[03:10] this might be a dumb question, its been a little while since i've tried juju... is there no longer a way to use lxc containers? [03:17] nvrmind, found my answer. looks like lxc is on the way [03:38] hey there, ive got some charms deployed a custom django app, gunicorn, apache2, haproxy, and postgresql. ive gotten everything deployed and relations created and I can access the web application after getting guincorn deployed. however when i attempt to build a relation between apache2 and haproxy and load my template {{ haproxy_gunicorn }} isnt being filled in even though i've done a relation-set services_all in the website-relation-joined [03:39] negronjl: you awake/busy? [03:39] stokachu: are you trying to add a relation between the haproxy that is already attached to gunicorn? [03:40] so i attached gunicorn to my django app [03:40] one sec ill paste my steps [03:40] stokachu: I'm not long for this world, the bed is calling, but I'll try to help you out until then [03:41] marcoceppi: thanks :D http://pastebin.ubuntu.com/5866972/ [03:41] those are my steps and and ive got a vhost template that has {{ haproxy_gunicorn }} to pull that information [03:43] How would haproxy know of gunicorn? It looks like seg-dash -> gunicorn then haproxy -> apache2 haproxy and gunicorn don't seem to intercept [03:45] stokachu: ^ [03:45] marcoceppi: hmm [03:46] so would i add relation from apache2 -> gunicorn then haproxy -> apache2? [03:46] also im not sure the benefit for haproxy + apache2 [03:46] You could, I'm not seeing the point of apache in this equation [03:47] or haproxy [03:47] or gunicorn [03:47] all three act as proxies (verifies) for the most part [03:47] so i was doing apache for the static files and gunicorn for the application [03:47] stokachu: ah, I see [03:47] So that will be promblematic [03:47] but i saw all these other charms and got excited [03:48] so when i manually setup an app i have nginx/apache reverse proxy to gunicorn [03:48] and thats it [03:48] You, you'd ideally want to serve them all from one IP/fqdn, but haproxy can't talk to multiple charms at once since it won't know which address to talk to [03:49] So you'd effectively have one ip address that maps to the apache2 static (via haproxy) then the gunicorn would be the app server [03:49] which could be put in front of another haproxy, if you so wanted [03:49] actually, it could not [03:49] since it dones't provide a website relation, for some weird reason. [03:50] that is unfortunate [03:50] so if ive just got one application i want to deploy should i ditch haproxy or apache and just proxy it through to gunicorn? [03:50] So, are you actually using gunicorn to deploy a wsgi? [03:51] I'm not sure what the seg-dashboard is [03:51] thats the django app [03:51] so i was using gunicorn in place of mod_wsgi [03:52] One sec, I need to check one more thing [03:52] OKAY [03:52] sorry, was confused that gunicorn was a subordinate [03:53] So you've got seg-dashboard that's using gunicorn deployed on top of it as it's web server [03:53] yea [03:53] so far [03:53] my next step is to proxy it through apache [03:53] I'm not sure why, I feel like gunicorn should *provide* a website relation over http interface, so that you can actually hook it up to a proxy [03:53] I'm not sure why it doesn't* [03:54] let me check a few more charms to see where the onous is for website relation and gunicorn [03:54] ok [03:55] so far ive been basing my stuff of apache2,gunicorn,and python-django [03:55] and the apache2 one mentioned haproxy+gunicorn [03:55] So, the example gunicorn provides, python-moinmoin, python-moinmoin (and python-django) has the website relation on that charm [03:55] yea so i can get python-django and gunicorn to work [03:55] So, what you could do is make sure seg-dashboard provided a website relation over the http interface, it would then set the "hostname" and "port" that the proxy would use for it's request [03:56] ahh [03:56] i think im missing a provides [03:56] So, as far as connections go, seg-dashboard -> apache2 (put static here as well via custom vhost) -> haproxy [03:56] where haproxy would push to apache2, which would (could) serve static or push down in to the app layer [03:57] gotcha, lemme go back to messing with it and ill let you get some sleep [03:57] thanks for the help :D [03:57] if you're not worried about static content, or if static could live in a CDN/on the seg-dashboard, just wire up seg-dashboard -> haproxy and cut apache2 out of the loop [03:57] stokachu: no problem! I'll be back EDT AM if you want to drop questions here or on the list [03:57] good luck! [03:57] marcoceppi: thanks will do :D [03:58] cheers === defunctzombie_zz is now known as defunctzombie === defunctzombie is now known as defunctzombie_zz === CyberJacob|Away is now known as CyberJacob === defunctzombie_zz is now known as defunctzombie === defunctzombie is now known as defunctzombie_zz === defunctzombie_zz is now known as defunctzombie === defunctzombie is now known as defunctzombie_zz === mwhudson_ is now known as mwhudson === jam1 is now known as jam [12:00] What is JuJu? And how does it affect a Google App Engine developer? [12:00] And how does it speed up the development process? [12:17] amanSharma: It's a tool to deploy and configure services on clouds like EC2, OpenStack, MAAS [12:17] TheMue : not related to appengine? [12:17] amanSharma: The GAE is too different from those as it is more a development platform and runtime environment. [12:18] amanSharma: The others are IaaS, they provide a pure infrastructure. [12:20] amanSharma: There are charms for a large number of services. They describe how the software has to be deployed and how it can be connected and configured. === andreas__ is now known as ahasenack === cmagina-away is now known as cmagina [13:43] when deploying a django app with gunicorn, apache, and haproxy i cant figure out how to link gunicorn and apache together in a reverse proxy setup [13:44] apache talks about {{ haproxy_gunicorn }} and setting up that relation in website-relation-changed but when i deployed the template it isn't getting filled in [13:44] o/ [13:45] marcoceppi: hey there! :D [13:45] so im using the minimal amount of charms to try and get the proxy working [13:45] * marcoceppi pokes at the apache2 charm [13:46] right now im trying to figure out how to get apache2 to proxy to gunicorn [13:46] stokachu: do you actually need apache2? I know you mentioned static serving, but is it really needed? [13:46] i could force gunicorn to listen to port 80 [13:46] and bypass apache all together [13:47] stokachu: well, you don't need gunicorn to listen on port 89 [13:47] err 80, since haproxy can proxy to non port 80 web servers [13:47] ah ok [13:48] Ass long as website-relation-changed provides the hostname and port (typically `unit-get private-address` and the port number respectively) [13:48] then any charm that consumes the http interface will know how to talk to that charm [13:48] ah that may be my missing piece [13:49] right now it checks for /etc/gunicorn/${UNIT_NAME}.conf but that never gets generated [13:49] and i couldnt figure out how to generate that [13:50] Right, so you can deploy gunicorn on top of the apache2 charm (since it's a subordinate) and use gunicorn + apache2 [13:50] ah [13:50] lemme try that real quick [13:50] but you've already got gunicorn on top of your dashboard charm, and to be honest it's probably fine there [13:50] Here's an example website-relation hook: http://jujucharms.com/charms/precise/wordpress/hooks/website-relation-joined [13:51] ok lemme check that file [13:53] ok gonna try a few ideas [14:02] hello, is there any documentation for debugging "agent-state: down" ? [14:10] so i noticed u'haproxy_all_services': u'- {service_name: haproxy_service, service_port: 80}\n' which doesn't include my additional services [14:10] and apache2 template can't pick up the gunicorn settings [14:10] http://paste.ubuntu.com/5868173/ [14:11] that is my website-relation hook [14:13] http://paste.ubuntu.com/5868181/ and here is my status output [14:28] negronjl: you around? [14:36] according to haproxy i should be able to set all_services during a website-relation-changed but it isn't getting picked up when doing an add-relation apache2:reverseproxy haproxy:website [14:39] here is the template http://paste.ubuntu.com/5868244/ [14:42] stokachu: haproxy_gunicorn isn't going to be filled because gunicorn isn't deployed on apache [14:42] * marcoceppi pokes apache2 charm a little more [14:45] marcoceppi: if you're poking apache2, I was wondering if it would be a good idea to expose in the relation the SSL certificate it generated when SELFSIGNED was used [14:45] marcoceppi: it would help setting clients up to talk to that apache without having to use --no-check-certificate and stuff like that [14:47] ahasenack: I wouldn't think that you'd need to talk over SSL when "inside" the network though, right? [14:47] marcoceppi: it's used for testing, the "final" and "real" service would use a real cert [14:47] marcoceppi: so I need to use ssl [14:47] gotchya [14:48] in my case, I can't even disable the SSL check [14:48] so I have to copy the cert over to the clients or else it won't work [14:48] it wouldn't hurt to add the cert to be sent over the wire, you'd just have to make sure your charm knew to read the relation value [14:53] i used this to base how to setup gunicorn and my web app https://bazaar.launchpad.net/~charmers/charms/precise/apache2/trunk/view/head:/README.md#L41 [14:54] but i dont see anywhere in the logs where website-relation-joined gets called from my charm [15:13] hi everyone -- does anyone have a guide on deploying openstack HA w/ ceph w/ < 28 servers? (less than 16 is what i really need -- and i'd like multiple compute nodes as well) [15:14] also, w/ openstack HA, the compute nodes still use locally attached storage for the locally running VMs, correct? they're not using ceph? [15:15] marcoceppi: how would i deploy gunicorn on top of apache2? [15:18] i get error no matching endpoints when i do juju add-relation gunicorn apache:reverseproxy [15:18] stokachu: The same way you deployed it on to your dashboard, just juju deploy gunicorn apache2-gunicorn (so you have a different version of the subordinate) then add relation gunicorn_apache2 to apache2 [15:18] Are you trying to deploy the subordinate or trying to connect the two via relations? [15:19] i have gunicorn deployed just trying to add a relation [15:19] stokachu: gunicorn doesn't have an http interface, so you don't directly connect gunicorn to anything but the unit it's running on [15:20] you instead want to connect your dashboard, which has a website:http relation to apache2. Though I still don't think you need apache2 for this set up [15:21] how is juju 2.0 coming along? is there an official release date? [15:21] halloween? [15:22] TheChistoso|2: no official release date yet, that I know of, lots of progress being made with 1.X releases though! [15:23] marcoceppi: i just did add-relation haproxy seg-dashboard [15:23] and it finally pulled those variables in [15:23] \o/ [15:23] yayay [15:23] marcoceppi: thanks -- i'd really, really like to use maas+juju to deploy openstack HA but i really need to see something that uses < 28 machines. i have 16 available to me right now. [15:23] ok lemme see if i can get farther now [15:24] stokachu: good luck, we'll be here if you need more help [15:24] marcoceppi: thanks! im going to write a blog post about it once i get it all done [15:24] * stokachu spent 8 hours on this [15:24] can folks explain the ceph setup w/ openstack ha using juju? [15:25] TheChistoso|2: I'm not sure the state of the HA OpenStack charms, jamespage or adam_g might be able to provide more clarification on it [15:25] I know they've been doing a lot of work to make the OS charms a lot more robust [15:26] do you or anybody know how the OS charms configure the network interfaces? right now i only have 1 nic on each machine, but in production i'd have a few going to different switches -- so i'd need to configure provisioned machines accordingly... [15:27] is it assumed all machines are completely homogenous? [15:28] s/homogenous/homogeneous/g [15:28] TheChistoso|2: I don't personally know, hopefully one of the two I pinged above will pop in to answer [15:31] marcoceppi: thanks anyway :D [15:32] TheChistoso|2: from just poking at a few of the OS charms, it looks like most have configuration options for managing the networking [15:35] hey m_3_ around? [15:35] jcastro: yup [15:35] ssup? [15:35] I want to try to put the mac tarbal on github.com/juju/juju-core [15:35] in the releases section [15:35] k [15:36] but I want you around in case I clobber the repo [15:36] :) [15:36] which I am confident it won't do [15:36] np [15:36] hey so is this mirror the one off or is it auto updating? [15:36] jcastro: it'll be good to test if that change to the mirror target screws up the mirroring [15:36] it _should_ be auto-updating [15:36] but I haven't checked since yessterday [15:36] every 6 hrs [15:37] m_3_: there's a commit on there from 4 hours ago [15:37] so it would appear to be working [15:38] jcastro: http://paste.ubuntu.com/5868387/ [15:38] jcastro: first is an example of a barf... subsequent ones succeeded [15:38] timestamped [15:38] marcoceppi: ack, cool [15:39] https://github.com/juju/juju-core/releases/tag/1.11.2 [15:39] how's that? [15:39] jcastro: lemme kick it and see if it barfs [15:41] we should really mirror tags too [15:42] jcastro: that seemed to be ok [15:42] m_3_: +1 [15:42] m_3_: if you mirror tags though, jcastro's release will probably piss off the importer [15:42] marcoceppi: it tries, but can't catch all the tags b/c of tag formatting allowed in bzr -vs- git [15:43] (see that pastebin above for the warnings about tag conversion) [15:43] m_3_: I only see 0.x tags though, where are the 1.X tags? [15:43] jcastro: so that looks like it's keeping the releases in a bucket entirely independent of the source [15:43] jcastro: so subsequent pushes don't require pulls/merges first [15:44] m_3_: nice! [15:44] jcastro: so I think we're good if we just wanna keep it all on juju-core [15:44] we've gotta change the readme to explain how to get the mac client [15:44] and make it markdown :) [15:44] that's pretty much an ideal situation [15:44] jcastro: but that's gotta happen on lp... _not_ gh [15:45] marcoceppi: gh might take rst if we named it such... dunno [15:45] marcoceppi: it's easy enough to convert it to md [15:45] imo [15:45] :) [15:45] m_3_: yeah, you're right, if it's .rst it'll render [15:46] but, but, but, markdown :) [15:46] I know [15:47] ok so all we need to do is blow away github.com/juju-client [15:47] jcastro: shall I do that now? [15:47] also does anybody know the status of hpcloud az-2 for juju 1.11.2? [15:47] (tools) [15:47] not yet [15:47] let me announce first [15:47] jcastro: k [15:47] since I sent that URL to the list. [15:47] ack [15:47] m_3_: I created an az-3 bucket [15:48] and have 1.11.1 tools in it [15:48] Anybody else with a Mac want to test this? [15:48] marcoceppi: I'm working with hp peeps in az-2 in a couple of hours [15:48] marcoceppi: and gonna bail back to juju-0.7 I guess [15:48] m_3_: I'm not aware of any change there, you'll probably have to generate the image-metadata, etc === defunctzombie_zz is now known as defunctzombie [16:02] jcastro: we modeled the docs heavily after stripes documentation [16:02] evilnickveitch: hey so I think it's time to redirect juju.u.c/getting-started to new docs [16:02] that's how they organize "code examples" in different languages [16:02] http://i.imgur.com/1HAkvdd.png [16:02] figured we could do something similar when we have instructions for multiple versions of an OS [16:02] yeah, it is pretty easy really [16:02] and a good idea [16:03] ok show me how! [16:03] jcastro - yes please [16:03] evilnickveitch: I can push up a quick merge proposal with the CSS [16:03] Actually I was going to do that for you Jorge. [16:03] jcastro: * [16:03] Thanks man! [16:03] marcoceppi, do it [16:03] mgz: do you know what happened to alternatives in the current juju-core ppa? [16:03] hmm, better push my work today too [16:04] m_3: was never in the ppa, works fine in the distro release (saucy only atm) [16:04] as we have golang and mongo backport issues [16:04] (mongo now resolved) [16:04] mgz: right... trying to work with users on precise (using the ppa) [16:04] ah, gotcha [16:05] mgz: should I file a bug to get it in the ppa? [16:05] or was that explicitly decided? [16:05] fixing the ppa is doable now we've made it somewhat working again [16:05] ack... mgz thanks! [16:05] but remember it'll clash with pyjuju < 0.7 anyway [16:05] yup [16:05] so, they'd need the pyjuju paa too [16:05] *ppa [16:05] right... which was really the case [16:06] most peeps worked precise juju from the ppa [16:06] distro was so old for so long [16:07] and our docs always said "use the ppa" [16:08] evilnickveitch: related to the mailing list posts about consolidating those pages, I sort of moved metadata.yaml stuff over to one from the other already [16:08] sort of makes sense to get alternatives everywhere anyways... who knows what the transition from 1.xx to 2.xx will look like [16:10] jcastro, okay, I am rewriting them all anyhow [16:11] marcoceppi: lmk when you push and I'll start on the OSX instructions [16:12] jcastro: ack [16:12] going to lunch first [16:13] no worries [16:14] jcastro, re OSX default series question [16:15] it looks like the behavior is juju defaults to precise if nothing is entered. I think the same is for Ubuntu except I *think* juju will try to read what the client version is at [16:15] arosales: I was under the impression juju will _always_ default to LTS unless otherwise specified [16:15] mgz, can you confirm if no default series is present and a user is deploying from raring will juju try to deploy precise or raring? [16:15] but I could be wrong [16:16] marcoceppi, that very well be the case [16:17] jcastro, but I think I was incorrect in stating osx users need to specify precise. They would need to specify a default-series if they want something other than the latest LTS [16:19] which as marcoceppi said is the same behavior on Ubuntu [16:21] arosales: precise [16:22] mgz, thank you for confirming. [16:22] that behaviour did change :) [16:22] marcoceppi, no surprises you were correct :-) [16:22] mgz, thanks for helping try to save face lol :-) [16:24] jcastro, so aside from downloading the osx client from github the instructions are the same in set up. === dannf` is now known as dannf [16:26] jcastro, we may want to add the public bucket URL for Hp cloud in the instructions until we get simple streams fully implemented [16:26] thats applicable to all client [16:27] jcastro, ping me if you have any other questions on the osx instructions. [16:55] hi there [16:55] does someone know where is the Hangout for today's event? === hatch_ is now known as hatch [16:57] I am setting it up now [16:57] it'll be on ubuntuonair.com in a few minutes [16:59] jcastro, can you also post the URL for folks to join if they are interested in being an active participant? [16:59] yep [16:59] https://plus.google.com/hangouts/_/9a58ebb993347b7d0f87bbd2844942c3daffc22f?authuser=0&hl=en [16:59] jcastro, thanks [17:00] note ubuntuonair.com for viewing, above plus.google.com url for participating. [17:00] questions in here [17:00] mims had to reboot so waiting on him [17:06] jcastro: has it started? ubuntuonair.com doesn't have it yet [17:06] http://youtu.be/vPBrpMcXHN0 === andreas__ is now known as ahasenack [17:06] jcastro: why does it always lag behind? :) [17:06] does that work? [17:06] ok [17:07] jcastro: the youtube one does [17:07] jcastro, ya ubuntuonair not working for me atm either [17:07] checking [17:07] jcastro: youtube is ok, ubuntuonair.com has nothing [17:08] ok try now [17:08] jcastro: it's good now === defunctzombie is now known as defunctzombie_zz === defunctzombie_zz is now known as defunctzombie === defunctzombie is now known as defunctzombie_zz === bloodearnest__ is now known as bloodearnest [17:29] ok so ive taken out haproxy and setup apache -> gunicorn -> django app [17:30] http://paste.ubuntu.com/5868704/ - this is the relation setup === defunctzombie_zz is now known as defunctzombie [17:30] everything works except when i attempt to access login.launchpad.net through our openid setup it is using the unexposed proxy ip rather than the public apache ip [17:31] http://paste.ubuntu.com/5868709/ - this is my vhost file [17:31] when accessing thirdparty sites like launchpad am i missing a apache setting to force everything to use the server ip? [17:32] FYI, somebody should probably get juju-core into Debian. juju and juju-jitsu are in the process of being removed http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=712724 [17:33] Yeah I need to find a volunteer [17:33] SpamapS: but we have our hands full with the 12.04 backport [17:33] and also if my static files are on the proxy server how do i serve those out of apache [17:39] looks like ProxyPreserveHost worked [17:41] stokachu: Was that something you had to add to your vhost.tmpl or that you had to modify the charm to include? [17:44] i put directly in the vhost template [17:44] marcoceppi: http://paste.ubuntu.com/5868744/ [17:44] thats my latest [17:45] even with forced ssl openid still wants to callback to http [17:45] stokachu: right, so it wasn't nessiarily somethign that needs to be patched in the charm [17:45] that may be something with the application code itself [17:45] marcoceppi: oh nah [17:45] unless we wanted to add configurable proxy support in apache charm [17:46] but right now i just use enable_modules="proxy proxy_http" [17:46] and that stanza [17:46] stokachu: awesome [17:47] so can i create a custom site.yaml file and just include all the options for different charms? [17:48] and do i have to specify which charms go with which config option? [17:48] my django app charm has a config.yaml but i need one for deployment to cut back on juju commands i have to type [18:34] Juju deployed apache+gunicorn+django app over SSL, openid fails with No Mode | http://askubuntu.com/q/319382 [18:41] arosales jcastro about to commit this, any tweaks? [18:41] http://i.imgur.com/AwaZ20T.png [18:42] marcoceppi: hide windows for now pls [18:42] jcastro: ack [18:42] * arosales looking [18:43] jcastro: I'll leave it to you guys to write the mac osx stuff [18:43] yeah I just need the blank canvas [18:43] this is just a css skel/preview of what it'll look like [18:43] marcoceppi, looks good [18:43] marcoceppi, I take it we can edit that via wp admin [18:43] no it's on docs isn't it? [18:43] arosales: nope, it's on the docs, so you can edit it with the power of html [18:43] it's pretty straight forward though [18:44] I was just talking about just redirecting get-started to docs once we do this [18:44] marcoceppi, even better thanks [18:44] have the get started in one place instead of two [18:45] "Backports are usually enabled by default" really? [18:45] yeah [18:49] http://askubuntu.com/questions/319382/juju-deployed-apachegunicorndjango-app-over-ssl-openid-fails-with-no-mode [18:49] in before the bot [18:50] jcastro: heh, bot beat you by 15 minutes :) [18:53] i figure i'd post there for others to get some help too, but once i get this all working ill writeup a nice long tutorial [18:53] that would be awesome [18:53] ive spent 8 hours on this so far so for me personally i dont want to forget it [18:53] tbh this was my first task at juju [18:54] so ive come quite a ways [18:56] ".. punches hamsters", haha [18:58] LOL [18:58] i forgot about that === defunctzombie is now known as defunctzombie_zz [19:03] marcoceppi: ready to push those doc changes? [19:03] jcastro: just working out the javascript so you can actually click on the tabs [19:03] * jcastro nods [19:05] is juju 1.x still python? [19:05] this is the version iv been using 0.7+bzr628+bzr631~precise1 [19:06] that's the python version [19:06] 1.x and up are golang [19:06] ah ok [19:06] i heard 1.x doesn't support lxc yet? [19:07] no, next monthish? [19:07] ok cool [19:07] thats just what ive been using for testing etc [19:07] yeah [19:07] once we have containers it will be so nice [19:07] definately, makes it easy to develop apps with. before i was using varnish but this is too cool [19:09] you mean vagrant? [19:11] ah yea that other v one [19:14] jcastro: pushed [19:17] arosales: for the osx instructions [19:17] how do you untar something on OSX? [19:19] tar -xvf [19:19] unix linix pretty similar here [19:19] via the command line [19:20] I have problem with this guide [19:20] some of these aren't very black and white [19:20] guide == qa raiting [19:20] rating [19:20] it must be friday, brain is failing me [19:21] marcoceppi, there are some gray areas [19:21] marcoceppi, let me get you a link to existing rating reviews [19:22] also, some of these are pretty opinionated [19:22] like "Maintainable, easy to read and modify", if I find a charm in chef-solo I would answer no, but a charm in bash would be a yes [19:24] think more code comments there [19:24] ugh, I need to work on the wordpress charm [19:24] marcoceppi: are the new CSS/instruction things supposed to render locally? [19:25] don't work for me [19:25] don't know, haven't tried. I run a local apache instance when testing [19:25] ok [19:25] and it worked? [19:25] yeah, it doesn't work without a webserver [19:25] ok I'm done, can you test when I push to make sure? [19:26] probably why evilnick is having problems with ask ubuntu integration [19:26] jquery doesnt' load properly, and it's all down hill from there with jquery [19:26] ok [19:26] pushed, can you pull and see if it renders right? [19:27] ack [19:28] jcastro: only one bug http://i.imgur.com/ds6mmGW.png I'll patch it up though [19:29] marcoceppi: [19:29] needs a around juju instead of ` too [19:29] jcastro: also, you can use in liue of `` [19:29] heh [19:29] yup :D [19:35] I was actually planning on adding these tabs to begin with for the Amulet reference doc, so it's nice to get them out of the way now [19:44] arosales: do you remember when the docs regen? [19:47] jcastro, cron setup to run at: 6,18 30 * * * [19:49] darn, just missed the 18:30 utc window [19:58] marcoceppi: do you remember how we did redirects? [19:58] jcastro: I do [19:58] I copied the survey one and called it template-getstarted-redirect.php [19:58] Oh, you mean in WP [19:58] pushed, and then when IS updates, in WP .... ? [19:58] I just select that PHP as the template? [19:59] You just change the page's template to that [19:59] that *should* be it [19:59] ok === defunctzombie_zz is now known as defunctzombie [20:55] hi guys -- are the openstack guys around? [20:56] not that i know of [21:05] Not sure if anybody has had this issue before. I'm trying to deploy gunicorn on a local LXC environment, and it can't find endpoints. After checking my environment/configs, I decided to create a new user and try there, and deploying gunicorn worked perfectly. [21:06] Are there files I should remove to 'clean' my environment? I removed .juju/cache and even my local lxc machines === defunctzombie is now known as defunctzombie_zz === defunctzombie_zz is now known as defunctzombie === cmagina is now known as cmagina-away === defunctzombie is now known as defunctzombie_zz [23:40] marcoceppi, jcastro do you know of the github mirror goes both ways or just one way [23:40] ie do changes on the github side get mirrored to lp? [23:40] or just lp --> github ? [23:47] * arosales thinks the latter, but wanted to confirm