[00:06] jose: I'm in the middle of something ... give me a few minutes ( about 30 I'm afraid ) [00:06] I have time :) [00:07] ping me when you're available :) [00:23] jose: I'm back ... what's up ? [00:39] negronjl: I'm having some troubles with the charm [00:39] because of the err_and_quit you defined instead of the set -x [00:39] now I'm getting some weird django errors, too [00:40] jose: let me deploy it so I can see as well [00:40] second [00:40] oh, go ahead [00:41] jose: What are you deploying on ? precise? trusty? [00:41] precise [00:41] jose: ok ... hold on ... bootstrapping [00:41] np :) [00:43] jsoe: Where are you deploying this ? ec2? something else ? [00:43] jose: ^^ [00:43] ec2 [00:44] can you ssh into the instance ? [00:44] but the error I get is an unbound variable one [00:44] I'm in right now [00:44] oh, no [00:44] lemme redeploy [03:26] jose: changes pushed [03:27] negronjl: ack, thank you! === menn0 is now known as menn0-afk === vladk|offline is now known as vladk === thumper is now known as thumper-afk === CyberJacob|Away is now known as CyberJacob === thumper-afk is now known as thumper === CyberJacob is now known as CyberJacob|Away === menn0-afk is now known as menn0 [08:17] ERROR bootstrap failed: waited for 10m0s without being able to connect: Permission denied (publickey,password). [08:18] hi, I'm trying to bootstarp a private cloud (openstack) with juju, but getting above error. where should i set the user/pass for the image that I'm using for bootstrap? [08:29] Hi, I'm trying to bootstarp a private cloud (openstack) with juju, but getting above following error. where should i set the user/pass for the image that I'm using for bootstrap? ERROR bootstrap failed: waited for 10m0s without being able to connect: Permission denied (publickey,password). === isviridov_ is now known as isviridov [09:54] ls [09:57] hi, after successfully bootstraping juju with my private openstack, when I try to deploy a charm (wordpress for example), it's complaining that network is unreachable. because it's trying to use private netowrk IP, not the public (floating) one. [10:05] mhshams: did you set "use-floating-ip" to true in the environments.yaml? [10:06] yes, i did. in fact bootstrapping is done with public ip, but only deploy is trying to use private one. [10:38] marcoceppi: even when i try to run "juju status", the result is the same: ERROR state/api: websocket.Dial wss://192.168.0.2:17070/: dial tcp 192.168.0.2:17070: network is unreachable [10:39] mhshams: bootstrap wasn't done with the public IP address if deploy is trying to use the private one [10:45] marcoceppi: but the private network is not visible from my machine (juju installed). how can juju bootstrap with ip that is not visible in the network? [10:47] mhshams: when juju bootstrap runs, it connects to the OpenStack install and uses the API to spin up an instance. If use-floating-ip isn't set to true before you bootstrap for the first time, then when it requests metadata about the instance to cache where the bootstrap node is, it'll just get the first address it responds with which will likely not be the floating-ip [10:50] marcoceppi: in my configuration, use-floating-ip is set to true, and I can see the juju-openstack-machine has the public ip assigned. i can connect to the machine with it's public ip. === natefinch is now known as natefinch-afk [12:09] hi rbasak === BradCrittenden is now known as bac [12:18] mhshams: if it's not production, try destroying the environment and re-bootstrapping. easiest way. [12:21] bac: hi! [12:21] bac: let me just read your email again [12:21] rbasak: cool. hey could you join us in #juju-gui in case someone else there is interested? [13:10] @jose: i did tried that, and the result was the same. one thing to clarify is that, i'm creating the image metadata locallay and using this bootstrap with --metadata-source. ( juju bootstrap --metadata-source=/home//.juju/metadata) [13:11] I'm no openstack expert [13:11] someone else should be able to help [13:13] Hey guys, I''m going to move the LXC/juju session from thursday to wednesday so didrocks and I can be there === natefinch-afk is now known as natefinch === rharper_ is now known as rharper [14:19] jamespage: ping [14:19] lazyPower, OMG it must be ping jamespage hour [14:19] lazyPower, hey! [14:19] oh, i'm sorry :( [14:20] I was curious if you had a chance to review teh vxlan-gateway that we sync'd on a while back, and what your notes were if any. Its come back up in the rev queue [14:22] lazyPower, I've not looked yet [14:22] lazyPower, (and no problem about the ping :-)) [14:23] * lazyPower takes notes to not ping jamespage before 10:45 EDT [14:23] can i interest you in a coffee, or a crepe while I have your attention? === isviridov is now known as isviridov|away === isviridov|away is now known as isviridov [15:07] if you have a charm that has peers, how would you instantiate an additional instance of the peer on a particular host/vm/container? Is it just juju add-unit --to ? [15:09] tech2: yes [15:10] More to the point I suppose. If I had a setup akin to US East, Europe (in terms of data-centres), how would I add units to one of those collections of VMs rather than to a particular machine? [15:10] though, you will probably want to use add-unit --to (kvm:lxc):Machine# [15:10] thanks Marco. [15:10] tech2: that's something you would do with constraints, there is a "zone" constraint landing in juju which will allow you to dictate which zone to create a new machine [15:11] marcoceppi: zones aren't here yet (assuming that's what you mean with the future-tense word "landing")? [15:11] tech2: correct, code just landed in trunk for zones [15:11] so it'll likely be in the next release of juju [15:12] marcoceppi: thanks [15:13] marcoceppi: is there a writeup about this anywhere (something I can show to others). I'm trying to upsell the virtues of juju at the moment and these are the kinds of issues being raised. [15:14] tech2: there's no real document on zones, it just follows the normal constraint pattern, https://juju.ubuntu.com/docs/charms-constraints.html [15:14] thanks [15:14] Once it lands in a release the docs will be updated [15:17] tech2: for the exact format, here's the bug against the doc: https://github.com/juju/docs/issues/86#issuecomment-45469055 let us know if you and your team have any other questions about juju! Happy to help [15:18] marcoceppi: no doubt I'll be in and out over the coming months, thanks again for the help. [15:37] Is there a tool for web-managing multiple environments (much like juju-gui is for a single env), or is that Landscape? [15:43] tech2: Landscape does a bunch of stuff, but I don't think it does management for multiple juju environments, it can and does know about multiple but won't give you the same experience as the GUI [15:43] marcoceppi: so the only way of really using a web-ui for multiple environments at the moment would be to have multiple instances of juju-gui (one per environment)? [15:45] tech2: correct, that's the way at the moment [15:45] marcoceppi: My issue is that we have an obligation to keep different clients' data on different instances such that we could do things like provide them with ssh access for an audit etc. I figured environments might be the cleanest way to do that but that sadly increases the devops overhead a bit. [15:46] tech2: I can see your point in increasing the overhead a bit. However, juju is fully drivable via a websocket per environment [15:46] so you could append functionality to easy the devops world by building tools to help drive deployments [15:47] Right, there's always a cost, it just depends who has to pay it :) [15:47] hello there :). can someone help me on how to connect to the mongodb server, on the juju bootstrap node? [15:47] core is also pretty responsive to users needs, so if there are shortcomings for your team feature requests can be opened,e tc [15:47] avladu: any reason why in particular? it's strongly discouraged given the schema changes and it's not guarenteed to work between releases [15:48] I just want to see if some relation ids are correct [15:48] judebug reasons only [15:48] debug reasons only [15:49] avladu: there's a python tools called juju-dbinspect which will allow you to connect to the mongodb and pull data out with Python [15:49] avladu: https://github.com/kapilt/juju-dbinspect [15:49] marcoceppi: thanks [15:49] alternatively, you can just `juju ssh 0` then connect to the mongo instance running there [15:51] marcoceppi: yes, I have done: juju ssh 0 [15:51] marcoceppi: I do not know the credentials to connect to the jujudb [15:55] avladu: that information is stored somewhere... not entirely sure, but the juju-dbinspect gets you around that [15:56] nice session jcastro [15:56] ugh local provider [15:57] yeah, minor snafu [15:59] and when lazyPower wanted it to fail, it didn't [15:59] * lazyPower flips tables [16:17] marcoceppi: thanks, with a few hacks, the juju-dbinspect worked [16:27] lazyPower, remind me, did you solve your logstash problem or just figure out what was wrong? [16:27] iirc a version mismatch? [16:29] jcastro_, rick_h_, do you guys have some slides for the roadmap discussion? [16:29] for UDS [16:29] I am trying to figure out how I should prep [16:29] some people do slides, other just do a status report [16:30] up to you [16:30] alexisb: I was just going to walk through and try to demo a little bit. === isviridov is now known as isviridov|away [16:44] I assume you guys know that upstream lxc has go bindings to the lxc api? [16:57] jacekn: i figuredo ut what was wrong, i have WIP on fixing it [16:57] s/jacekn/jcastro/ [17:14] marcoceppi, I'll need you at the top of the hour for "State of the Charm Store" === natefinch is now known as natefinch-afk === roadmr is now known as roadmr_afk [17:30] lazyPower: cory_fu: hi there, do you think you'll get a chance to look over https://bugs.launchpad.net/charms/+bug/942032 this week? [17:30] <_mup_> Bug #942032: Charm Needed: Review Board [17:30] sparkiegeek: you bet! [17:31] I'll promote it to first thing when i switch off from what i'm working on and move back into the rev queue [17:31] lazyPower: excellent :) looking forward to seeing the feedback [17:37] sparkiegeek: I'll go ahead and do a community review at the moment (I'm not a charmer), taking a look at it [17:37] jose: ah great, any and all reviews gratefully received [17:38] :) [17:38] jcastro, rick_h_, I know have a conflict for the 15:00 UTC time [17:38] would it be possible to move the UDS gui/core roadmap session to 16:00 UTC? [17:39] otp, one sec [17:39] or up to 13:00 UTC [17:39] arosales, can you move it? [17:39] nws [17:40] I can take a look [17:40] I had some trouble using Amulet with the PostgreSQL charm. The latter has a concept of "allowed_units" which charms that relate to it should poll to see if $UNIT_NAME is in. [17:41] the way Amulet inserts sentries between the two relating services means it'll never see it's own UNIT_NAME in allowed_units [17:42] sparkiegeek: its not a perfect solution, and thats oen we are aware of [17:42] theres a project called juju-db-inspect that we are looking at dropping in to aid in relationship inspection vs using sentries [17:43] alexisb: are you free at 16:00 UTC on Wed? [17:43] lazyPower: ok, sounds interesting [17:43] sparkiegeek: because it is a bit troublesome that the sentry can interfear with the relationships being defined. [17:43] lazyPower: indeed [17:43] alexisb: looking at switching "Juju with LXC containers for local dev" with "Juju Core and GUI Roadmap" [17:44] arosales, thanks [17:44] alexisb: so to confirm that time slot works for you, correct? [17:44] yes it does [17:45] ok [17:47] what's the next charm store session going to cover? [17:52] jose: ack [17:52] jcastro: ack [17:57] https://plus.google.com/hangouts/_/hoaevent/AP36tYdk_oApTaoYDWAv5eeR9zK3n8oUHTI8sE1WL4RptTN0lOaLpQ?authuser=0&hl=en [17:57] marcoceppi, ^^^ [17:57] jcastro: ack [17:57] and whoever wants to join [18:02] jcastro: what's the topic on this one? [18:03] arosales, let me know if you were able to make the switch, otherwise I will have to try and delegate the core roadmap discussion [18:05] sparkiegeek: mind a PM? [18:05] alexisb: the schedule is locked so I am pinging mhall119 [18:05] jose: WFM [18:05] arosales: it's not locked, track leads can edit it [18:06] mhall119: ah looks like jcastro or marcoceppi will have to make the edit [18:06] mhall119: thanks. I forgot I wasn't a lead this go around [18:06] arosales: yeah, you're only a lead on Ubuntu Development, if it's a devops session they can do it [18:06] arosales: also,sorry for the delay in replying, was hosting Mark's keynote [18:07] mhall119: no worries [18:08] mhall119: could i bug you to switch or add me as a lead to the Cloud DevOps? [18:09] arosales: sure [18:09] mhall119: thanks [18:09] give me one minute to do that [18:10] mhall119: ok thanks [18:12] arosales: done [18:13] mhall119: thanks [18:13] :-) === BradCrittenden is now known as bac [18:21] alexisb: all set for wed [18:22] thanks again mhall119 :-) [18:22] arosales, your the man, thank you! [18:23] alexisb: no worries, easy fix === roadmr_afk is now known as roadmr === roadmr is now known as roadmr_afk [19:01] mbruzek: you know how to run a test in an already bootstrapped env, with constraints? [19:16] jose you might be able to put the constraints in the Deployer class [19:16] method [19:16] oh, thanks [19:17] I don't think I have done constraints within Amulet yet. [19:17] if you can't find it please let me know. [19:19] jose It might be possible to set constraints in bootstrap and then all future deploys will have same constraints [19:19] https://juju.ubuntu.com/docs/charms-constraints.html#working-with-constraints [19:21] I'll check then :) [19:30] marcoceppi, let's talk local template lock LXC-hates me [19:30] okay [19:30] hangout or in here? [19:31] either [19:31] actually, hangout [19:31] go to the daily standup hangout [19:33] jcastro: session set up, youtube set up, will set up the page tomorrow morning === natefinch-afk is now known as natefinch === CyberJacob|Away is now known as CyberJacob === roadmr_afk is now known as roadmr === hatch__ is now known as hatch [20:31] lazyPower: ping [20:39] sebas5384: pong [20:39] lazyPower: hey! [20:39] Whats up? [20:40] we sell the first juju project! :) [20:40] to a client hehe [20:40] Thats awesome news! [20:40] congrats [20:40] so now i can be actually working in something thats goes to production [20:40] yeah! thanks! :) [20:40] so thats why i was lost these days [20:41] np :) [20:41] I've been occupied i assure you [20:41] yeah i imagine! [20:42] but hey! we must have our meeting to talk about vagrant and juju [20:42] and other devops stuff hehe [20:42] can we schedule it for end of week again? UDS has set me back in terms of time [20:43] yeah absolutely :) [20:43] let's talk trough the week then [20:48] sebas5384: surely === vladk is now known as vladk|offline === vladk|offline is now known as vladk === vladk is now known as vladk|offline === vladk|offline is now known as vladk === vladk is now known as vladk|offline [21:55] cmars: ping [21:55] sebas5384, pong [21:55] hey! cmars o/ [21:55] hi [21:55] i'm doing a drupal charm in ansible [21:55] oh, neat [21:56] i really like the ansible helpers [21:56] and i wonder if you can help me with a doubt [21:56] yeah!! thank you for your work on ansible+charms [21:56] :) [21:56] well [21:57] the thing is i was reading the helper code [21:57] oh, that's not mine.. [21:57] to whats all about the tags [21:57] i'm just a fan [21:57] ooh hehe [21:57] yeah thats right duh!! [21:57] hehe [21:58] sorry about that [21:59] no, i just don't want to take the credit. you should thank noodles775 [21:59] :) [21:59] well the doubt i have is, there's some way i can declare some default tags for the playbook [21:59] thanks noodles775 !! :) [21:59] hehe [22:00] btw here is the repo https://github.com/sebas5384/charm-drupal [22:00] :) [22:00] maybe noodles775 can help me with these then :P [22:01] you might be able to find help in #ansible as well. could be a way to set tags in the ansible env [22:01] yeah i was looking some info about that [22:02] thanks cmars :) [22:02] oh! and i didn't forgot to do the changes in the pull request for the juju-nat === hmatlock is now known as heath [22:02] i'm still going to do that, just need time hehe === james_w` is now known as james_w === CyberJacob is now known as CyberJacob|Away