[00:55] marcoceppi, ping about some emails not getting sent out? [02:35] anybody with some familiarity with nova-cloud-controller charm? === ray is now known as rawang === _thumper_ is now known as thumper === dspiteri is now known as darrenS === darrenS is now known as DarrenS === CyberJacob|Away is now known as CyberJacob === rogpeppe1 is now known as rogpeppe [08:37] I've been playing with Juju on AWS. Looks like a great way to scale. [08:38] Problem I have is money. I'd like Wordpress, MySQL, Varnish, Nagios deployed, but then that amounts to 4 instances and volumes. [08:39] is it an idea to pack all that functionality onto one meaty instance? [09:08] makara: yes, that can be done [09:08] juju deploy mysql --to 0 [09:08] will put mysql on the bootstrap node [09:09] thumper, i have this problem that I'm behind a firewall, so any web interface can only be port 80 [09:10] this creates the problem that if I install Wordpress, Juju-GUI, Nagios, and another simple static page server, they are not accessible [09:10] i could create multiple elastic IPs to that single machine [09:11] then I need to edit the Apache confs [09:12] can Juju handle this for me? [09:14] not out of the box === gary_poster|away is now known as gary_poster === FourDollars_ is now known as FourDollars === xerxas_ is now known as xerxas [14:57] wait.. you mean i could have deployed juju-gui to node 0? [14:57] Aye de mi thumper! i could hug you. [15:00] <_spacemonkey_> Greetings fellow jujulians. Er, jujubees. Uh, oh whatever. [15:00] <_spacemonkey_> Anyone here playing around with the jenkins and jenkins-slave charms? [15:14] <_spacemonkey_> So I used juju to deploy jenkins which works just fine. [15:14] <_spacemonkey_> Then I deployed jenkins-slave on another machine, and it is green as well :) [15:15] <_spacemonkey_> However from within jenkins they refuse to talk to each other. [15:15] <_spacemonkey_> When I ask jenkins to configure the jenkins-slave via java webstart, it gives me a different URL string than what is actually running on the slave (via juju). [15:15] <_spacemonkey_> I remembered the relationship, but that seems to be the part that is bustomicated. Any ideas? [15:17] lazypower: check out quickstart as well. It deploys the gui to 0 for you, does some new hotness. http://jujugui.wordpress.com/2013/12/05/new-juju-gui-0-14-0-juju-quickstart-0-5-0/ [15:19] _spacemonkey_: We use jenkins and jenkins-slave quite a bit, but I've found some odd stuff happens when you add relation it doesn't always wok [15:19] <_spacemonkey_> Yeah, definitely no wok here. ;-) [15:19] _spacemonkey_: fwiw, the charm store verison of the jenkins charms are a bit out of date, I think we've done work in a perosnal branch with the intention of perfecting it and merging it back in. [15:19] <_spacemonkey_> Ok will manually config the slave agent on the second machine. [15:19] _spacemonkey_: let me see if I can find these [15:19] <_spacemonkey_> Oh, cool! [15:20] * _spacemonkey_ waits patiently [15:21] <_spacemonkey_> Grazie mille. [15:41] rick_h__: Awesomeness. Do you have a link for quickstart or is it part of juju core? [15:41] lazypower: it's in the stable ppa I believe. The project is at https://launchpad.net/juju-quickstart [15:42] * lazypower hat tips === freeflying is now known as freeflying_away [19:12] what is the best method for getting juju tools for a MAAS environment w/o access to S3? the environment has access to a public swift bucket containing all the tools i need [19:13] mgz, ^ [19:15] adam_g: you should be able to use sync-tools and give it a local dir [19:15] mgz, dont suppose i can sync tools from a URL? :) [19:15] ...but if it's in swift already, you can just set the tools-url config [19:15] ah [19:15] or whatever we renamed it too [19:16] yeah, trying to find what that setting is called [19:19] adam_g: tools-metadata-url, needs the simple streams bits as well [19:19] which you can generate with the juju-metadata plugin [19:22] mgz, simple stream bits == a populated image-metadata-url ? === ajmitch_ is now known as ajmitch [20:07] <_bjorne> Hello, are that someone who know why only one node installing lxc, mongodb and run a dist-upgrade? i have two nodes and that is function on only one node? [20:08] <_bjorne> is that someone here? [20:09] _bjorne: you may need to hang around for a little while [20:11] <_bjorne> oki.. see if they have some information about my problem :) im get without hair soon :) if im cant solve this. [20:14] bjorne: sorry.. what is your question again? [20:15] <_bjorne> Im have some problem with maas and clients, should every clients (node) run a dist-upgrade and install lxc, monogodb? after im set up new node? only one node runs that not the others... i wounder why?! [20:16] <_bjorne> im use ubuntu 13.10 as server and clients. [20:16] so.. maas.. when you go to the web ui.. what is the status of your computers? [20:17] <_bjorne> bofh is allocated to root if you mean that? [20:17] they should be in ready state, i think.. disclaimer: i experimented with maas a bit a few weeks ago.. [20:18] so i do not know much about it.. [20:18] _bjorne: as someone who's been down that path... go with 12.04 LTS as the charms are all written for that. This link will get you the latest juju/maas for 12.04 https://wiki.ubuntu.com/ServerTeam/CloudToolsArchive [20:18] and then when you do juju bootstrap one node should be assigned to juu [20:18] <_bjorne> no i new on this with maas, and not new with computers :) [20:19] +1 on zradmin's point.. use 12.04 [20:20] <_bjorne> why only one node run dist-upgrade and installing lxc and mongodb? should that not installs on every node? with some automatic script? [20:21] only the bootstrapped node has mongodb.. this mongo stores all the configuration and state data, i think [20:22] <_bjorne> ok if you dont have lxc and mongodb juju cant find them? or have wrong with that? [20:23] jujud daemon should run on all nodes.. sorry.. i'm new with juju, just trying to help with what i think i know :) [20:24] <_bjorne> in maas server i have installed juju-core and juju bootstrap and after that i can see the first node in juju status and not the other node, becuse for som script is not run like it should... in the second node... [20:25] <_bjorne> in the first node after fresh install, the node run a dist-upgrade and installs lxc and mongodb and starts it.. and that is not runs in the second node... im dont find any information about it... [20:26] when you say juju bootstrap it will allocate a new node by default.. [20:27] brb === negronjl_ is now known as negronjl [22:46] Hey, I have a dumb question... when I bootstrap with juju, does the bootstrap node always use the same Ubuntu release as the juju controller? or can I specify a different release... e.g. can I have juju on a Saucy laptop but have it bootstrap to another machine using Precise as the OS on the bootstrap node? [22:48] bladernr_: that should work well; juju will pick precise by default, and I expect once 14.04 LTS is out, juju will probably default to trusty instead. (but don't quote me on -that- :) [22:49] how would I specify to bootstrap using precise? [22:51] bladernr_: ah, here we go, "default-series" key in the environment: https://juju.ubuntu.com/docs/config-aws.html [22:52] ahhh.. thanks. I was looking for that but kept finding more basic stuff... [22:52] much appreciated