[09:28] <gnuoy> jamespage, do you think it makes sense to add functionality to the quantum-gateway charm to configure the IP, gateway etc of the ext-port ?
[11:15] <Roconda> Hey, can anyone point me in the right direction with JuJu? Looking for a solution to use juju on multiple servers with lxc(similar to multiple local env.). Lets say I´ve got multiple local JuJu enviroments with lxc, how do they communicate with each other?
[11:17] <marcoceppi> Roconda: They don't at the moment. We have an item on the roadmap to allow for "cross-environment relations" it's not implemented yet
[11:19] <marcoceppi> Roconda: we have limited contanier support for deployed servers/services
[11:20] <Roconda> marcoceppi: what do you recommend me, based on my situation?
[11:20] <Roconda> marcoceppi: good to know this issue is something being worked on :-)
[11:20] <marcoceppi> Roconda: so you can create LXCs in deployed environments and co-locate/create more dense services
[11:22] <Roconda> marcoceppi: using JuJu of without?
[11:23] <marcoceppi> Roconda: so containers is a great way to achieve what you want to do, and there is support for this currently, however there are still issues with this solution as there are networking issues
[11:27] <Roconda> marcoceppi: Okay, I guess I will use lxc at the moment and keep an eye on the cross-env. relation. Once cross-env. works I can migrate to JuJu
[11:28] <marcoceppi> Roconda: if you're not already, subscribe to the juju mailing list and watch the [ANN] emails
[11:28] <marcoceppi> Roconda: https://lists.ubuntu.com/mailman/listinfo/juju\
[11:28] <Roconda> marcoceppi: Thanks! Will do. Thanks for your help
[11:33] <marcoceppi> Roconda: while I don't have exact dates, we're looking to address these two features (and many others) within the next 6 months
[11:39] <Roconda> marcoceppi: Ah, in that case I´ll currently use the local environment so I can easily migrate to the cross-environment. I case I do need to scale out I could migrate to Amazon. The reason why I would like to use LXC is about costs. Working on a startup so I can´t afford to run many servers on amazon atm.
[12:04] <marcoceppi> Roconda: there are more affordable cloud providers that Juju supports available. Last time I checked HP Cloud was pretty low in cost compared to AWS
[15:29] <kurt_> jcastro: ping
[15:53] <jcsackett> sinzui: so we're hitting the request limit on the askubuntu API, thus no items.
[15:53] <sinzui> yay!
[15:53] <jcsackett> sinzui: there's a way to get more requests, requiring auth fun. thought we could avoid it, but evidently even once a day is too often.
[15:54] <sinzui> We disable staging and give prod a chance to do its job
[15:54]  * jcsackett nods
[15:54] <jcsackett> that's a thought.
[15:55] <jcsackett> we'll disable it on staging, see if prod picks it up tomorrow, and i'll work on the better version of api querying.
[15:55] <sinzui> lets do that for now. We saw it work on staging. We can discuss this with charmers next week
[15:55]  * jcsackett nods
[15:58] <jcsackett> sinzui: disabling on staging is done, and a card has been added.
[15:58] <sinzui> thank you
[16:44] <gary_poster> paulczar, last night's release of the GUI should have fixes for the bugs you reported.  please let me know if you encounter other issues.
[16:44]  * gary_poster working on announcements
[16:46] <paulczar> yessir appears to have fixed my issues with the gui
[16:46] <paulczar> now I have bug with juju-deployer :)
[16:46] <paulczar> https://bugs.launchpad.net/juju-deployer/+bug/1241721
[16:46] <_mup_> Bug #1241721: juju-deployer never finishes <juju-deployer:New> <https://launchpad.net/bugs/1241721>
[16:46] <gary_poster> paulczar, :-P looking
[16:46] <paulczar> it's still in bug-state ...so I can dig out extra debug info if you want anything
[16:46] <gary_poster> paulczar, plus small internal "yay" for gui ;-)
[16:50] <gary_poster> paulczar, hazmat probably is person to investigate eventually.  Out of curiosity though, do I understand correctly that, according to status and gui, you have all expected services but no relations?
[16:50] <paulczar> looks like rabbit is in pending state
[16:50] <gary_poster> huh
[16:51] <paulczar> so that might be holding it up
[16:51] <gary_poster> that smells more like a juju issue, yeah :-/
[16:51] <hazmat> odd.. still deployr should timeout w error
[16:51] <paulczar> oh I found it!
[16:51] <paulczar>   "4":
[16:51] <paulczar>     agent-state-info: '(error: invalid URL "http://cloud-images.ubuntu.com/releases/streams/v1/index.sjson"
[16:51] <paulczar>       not found)'
[16:51] <paulczar> I'll add the whole juju status
[16:52] <paulczar> added as a comment
[16:52] <paulczar> but that url works fine for me
[16:55] <gary_poster> paulczar, I'm fishing for help for you...
[16:55] <paulczar> thanks :)
[17:05] <paulczar> finally hit some sort of timeout and juju-deployer continued through
[17:05] <paulczar> but obviously still a failed deploy
[17:06] <gary_poster> paulczar, I guess that's good, in that you can probably work around it after the deployment.  So far everyone in the juju-dev pool I asked were at their EoD.
[17:08] <paulczar> yeah, just going to suck if I submit a charm contest entry which can't follow the contest instructions
[17:08] <paulczar> will need big walls of text in the readme :)
[17:19] <marcoceppi> paulczar: 1.16.0 ?
[17:24] <paulczar> yessir
[17:24] <paulczar> appears to be intermittant
[17:25] <paulczar> I ran again and it went through
[17:25] <paulczar> so guessing its not error handling well when a URL doesn't respond briefly
[17:27] <paulczar> the juju-gui export doesn't seem to export which services are exposed.  is this on purpose? or should I file bug ?
[17:30] <gary_poster> paulczar, hm.  that sounds like a bug, yeah, assuming normal deployer formats support this. :-/ thank you
[17:30] <gary_poster> should be easy to do, but inconvenient timing
[17:31] <paulczar> I'll test through again to make sure ... I didn't miss setting it
[17:31] <paulczar> then will file
[17:31] <gary_poster> ack
[17:59] <adam_g> is it possible to configure the network used by the local provider?
[18:24] <marcoceppi> adam_g: not that I know of at this time
[18:25] <adam_g> that stinks
[18:40] <paulczar> in my experience if you can get the local provider to work at all you should be happy :)
[18:52] <adam_g> paulczar, its working good for me.. maybe too good. i have a local provider running on my local system and a remote system, but would like to be able to address them both
[18:54] <paulczar> mine doesn't seem to be able to get working ssh keys in the lxc containers ... so I can't juju debug-log or juju ssh or anything
[18:56] <adam_g> paulczar, oh, maybe you're hitting this? https://bugs.launchpad.net/ubuntu/+source/lxc/+bug/1236577
[18:56] <_mup_> Bug #1236577: container's /home/ubuntu/ spawns with incorrect permissions, preventing SSH access <theme-oil> <lxc (Ubuntu):Fix Released> <https://launchpad.net/bugs/1236577>
[18:56] <adam_g> paulczar, i had the same issue
[18:58] <paulczar> it says they fixed ...  did you have to pull something from a master branch?   or was it fixed through to the apt-repos ?
[19:00] <adam_g> paulczar, it was fixed in LXC, not juju
[20:10] <sidnei> hazmat: patch landed in lxc, so progress
[20:10] <hazmat> sidnei, awesome
[20:11] <ahasenack> is there a way to use, say, the "ubuntu" charm, but on saucy?
[20:11] <ahasenack> the charm seems to exist only in the precise namespace
[20:12] <sidnei> ahasenack: if you check it out locally, put into a saucy directory in your local repo, then --repository=... local:ubuntu?
[20:13] <ahasenack> that I didn't try yet
[20:13] <ahasenack> I was looking for command-line options I may have missed
[20:16] <ahasenack> weird
[20:16] <ahasenack> I had to use this:
[20:16] <ahasenack> root@server-1e8400be-2e1c-469a-b794-62a5c587da47:~# juju deploy  --to 2 --repository $(pwd)/charms local:saucy/ubuntu ubuntu-saucy-container
[20:16] <ahasenack> without the saucy bit in local:saucy/ubuntu it would still try precise
[20:16] <ahasenack> I have charms/saucy/ubuntu/resto-of-the-charm
[20:17] <sidnei> ahasenack: probably from default-series in the environments.yaml defaulting to precise
[20:17] <ahasenack> could be. There was no default-series under the "local" environment though, only the others, so i assumed it didn't apply
[20:17] <ahasenack> since local will use whatever I'm running
[20:18] <ahasenack> and indeed, the bootstrap node is saucy
[20:18] <sidnei> i think that was the case in pyjuju
[20:18] <ahasenack> hm, I wanted to relate landscape-client to this one, but doesn't work, also because landscape-client only exists in the precise namespace
[20:19] <ahasenack> bummer, not friendly to non-precise workloads
[20:20] <sidnei> file a bug! i think this could well be a command-line option
[20:21] <ahasenack> the question is if all charms are supposed to work on all series
[20:21] <ahasenack> probably not
[20:29] <sidnei> nope
[20:31] <Nik_> does anyone know if juju clients newer than 0.6.1 (precise) would be backwards compatible with agents running on machines running precise. Since I can't easily move up the release for juju (using MAAS), I wonder if I can upgrade the clients for deploy --to functionality and some bugfixes
[20:36] <cotton> how does os x juju work.. does it create containers directly inside OS X?  or does it require a linux VM for an LXC host..
[20:58] <sarnold> cotton: there are a variety of "providers" that actually speak to whatever "cloud" backend you have. it might be aws or azure or openstack or "local" (lxc, linux-only) or "null" (bad name, hope that changes, that uses ssh)
[20:59] <cotton> ah ok
[20:59] <sarnold> cotton: the ssh-based provider may be able to host workloads on OS X alright, but I'm not sure how far that has been tested.
[20:59] <sarnold> cotton: using the juju frontend on OS X ought to be well-tested, or at least intends to be well tested. :)
[21:00] <cotton> so i can run juju on my mac via brew, then control a 'local' lxc cloud inside an ubuntu vm also?
[21:01] <cotton> also, does juju have charms that take care of HA clusters?  aka corosync/drbd/etc
[21:01] <marcoceppi> cotton: yes, so for the local provider on Mac/Windows, we utilize Vagrant to spin up a VM and run a special Ubuntu image inside that
[21:01] <cotton> yea ok i see HAcluster http://d.pr/i/8QEH
[21:01] <cotton> ah ok
[21:01] <cotton> cool
[21:01] <marcoceppi> cotton: so you need to vagrant up/vagrant ssh, then run juju commands inside the VM
[21:01] <marcoceppi> cotton: it's not a perfect story yet, just a stop gap :)
[21:02] <cotton> ok hah
[21:03] <marcoceppi> cotton: but outside of the local provider, all other providers/commands work on OSX with the native OSX juju client from brew
[21:03] <cotton> nice thanks
[21:08] <sarnold> marcoceppi: ha! you've got vagrant support going already? damn you guys are busy :)
[21:08] <marcoceppi> sarnold: the other way around
[21:08] <sarnold> marcoceppi: ha! vagrant has juju support already? damn those guys are busy! :)
[21:09] <marcoceppi> sarnold: we have a juju vagrant image which has local provider dependecies. So you can up a vagrant instance and get juju all installed witha  deployed local environment and the gui
[21:09] <marcoceppi> sarnold: hah!
[21:09] <sarnold> marcoceppi: that's pretty cool :D