your lxc mount settings?
[18:50] <skay> is anyone around who works on the python-django charm?
[18:54] <jose> rcj: ping
[18:54] <jose> skay: hey, what's up? probably can give you a hand
[18:55] <skay> jose: there are some changes I'd like to propose, but it looks like there is a big merge request about to land. does Patrick hang out here? I was wondering when it might land
[18:56] <skay> jose: I'd like to have an option to pip install from local files
[18:57] <jose> avoine: ^
[18:57] <skay> jose: so basically, for example, pass --no-index --find-links=wheelhouse -r requirements/ci.txt
[18:57] <jose> he hasn't been around for about 4h, so probably not here
[18:57] <jose> if the MP is open, then it should be landing soon (if no errors are found)
[18:58] <skay> and while I was looking at that, I see that you can import some things from pip, and maybe could call it directly. on the otherhand, it's not a public api so that could be a bad idea
[18:59] <skay> jose: there's a failing test for hte merge request, and I don't know where he is on fixing that. no new comments
[18:59] <jose> skay: lemme check, one sec!
[18:59] <skay> I could clone his branch and do what I need and then hop the merge request lands soon, then propose changes
[19:00] <skay> but I'd rather not
[19:00] <jose> skay: that will probably create a merge conflict, since the lander is usually someone different and revisions and stuff
[19:00] <jose> skay: could you PM me your email address? I'm going to send an email to Patrick and CC you in, asking for the status. sounds good to you?
[19:00] <skay> sure
[19:01] <jose> cool :)
[19:22] <avoine> skay: hi
[19:23] <skay> avoine: hi
[19:24] <skay> avoine: I'm interested in having python-django be able to install dependencies from local files. is this something that you'd like in python-django?
[19:24] <avoine> skay: you could probably use the pip_extra_args config variable
[19:24]  * avoine is reading the backlog
[19:25] <skay> avoine: oh thanks, I missed that! that may be all I need!
[19:26]  * skay is looking through source code
[19:27] <avoine> skay: hmm, It won't with the charm currently in the store I think
[19:27] <avoine> won't work
[19:27] <skay> okay, looking through pure-python branch now (that one?)
[19:28] <avoine> yeah this is the target for charm store
[19:29] <skay> avoine: I'll go ahead with using that one for now and maybe post to the juju mailing list if I need help (or chat here)
[19:30] <jose> marcoceppi: is there any way to force-update the review queue? it's taking me through several items that have already been promoted
[19:31] <avoine> skay: the pip_extra_args was added in an other version of the charm but the code was never backported to the pure-python branch
[19:31] <avoine> but that would not be too hard to do
[19:32] <skay> avoine: oh. I see the config file with the option is in the pure-python branch. (I just checked)
[19:32] <skay> I haven't checked the source code yet to verify that it uses it though
[19:32] <skay> I guess if it is not backported, I could make my own branch and use that for now?
[19:35] <avoine> skay: yes you could do that
[19:35] <skay> avoine: which branch is using the extra args feature?
[19:36] <avoine> skay: this one: http://bazaar.launchpad.net/~patrick-hetu/charms/precise/python-django/ansible/files
[19:37] <jose> whit: ping
[19:37] <whit> hey jose
[19:37] <jose> whit: have a min?
[19:37] <whit> jose, kinda in the middle of something
[19:38] <jose> whit: was just wondering where you had found the typos on https://bugs.launchpad.net/charms/+bug/1374085/comments/5, since I cannot find them
[19:38] <whit> jose, ah, whatever. what's up?
[19:38] <mup> Bug #1374085: New Shoutcast Charm <Juju Charms Collection:In Progress by lazypower> <https://launchpad.net/bugs/1374085>
[19:38] <jose> :P
[19:39] <jose> found the '(p)assword' one, but cannot find the admin one
[19:39] <whit> jose, in the readme?
[19:39] <lazyPower> oh my, i put assword as a config options. Clever
[19:40] <jose> :P
[19:40] <whit> iirc, I noticed it because it broke cutting and pasting
[19:40] <jose> hehe
[19:40] <jose> erm, let's check...
[19:40] <jose> found it
[19:40] <whit> cool
[19:40] <jose> thanks :)
[19:40] <whit> nw
[19:53] <skay> avoine: from a cursory glance, it seems only people who are using ansible in conjunction with juju will be able to use the pip_extra_args feature
[19:53] <skay> avoine: please let me know if I've misread things
[19:53] <avoine> skay: no your right
[19:54] <skay> avoine: maybe I could fork teh pure-python branch and then make a change so that I can use that with the hook.py code. would that be something helpful to contribute later? or perhaps someone is doing that already
[19:55]  * skay wants to use a charm store charm, ultimately
[19:56] <avoine> skay: I'll merge your changes for sure
[19:57] <jose> lazyPower, mbruzek, marcoceppi: authorization to promulgate shoutcast (https://bugs.launchpad.net/charms/+bug/1374085), haven't found any errors
[19:57] <avoine> skay: as soon as the branch is merge in the store we could add feature to it
[19:57] <mup> Bug #1374085: New Shoutcast Charm <Juju Charms Collection:In Progress by lazypower> <https://launchpad.net/bugs/1374085>
[19:58] <lazyPower> I'll defer to mbruzek as i'm the author
[19:58] <jose> ack
[19:58] <lazyPower> jose: did you fix up those typos for me?
[19:58] <lazyPower> i dont think we want end users looking for ***word
[19:58] <jose> lazyPower: I did, don't worry. people won't have to set that config option :P
[19:58] <lazyPower> solid
[19:59] <rcj> jose, I'm at a sprint this week but I can't get to work on the nrpe charm test failure next week if not sooner.
[19:59] <jose> rcj: sounds good, thanks :)
[20:00] <rcj> jose, that was the question, right?  not sure now reading the backlog if I confused this with conversation you were having about the  python-django charm.
[21:34] <mgz> gnuoy: try deleting /var/lib/lxc/juju-trusty-template or similar, Ian suggests
[21:35] <gnuoy> ack, will do
[21:38] <gnuoy> mgz, wallyworld_ fixed !
[21:38] <gnuoy> thank you
[21:39] <wallyworld_> gnuoy: that's good news, and highlights something we should deal with
[21:39] <gnuoy> tip top
[21:39] <wallyworld_> thanks for finding the issue
[21:40] <jrwren> 150% faster mongodb deploys: https://code.launchpad.net/~evarlast/charms/trusty/mongodb/trunk/+merge/242136  eco folks ;]
[22:42] <rcj> jose, Just re-read my earlier comment and realized that I mis-typed.  I can't get to work on the nrpe charm until next week most likely.
[22:46] <lazyPower> jrwren: merged
[23:03] <jrwren> lazyPower: ! thanks!