[00:05] <bodie_> anyone happen to know offhand whether there's an oh-my-zsh plugin for juju?
[00:06] <rick_h_> bodie_: http://askubuntu.com/questions/108689/how-can-i-get-tab-completion-with-juju-in-zsh
[00:06] <bodie_> I did see that, but I was thinking of a simple dropin for the plugins spec on oh-my-zsh (it's nice)
[00:07] <rick_h_> bodie_: heh, while I <3 zsh benji and I aren't big fans of oh-my-*
[00:07] <rick_h_> bodie_: so no official thing for that I know of
[00:08] <bodie_> it's a little bloaty, but the lazy in me approves
[00:08] <rick_h_> yea, I've had enough fun debugging broken things that I had no clue which thing did the breaking so vim, zsh, git, etc I go with my own config
[00:09] <bodie_> "oh-my-abbynormal"
[01:17] <bodie_> what's juju-mongodb?
[01:17] <bodie_> "mongodb for juju" doesn't exactly answer that question lol
[01:18] <sarnold> bodie_: iirc, it's a mongodb with the javascript engine removed
[01:18] <sarnold> don't ask me how that can work :)
[01:18] <bodie_> de juju
[01:18] <sarnold> but we're not interested in supporting a javascript engine for five years, hence this thing..
[01:18] <sarnold> .. so that juju can be supported in main
[01:19] <bodie_> nifty :)
[01:20] <marcoceppi> it's also mongodb with ssl enabled, which precise didn't have
[01:22] <jcastro> Just Enough Mongo
[01:22] <sarnold> really? why wasn't it enabled...?
[01:23] <jcastro> iirc it was the TLS licensing thing
[01:23]  * jcastro uses technical terms at this time of night
[01:25] <sarnold> jcastro: oh. that whole mess. ugh. :)
[01:37] <tvansteenburgh> jcastro, marcoceppi, et al: as someone getting ramped up on juju, i just want to say these Charm School videos are great, thanks for doing them!
[01:38] <jcastro> wow awesome!
[01:38] <jcastro> I am glad people are actually watching them!
[02:01] <davecheney> jcastro: yup, licencing
[02:02] <jcastro> I love how I have to give no details
[02:02] <jcastro> "the TLS thing"
[02:02] <jcastro> and everyone knows what that is
[02:02] <davecheney> jcastro: bad vibes, man
[02:02] <davecheney> although ironically we don't use TLS to talk to the mongodb server anymore
[02:02] <davecheney> all the clients talk to the api server
[02:02] <davecheney> which runs mongo locally
[02:02] <davecheney> it still talks TLS
[02:03] <davecheney> but this isn't as important
[02:03]  * davecheney waits for someone to raise a small pedantic exception to my prevoius statement
[02:07] <sarnold> hrm, the link to the copyright file from http://packages.ubuntu.com/trusty/juju-mongodb is broken
[02:13] <davecheney> sarnold: broken on every single package
[04:46] <hatch> Hey, is anyone working on an nginx charm now that it's going to be in Trusty?
[12:24] <rbasak> frankban: thanks for looking at that juju-quickstart bug. I have another issue with a test failure - are you able to look at this for me for the MIR, please?
[12:26] <frankban> rbasak: sure, what's the probem?
[12:28] <rbasak> frankban: http://paste.ubuntu.com/7084383/
[12:28] <rbasak> frankban: looks to me that the difference is ToMachineSpec
[12:29] <frankban> rbasak: yeah, this should be fixed in trunk
[12:29] <rbasak> frankban: do you know which rev, please, so I can cherry-pick for the distro package?
[12:29] <frankban> rbasak: what's the output of .venv/bin/pip freeze | grep jujuclient ?
[12:30] <rbasak> frankban: I also noticed that requirements.txt specifies jujuclient==0.11, but we have 0.17.5-0ubuntu1 in Trusty. I don't know if that's relevant.
[12:30] <rbasak> frankban: there is no virtualenv here. The test is running as part of the package build (an MIR requirement)
[12:30] <frankban> rbasak: OIC so that's relevant
[12:31] <rbasak> BTW, I'm on a bad connection. Workmen outside just removed my telegraph pole :(
[12:31] <rbasak> I may drop.
[12:31] <frankban> rbasak: http://bazaar.launchpad.net/~juju-gui/juju-quickstart/trunk/revision/57
[12:32] <frankban> rbasak: we plan to make a new pypi release asap, with updated dependencies and some fixes required by changes in juju-core
[12:32] <frankban> rbasak: we just need to wait for the new juju-core release
[12:33] <rbasak> frankban: OK, thanks. I'll work with a cherry-pick of that revision if that works, then, and we can update Trusty when you can do a release.
[12:33] <rbasak> frankban: please can you let me know when you do?
[12:33] <frankban> rbasak: sure, and thanks for your help
[12:35] <marcoceppi> Can you pass multiple units to juju remove-unit?
[12:50] <rbasak> frankban: test-requirements.pip was also missing, so I had to drop that part of the patch. I also dropped the bumping of quickstart.VERSION. Does that sound OK?
[12:51] <rick_h_> rbasak: he's run to lunch for a few min. I think that sounds ok
[12:51] <rbasak> BTW, I'm a bit dubious about calling the top level module "quickstart". Using /usr/lib/python2.7/dist-packages/quickstart/ seems ripe for collision.
[12:52] <mgz> marcoceppi: yes, you can pass N units
[12:52] <marcoceppi> mgz: ta
[12:53] <rbasak> rick_h_: thanks
[12:53] <rick_h_> rbasak: I can understand. In checking pypi there's not much on that space https://pypi.python.org/pypi?%3Aaction=search&term=quickstart&submit=search
[12:54] <rick_h_> rbasak: if it's something we need to adjust we can jujuquickstart it. I don't think quickstart will be used as a lib so
[12:54] <rick_h_> rbasak: but in dist-pacakges isn't it juju-quickstart?
[12:54] <rick_h_> ah, guess not
[12:56] <rbasak> rick_h_: I don't have a strong opinion. I don't think it should block the MIR. But it's something that I think I should flag with a Debian Python person, to avoid any potential problems in the future.
[12:57] <rick_h_> rbasak: definitely, and it's something we can adjust if required.
[12:57] <rick_h_> rbasak: appreciate you looking out for us :)
[13:07]  * rbasak tries to switch back to his main connection
[13:48] <zchander> ping frankban
[13:48] <frankban> zchander: hi
[13:50] <zchander> Hi, sorry I am bothering you again, but do you (or someone you know) have any knowhow to use 2 nics in a charm? I have just deployed an apache2 charm I am intending to test (yes, again ;) ) reverse proxying traffic from our school lan to the cluster network
[13:50] <zchander> e.g. how do I set up the second nic in the node to enable traffic from our school lan
[13:54] <frankban> zchander: I am not sure about that, I guess it depends on whether the apache charm supports that or not. You can investigate looking at the charm's docs, sshing into the deployed service node or asking the charm's maintainer
[13:54] <frankban> rick_h_: any ideas? ^^^
[13:55] <rick_h_> yea, I don't think many charms support dual nics natively because most cloud environments give you one nic
[13:55] <rick_h_> you might be able to write a second nic subordinate I'd guess. Anyone ever try that marcoceppi?
[14:16] <zchander> Mmmmm, so much for another brainwave :D Seems the only option that is remaining for me, is to deploy a second server which isn't part of MAAS, to act as a reverse proxy (??) Or are there any other ideas to accomplish that
[15:11] <jcastro> marcoceppi, ok how should I fix these bundles, reupload to new branches?
[15:11] <jcastro> or can we copy over the ~charmers ones?
[15:11] <marcoceppi> jcastro: we need to create new branches
[15:11] <marcoceppi> and then considate multiple bundles
[15:11] <marcoceppi> like the mediawiki ones
[15:12] <marcoceppi> using simple namespaces
[15:12] <marcoceppi> like mediawiki/simple instead of mediaiki-wiki/mediawiki-simple
[15:12] <rick_h_> marcoceppi: jcastro make sure to do a test one to verify how it looks and such
[15:12] <marcoceppi> get them all in a branch somewhere, point me at it and I'll express promulgate after proof
[15:13] <jcastro> not all in one branch though, you mean in their own individual service branches right?
[15:13] <marcoceppi> jcastro: one branch per bundle(s) for that service
[15:14] <marcoceppi> jcastro: right, so there would be a mediawiki branch with the two mediawiki bundles in the bundles.yaml file
[15:14] <jcastro> right, ack
[15:27] <jcastro> marcoceppi, ok, pushed hadoop
[15:29]  * marcoceppi looks
[15:30] <jcastro> for naming, how about "single" and "scalable"?
[15:30] <jcastro> for say, mediawiki
[15:30] <marcoceppi> jcastro: sounds good to me
[15:35] <jcastro> marcoceppi, mediawiki and mongo pushed
[15:35] <jcastro> that's all of them, all that's left is rails and django
[15:36] <marcoceppi> jcastro: I'll push mediawiki first
[15:36] <marcoceppi> to see how it renders
[15:36]  * jcastro nods
[15:36] <marcoceppi> we'll see it in about 10 mins
[15:46] <marcoceppi> jcastro: what's the comming-soon address for jujucharms?
[15:46] <marcoceppi> apparently, I'm a little M friendly today
[15:47] <marcoceppi> jcastro rick_h_ looks awesome in the gui
[15:47] <rick_h_> marcoceppi: woot! and should make nicer quickstart commands
[15:49] <marcoceppi> jcastro: http://i.imgur.com/BIbDB3j.png
[15:49] <jcastro> marcoceppi, comingsoon.jujucharms.com
[15:50] <jcastro> marcoceppi, ok so now to remove old and busted .... ?
[15:50] <marcoceppi> jcastro: we can't yet
[15:50]  * rick_h_ hides
[15:50] <marcoceppi> not until the sometime next week
[15:50] <jcastro> oh, until rick, got it
[15:50] <marcoceppi> yeah, otherwise I'll start pushing the other new hotness up
[15:51] <jcastro> keep on pushing
[15:53] <jcastro> https://plus.google.com/hangouts/_/hoaevent/AP36tYdEL_9S543oWBGGJFmgI-6Vzdu2cLUpciThZmLeVlDfhr5Uhw?authuser=0&hl=en
[15:54] <jcastro> Juju Eco Status session will be at that URL ^^ if you want to find out what's happening
[15:54] <jcastro> we'll go  live in 6 minutes
[15:56] <jcastro> we're in the spill over room too, so you can ask questions here or in #ubuntu-uds-hallway
[16:44] <ryanlee> Hello all!  Anyone know what if there is a way to use Juju to deploy multiple Wordpress sites on the same server (using Virtual Hosts or Server blocks)?  Is this in the works?
[16:46] <mgz> ryanlee: if the charm doesn't support that, you could change the wordpress charm itself to allow for it
[16:47] <ryanlee> mgz, yeah, I actually was planning on contributing to it, but wanted to see if someone already had done something to make this possible.
[18:02] <jose> jcastro: can we host future ones at ubuntuonair? just to keep track of all of them in one place
[18:02] <jcastro> yeah sorry I just didn't have time
[18:02] <jcastro> we'll do tomorrow's on UOA
[18:06] <Fishy_> hi!
[18:07] <marcoceppi> Fishy_: o/
[18:10] <Fishy_> So looking at using juju, have a Q.  Say I have an app Foo i want to deploy. So I create a Foo charm.  The problem is, Foo needs java installed to work. Do I  a) Create a subordinate service of Java that Foo depends on?  b) Install java in my gems hooks   c) ????
[18:10] <mgz> Fishy_: most simply, your foo charm installs java in your install hook
[18:10] <Fishy_> I could easily put 3 services on 1 machine, that all need Java (or some common thing) installed to run
[18:11] <mgz> ubuntu/debian has package management for a reason
[18:11] <mgz> though, how well java plays with that has long been open to question
[18:11] <Fishy_> ya that too
[18:12] <Fishy_> but also:  say I have 20 instances of Foo around..  I want to upgrade java 1.7.x to 1.7.y.. but i want to roll it out slowly, to make sure it works
[18:12] <Fishy_> the hook way seems not helpful for that
[18:12] <Fishy_> but if i wrapped java in a charm
[18:12] <Fishy_> could roll it out
[18:13] <mgz> that's still perverse, even with this scenario
[18:13] <Fishy_> i guess if I just made a new Foo Charm and did an upgrade, the hook could upgrade it
[18:13] <mgz> java is not a charm it's a bit of infrastructure
[18:13] <Fishy_> ok so java dont make it a charm
[18:14] <mgz> the way people have tended to to rolling upgrades like that is have a config value that can be `juju set` on the service to pick a subset of units to upgrade
[18:14] <Fishy_> do I also need something like puppet to manage the underlying OS stuff like java ?
[18:14] <Fishy_> im trying to avoid puppet ;)
[18:15] <Fishy_> Ok so java.version a config value to my Foo charm.. then I can slowly upgrade the config value and roll out the new Java
[18:15] <mgz> no, you don't need it, but you can benefit from some management stuff off the side of a charm
[18:15] <mgz> Fishy_: that kind of thing
[18:16] <mgz> to do rolling upgrades you need more than just a version, because you want non-homogenous units
[18:16] <Fishy_> yah thats true
[18:16] <Fishy_> id rather tie version to the actual app not dependencies
[18:16] <Fishy_> but config seems to do that
[18:17] <mgz> a set of unit ids to upgrade would do
[18:17] <Fishy_> but 3 apps on the server, may end up with 3 different java.version variables i would need to have
[18:17] <jcastro> http://askubuntu.com/questions/433842/how-to-resolve-error-machine-is-already-provisioned-in-manual-provision-set-up
[18:17] <mgz> then the roll back is leave the same set, and revert the version config change
[18:17] <jcastro> any ideas here?
[18:17] <mgz> and a rollout is add all remaning unit ids to the config
[18:17] <Fishy_> i like
[18:18] <jcastro> Fishy_, so how I would do it
[18:18] <jcastro> is since deploying instances is an easy operation with juju I would do a java upgrade test in a seperate environment
[18:19] <jcastro> rather than doing it instance per instance in an existing environment
[18:19] <Fishy_> we can't big bang to prod sadly
[18:19]  * jcastro nods
[18:19] <jcastro> yeah so you could make it part of the charm and set the java version via config
[18:19] <Fishy_> see http://en.wikipedia.org/wiki/Knight_Capital_Group
[18:19] <Fishy_> for cost of failure ;)
[18:19] <jcastro> heh, that is awesome man
[18:20] <Fishy_> Ok so then the choice is does the charm hook use apt-get, or does it install a local java
[18:20] <Fishy_> local java lets us have 3 apps with diff javas
[18:20] <Fishy_> but is "wasteful"
[18:21] <jcastro> a subordinate could also work
[18:21] <mgz> you likely don't want your apps in the same charm or smashed as services on the same machine anyway
[18:21] <jcastro> but like are you apt-get installing your java currently or do you do the whole oracle manual install stuff?
[18:22] <Fishy_> right now we have a mix of 6 versions of se linux and 3 versions of ubuntu and a puppet server that is berserk and a bunch of hand configs
[18:22] <Fishy_> then the IT guy quit
[18:22] <Fishy_> i am trying to start fresh... thinking of a POC using juju to manage everything
[18:23] <Fishy_> ya in most cases I see 1 app to 1 VM or 1 MAAS
[18:23] <Fishy_> in which case I can just use apt-get like a sane person
[18:23] <jcastro> yeah, in that case, it's just part of the install hook
[18:23] <mgz> that sounds... much nicer
[18:24] <Fishy_> would you guys reco juju on 12.04 LTS or 13?
[18:24] <Fishy_> I figured deployment is cheap so we don't NEED to use LTS
[18:24] <Fishy_> if we automate things right
[18:24] <jcastro> LTS
[18:24] <jcastro> all our charms and stuff all do LTS
[18:24] <Fishy_> 14 is soon I know
[18:24] <Fishy_> ok, LTS it is
[18:25] <jcastro> yeah, but we're not even at 12.04's half way point in support, I'd base on 12.04, get the important things you mentioned in order
[18:25] <jcastro> and target 14.04 like in the fall or something
[18:25] <jcastro> we provide PPAs and updates back to 12.04 so you won't lose anything over choosing 13.x
[18:26] <mgz> right, heading home time for me
[18:26] <Fishy_> ya
[18:26] <Fishy_> some things 12 may not have that 14 will
[18:26] <Fishy_> tradeoffs..
[18:33]  * Makyo is away: Lunch
[18:40]  * Makyo is back (gone 00:07:43)
[18:48] <ev> is it possible to tell juju about an apt proxy? Cloud init lets you, but it doesn't seem possible to pass that through.
[18:49] <marcoceppi> ev: it does it automatically for local lxc deployments
[18:49] <marcoceppi> ev: not sure about other deployments
[18:50] <ev> marcoceppi: it's for a ctslab openstack deployment
[18:51] <jcastro> ev, there's a setting you put in environments.yaml that lets you define a proxy
[18:51] <jcastro> I am looking for it
[18:53] <ev> apt-http-proxy?
[18:57] <jcastro> 	"apt-http-proxy": "value-set"
[18:57] <jcastro> that seems to be it
[19:05] <jcastro> ev, I filed  a bug for thumper to document the value, I've CCed you on it
[19:05] <jcastro> he usually comes online in about an hour
[19:07] <ev> thanks
[19:46] <jcastro> marcoceppi, I pushed the rails bundle up even though it's not passing proof
[19:46] <jcastro> I wanted to get the readme and combining the 2 existing ones out of the way first
[19:52] <rick_h_> jcastro: we've got the RT in progress to deploy
[19:52] <rick_h_> jcastro: so hopefully have that ready in minutes vs hours now
[19:52] <jcastro> excellent
[19:52] <rick_h_> jcastro: actually, I'm told it's deployed
[19:52] <rick_h_> jcastro: so try running charm proof on it and should work
[19:52] <jcastro> I have the django one almost done, so really the only one we have to fix after is rails
[19:52] <jcastro> <3
[19:53] <jcastro> I am confused, how would running charm proof on it work now?
[19:53] <rick_h_> jcastro: because charm proof asks charmworld for help in validating
[19:53] <jcastro> oh!
[19:53] <rick_h_> jcastro: and that's been deployed with a new release
[19:54] <jcastro> you sir, are correct!
[19:54] <jcastro> passed first time!
[19:54] <lazyPower> boom!
[19:54] <jcastro> marcoceppi, check out the rails bundle I just pushed please
[19:54] <lazyPower> jcastro: he's oot, i'll +1 it
[19:54] <jcastro> you'll have django in a minute
[19:54] <jcastro> k
[19:55] <bodie_> I'm curious how much charm makers follow DigitalOcean docs :P
[19:55] <bodie_> I have a couple of buddies who do those full-time, so it might be fun to get you guys together
[19:55] <jcastro> hook me up! send them to me
[19:55] <jcastro> we work on DO no
[19:55] <jcastro> err, now
[19:55] <bodie_> yeah, I saw :D
[19:56] <lazyPower> jcastro: link to branch with bundle? i dont see it in the revq
[19:57] <jcastro> https://code.launchpad.net/~jorge/charms/bundles/rails/bundle
[19:57] <jcastro> I didn't file a bug yet
[19:57] <jcastro> lazyPower, I am reasonably certain the "sample-app" won't work ootb
[19:57] <lazyPower> ack
[19:58] <lazyPower> jcastro: is this going to be our common convention for bundles moving forward so i dont ask you this over and over again?
[19:58] <lazyPower> charms/bundles/<charm>/bundle?
[19:58] <jcastro> yeah, it's in the bundles doc page
[19:58] <lazyPower> ah ok
[19:58] <lazyPower> i haven't looked that over in a while.
[19:58] <jcastro> and then the name is in the bundle itself
[19:59] <lazyPower> ty
[19:59] <jcastro> so this one will be
[19:59] <jcastro> juju-quickstart bundle:rails/single
[19:59] <jcastro> juju-quickstart bundle:rails/scalable
[20:01] <lazyPower> jcastro: nix juju-gui
[20:01] <lazyPower> you left it in the bundle and quickstarts freaking out
[20:02] <jcastro> ah! good catch
[20:02] <jcastro> lazyPower, fixed, pushed
[20:03] <lazyPower> pulled & running
[20:04] <lazyPower> rick_h_: love the quickstart workflow btw <3 you guys are made of awesome
[20:05] <rick_h_> lazyPower: very cool, buy frankban a beverage next sprint. It's 99% his awesomeness
[20:05]  * lazyPower readies his credit card for rounds
[20:06] <jcastro> frankban truly crushed it this cycle
[20:06] <jcastro> lazyPower, django pushed, filing the bugs now to get them in the queue
[20:07] <jcastro> rick_h_, notice how I am telling people the URL stucture before you write it? That forces you to fix the ~charmers thing, mwahahah.
[20:07] <rick_h_> jcastro: card is on the board and frankban started working on it today :P
[20:07] <lazyPower> sneaky sneaky
[20:08] <rick_h_> jcastro: jujucharms.com updated and shoiuld show recommended bundles now
[20:09] <zchander> lazyPower, marcoceppi: thanks for all your help! I have just sent out a mail to the mail list with my question. I'll continue my search and (hopefully) will be able to provide us all with more info on my quest
[20:10] <lazyPower> zchander: i saw that update. If you have any further questions dont hesitate to ask. We'll do everything we can to help you out.
[20:10]  * zchander forgot to also mention frankban *shame on me*
[20:11] <zchander> lazyPower: Maybe I should go the 'easy' way and use a separate reverse proxy server, which won't be part of the MAAS/Juju environment
[20:11] <jcastro> rick_h_, well done sir, I was expecting to be up all night and here we are with 45 minutes to spare.
[20:12] <rick_h_> jcastro: :) we do our best to last-minute it as much as possible :P
[20:12]  * zchander is going to walk the dog and relax (playing BF4)
[20:12] <jcastro> ok so all that's left is pushing these and then tomorrow I'll feature them
[20:14] <jcastro> rick_h_, hey so ... the ability to feature something is on manage.j.c
[20:14] <jcastro> which only shows charms
[20:14] <jcastro> so how do I feature a bundle?
[20:22] <rick_h_> jcastro: should be the same checkbox like a charm
[20:22] <rick_h_> jcastro: on mjc
[20:25] <jcastro> yeah but how do I see the bundles? it only shows charms afaict
[20:25] <rick_h_> jcastro: search for bundle in the site search
[20:25] <rick_h_> http://manage.jujucharms.com/search?search_text=bundle&op=
[20:25] <jcastro> dude
[20:26] <jcastro> that is way better than jujucharms.com searching for bundles
[20:28] <jcastro> rick_h_, nope, no link to adding featured in a bundle
[20:28] <jcastro> tried to URL hack it too, no dice
[20:30] <rick_h_> jcastro: logged in?
[20:30] <jcastro> yep
[20:30] <jcastro> I can get to the charm ones no problem
[20:30] <rick_h_> jcastro: k, otp, will look in a sec
[20:42] <rick_h_> bac: do you recall how the bundle featuring was meant to work? I don't see links in the bundle template.
[20:43] <bac> rick_h_: a bundle can be featured via the admin gui on mjc if it is promulgated
[20:43] <rick_h_> bac: yea, but it's not showing and there's no <a> link in the bundle.pt like the charm.pt
[20:45] <rick_h_> oh, got a url that 403s on me
[20:45] <rick_h_> jcastro: try http://manage.jujucharms.com/bundle/mediawiki/single/featured/edit
[20:46] <rick_h_> jcastro: k yea that works for me ^
[20:50] <bac> rick_h_: http://manage.jujucharms.com/bundle/mongodb-cluster/mongodb-cluster/featured/edit
[20:51] <bac> rick_h_: sorry, i was trying to figure it out.  we need to add the link on the promulgated bundles.
[20:51] <rick_h_> bac: yep, added a card thanks
[21:02] <lazyPower> jcastro: so, i found the other issue
[21:02] <lazyPower> deployer isn't interpreting this correctly, its trying to deploy subordinate services to units - which is causing the issue.
[21:03] <lazyPower> jcastro: http://i.imgur.com/BIhknum.png
[21:03] <lazyPower> thats on rails-complex
[21:03] <hazmat> lazyPower, hmm
[21:04] <hazmat> lazyPower, config?
[21:04] <lazyPower> hazmat: http://bazaar.launchpad.net/~jorge/charms/bundles/rails/bundle/view/head:/bundles.yaml
[21:04] <hazmat> lazyPower, shouldn't do that there are lots of extant deploys with subordinates..
[21:05] <lazyPower> yeah, i'm looking at this, and thinking i may start deleting charms from the bundle until it deploys, and slowly re-add them back to find where its going awry
[21:05] <hazmat> lazyPower, ie .. that assumption isn't correct.
[21:05] <hazmat> afaics
[21:05] <lazyPower> well, its strange that was the error output. it may be something else causing the issue and bubbling that error message back up the stack
[21:06] <marcoceppi> lazyPower: is there a num_units key on kibana?
[21:06] <marcoceppi> if so, that's wrong
[21:06] <lazyPower> marcoceppi: ther eis
[21:06] <marcoceppi> err, logstash-agent
[21:06] <marcoceppi> the subordinate in other words
[21:06] <lazyPower> there is not on the logstash-agent
[21:07] <marcoceppi> lazyPower: the bundle looks good to me
[21:07] <lazyPower> it looks fine to me too, but i cant get it to deploy on hpcloud or local
[21:07] <marcoceppi> lazyPower: via gui or deployer or quickstart?
[21:08] <lazyPower> quickstart
[21:08] <marcoceppi> huh
[21:09] <lazyPower> also when dragging/dropping on the gui, just finished confirming that
[21:09] <lazyPower> let me move to my laptop
[21:09] <lazyPower> my environment may be wonky
[21:11] <lazyPower> ok trying again - 1 sec while i fetch the bundle
[21:18] <hazmat> marcoceppi, all roads there end at deployer
[21:18] <marcoceppi> hazmat: good point
[21:18] <hazmat> marcoceppi, num_units gets a log warning from deployer, but it ignores
[21:18] <lazyPower> jcastro: as an aside, the readme is full of lies
[21:18] <hazmat> marcoceppi, on  a subordinate
[21:18] <lazyPower> the bundle name is rails-complex, not rails-scalable
[21:19] <lazyPower> well, not lies, but its not correct :D
[21:19] <marcoceppi> lazyPower: the name is all wrong to begin with
[21:19]  * lazyPower pokes at the revq to see if his bugs are open yet
[21:21] <lazyPower> oo its in here, ok doing the writeup
[21:21] <lazyPower> thanks for the responses gentlemen o/
[21:21] <hazmat> lazyPower, what's the error output?
[21:22] <lazyPower> hazmat: do you want the info from the gui or the machine log?
[21:23] <hazmat> lazyPower, either or
[21:23] <hazmat> lazyPower, how bout both
[21:23] <lazyPower> hazmat: the GUI error is http://i.imgur.com/BIhknum.png
[21:23] <lazyPower> let me fetch the machine log, 1 sec
[21:23] <hazmat> lazyPower, okay the other one
[21:23]  * hazmat src dives into deployer
[21:24] <lazyPower> gah i destroyed env since that screenshot, let me run this again
[21:24] <lazyPower> sorry 1 moment
[21:26] <hazmat> hmm
[21:26] <hazmat> lazyPower, which version of core?
[21:27] <lazyPower> 1.17.4-saucy-amd64
[21:27] <hazmat> lazyPower, it does look like num_unit carries through on deploy
[21:28] <hazmat> which it really shouldn't on a subordinate, but that's the first time its been an issue.. num_units is always 1 on deploy.
[21:28] <hazmat> and deployer adds units subsequently which skips out on subordinates
[21:28] <lazyPower> ok i'm getting the same thing on 12.04 too - let me pastebin these logs
[21:29] <lazyPower> http://paste.ubuntu.com/7087033/
[21:30] <lazyPower> Saucy ^
[21:30] <lazyPower> precise > http://paste.ubuntu.com/7087044/
[21:37] <mbruzek> lazyPower asked me to deploy this bundle and I am also seeing this bundle fail, but it appears mine will not connect to the GUI.
[21:37] <mbruzek> http://pastebin.ubuntu.com/7087061/  <== I am using hpcloud
[21:38] <mbruzek> juju-gui unit shows as started and exposed
[21:39] <mbruzek> But I am unable to raise it either on port 80 or 443
[21:40] <mbruzek> Hey marcoceppi would this hp problem be related to the security groups being "full"?  Don't you have a script that cleans that up?
[21:40] <marcoceppi> mbruzek: probably
[21:41] <marcoceppi> I'm going to tourch the hp account, one min
[21:41] <mbruzek> should I destroy-environment?
[21:42]  * mbruzek thinks that marcoceppi should explain the magic he is doing and when to know to do it.
[21:42] <marcoceppi> mbruzek: yeah, go ahead and destroy
[21:43] <mbruzek> done
[21:43] <mbruzek> How can you tell if they are full?
[21:43] <marcoceppi> mbruzek: I'm just running a little go program that uses the nova API to delete security groups
[21:43] <lazyPower> oh man i never thought of that
[21:43] <lazyPower> the sec groups being full
[21:43] <marcoceppi> I mean, I don't really know, I just run it occasionally
[21:43]  * lazyPower hat tips
[21:43] <marcoceppi> juju is getting pretty good at cleaning up these days though
[21:43] <lazyPower> orly?
[21:43] <mbruzek> lazyPower, I only *just* thought of that.
[21:44] <lazyPower> mbruzek: that was faster than I came up with it :D
[21:44] <mbruzek> marcoceppi, does your magic script tell us how many it deletes?
[21:44] <marcoceppi> yes
[21:44] <mbruzek> Would that tell us if it is full?
[21:44] <lazyPower> hazmat: want me to pack the rest of the logs up and ship them or are you sated with just machine-0 logs?
[21:44] <marcoceppi> mbruzek: it'll tell us how many it deleted
[21:44] <mbruzek> For example if 2048 security groups deleted then it was full
[21:45] <hazmat> lazyPower, hmm.. what i need is the gui unit logs
[21:45] <marcoceppi> mbruzek: we have a very finite limit
[21:45] <marcoceppi> I think 75
[21:45] <lazyPower> hazmat: ack, inc
[21:45] <mbruzek> or some high number, but if only 5 were deleted then that was likely not the problem.
[21:45] <marcoceppi> haha, yeah 70 were deleted
[21:45] <marcoceppi> mbruzek: try to bootstrap now
[21:46] <lazyPower> hazmat: http://paste.ubuntu.com/7087105/
[21:46] <mbruzek> deploying that bundle
[21:48] <mbruzek> bootstrapping
[21:50] <mbruzek> OK I can now see the GUI
[21:50] <mbruzek> That must have been it, and 70 must be our limit
[21:51] <lazyPower> interesting
[21:56] <marcoceppi> I'll talk to HP to get it increased, so annoying
[21:56] <mbruzek> Perhaps marco should share the script too
[21:57] <mbruzek> Looks to me like Quickstart thinks it deployed, but I only see 3 things on my GUI.
[21:57] <mbruzek> http://pastebin.ubuntu.com/7087146/
[21:58] <mbruzek> Juju-GUI, kibana, and haproxy
[21:58] <mbruzek> Isn't there supposed to be a rails charm in there somewhere?
[22:00] <mbruzek> yeah I only see three charms deployed
[22:00] <mbruzek> http://pastebin.ubuntu.com/7087161/
[22:14] <lazyPower> thanks for that deploy test mbruzek. You've validated what I found and outlined in the review
[22:27] <jcastro> lazyPower, whoa! I appear to have pushed the wrong bundle to rails
[22:27] <jcastro> fixing
[22:27] <jcastro> it's supposed to be single and scalable
[22:28] <lazyPower> jcastro: have you tried to deploy the django bundle?
[22:28] <lazyPower> i'm getting the same error output ont rying to deploy the bundles about subordinate services not being able to deploy with units.
[22:28] <lazyPower> so i'm not convinced this is a bundle issue as much as i'm wondering whats going awry causing this to be such a p
[22:28] <lazyPower> ita
[22:29] <jcastro> we haven't touched the charm itself in a long time
[22:29] <jcastro> ok found the issue, pushed new rails
[22:29] <lazyPower> ty - pulling update
[22:29] <jcastro> it is no longer full of lies
[22:30] <lazyPower> haha
[22:30] <lazyPower> <3
[22:30] <jcastro> ... and left in juju gui again
[22:30] <lazyPower> jcastro: wipe juju-gui from scalable bundle
[22:30] <lazyPower> yeah
[22:31] <lazyPower> i want to stamp these as good but with them not deploying - i cant stamp them yet :( I'm sorry man
[22:31] <lazyPower> mbruzek tried too and couldn't get them to deploy
[22:31] <jcastro> yeah that's fine
[22:31] <jcastro> I just wanted to get them in line instead of blocking
[22:31] <jcastro> fixing them is no problemo!
[22:32] <lazyPower> mission accomplished
[22:32] <jcastro> pushed again
[22:32] <lazyPower> well, phase 1 anyway
[22:32] <jcastro> I'm not even sure pavel's sample-app for rails works
[22:33] <lazyPower> jcastro: a good app to do this with would be stringer.
[22:33] <lazyPower> its a RSS reader, and an excellent test bed
[22:33] <jcastro> ok so right now the rails charm is using a sample github url
[22:33] <jcastro> if you think we can just put stringer in there and that will work then nod
[22:34] <lazyPower> ack. when i hear back from hazmat that he's got something for me to do differently, i'll give it another go, and if it fails i'll stick in a new rails app for ya
[22:34]  * lazyPower EOD's
[22:34] <jcastro> it would not surprise me if either django or rails is broken
[22:35] <jcastro> neither has tests and hasn't been touched in a while
[22:35] <lazyPower> its not the charms
[22:35] <jcastro> wait huh?
[22:35] <lazyPower> did you see the screenie i posted ealier?
[22:35] <lazyPower> http://i.imgur.com/BIhknum.png
[22:35] <jcastro> huh what is that
[22:35] <lazyPower> the craziest thing i've ever seen in my life
[22:35] <jcastro> I didn't have that error
[22:36] <lazyPower> and i've encountered it on 3 machines so far
[22:36] <hazmat> lazyPower, oh.. i thought you resolved the issue to limits
[22:36]  * jcastro goes to EOD, we'll find the issue in the manana
[22:36] <lazyPower> agreed
[22:36] <lazyPower> hazmat: negative, that was with the GUI being unreachable
[22:36]  * hazmat follows the path of wisdom
[22:36] <lazyPower> the subordinate error message is still standing
[22:36] <hazmat> EOD ftw
[22:36] <lazyPower> haha :)
[22:37] <hazmat> marcoceppi, i'm walking to lost dogs .. if you've got any interest.. dial me.
[22:37] <marcoceppi> hazmat: I'm down, meet you there in 5 mins?
[22:37] <hazmat> eta 45m
[22:37] <marcoceppi> oh, see you in 45m
[22:37] <hazmat> marcoceppi, walking.. its a ways.
[22:37] <marcoceppi> gotchya, yeah I'll be over in about 30-45m
[22:37] <jcastro> sweet, we'll EOD, you guys go to the bar and fix the bundles
[22:37] <jcastro> sounds like a plan
[22:38] <rick_h_> lol
[22:38] <rick_h_> good try jcastro