[00:25] <lazyPower> jcastro: 1.14.1-raring-amd64
[01:26] <_mup_> Bug #1233924 was filed: cannot bootstrap azure because no OS image found <azure> <bootstrap> <juju-1.15.0> <juju:Triaged> <https://launchpad.net/bugs/1233924>
[01:26] <ubot5`> Ubuntu bug 1233924 in juju "cannot bootstrap azure because no OS image found" [Critical,Triaged]
[01:26] <_mup_> Bug #1233924: cannot bootstrap azure because no OS image found <azure> <bootstrap> <juju-1.15.0> <juju:Triaged> <https://launchpad.net/bugs/1233924>
[01:46] <mhall119> so my juju-status says I don't have a bootstrapped environment, but ps says it's running local instances anyway: http://paste.ubuntu.com/6182048/
[01:46] <mhall119> can somebody tell me how to clean up after juju?
[01:52] <omgponies> does juju expose details like hostname, ip address  to the local hooks ?    or should I call out to `hostname` etc
[01:57] <davecheney> omgponies: two secs
[01:57] <davecheney> omgponies: use unit-get
[01:57] <davecheney> that is the recommended way
[01:58] <davecheney> so
[01:58] <davecheney> unit-get private-address
[01:58] <davecheney> unit-get public-address
[02:01] <omgponies> sweet
[02:01] <omgponies> thx
[02:02] <davecheney> AWS Y U SO SLOW!?
[02:03] <omgponies> slow is relative ...  used to be we were thankful at a 4 week turnaround for a new server :)
[02:03] <davecheney> omgponies: so relative to other cloud providers ...
[02:04] <omgponies> go spend some time on azure ;)
[02:05] <davecheney> omgponies: touché
[02:07] <lifeless> PONIES
[02:27] <omgponies> linux why!
[02:28] <omgponies> oh wait ... because I told it wrong
[02:28] <omgponies> me why!
[06:00] <AskUbuntu> Problem with IP Openstack MAAS | http://askubuntu.com/q/352570
[06:00] <AskUbuntu> MAAS login error | http://askubuntu.com/q/352573
[09:31] <mattyw> is there a way I can run juju debug-hooks on a unit that has deployed succesfully - the aim is not to debug the hooks - but rather to enter an interactive session with a hook environment, so I can see the env flags that get defined and I can call things like relation-get/ unit-get
[11:07] <joachimhs> I have a set of servers in a datacenter. ON some of these machines I want to be able to separate out different applications. I have thought about using KVM, or Docker. How would do Ubuntu Juju running on the local machine on LXC compare with those two?
[11:07] <joachimhs> ^ I will mostly deploy custom-made software, as well as web apps via Nginx
[11:50] <marcoceppi> mattyw: launch debug-hooks, run `juju set` to change a config value and it'll catch the config-changed hook
[12:16] <joachimhs> I would really appreciate any comments on the above question :) ^
[12:18] <mattyw> marcoceppi, thanks, I was actually looking for a generic way of being able to interact with a hooks environment rather than changing config-settings
[12:18] <marcoceppi> mattyw: that's the only way atm
[12:18] <mattyw> marcoceppi, ok, thanks
[12:44] <joachimhs> If one of my charms require a JVM. Is that installed for the host-os, or separate for each of the charms ?
[12:51] <marcoceppi> joachimhs: so, for your first question, we have container support landing in Juju soon. It's currently there but pretty rough cut. The idea being you can safely isolate multiple charms on a single machine and still perform orchestration on them
[12:52] <marcoceppi> joachimhs: as for JVM, that would need to by the charm. The base operating system is the same for all charms. You use the charms to build/configure the OS to meet the needs of your service
[12:53] <joachimhs> marcoceppi: Thanks for your answers! So this means that each physical server will have one JVM installed, and that each charm will configure the java properties to match its requirements?
[12:54] <marcoceppi> joachimhs: that, again, depends entirely on you as the charm author
[12:54] <joachimhs> marcoceppi: OK. So each charm can have its own JVM (or any other installed software) ?
[12:54] <marcoceppi> joachimhs: each charm can do whatever it wants. It's executed as root on the box
[12:55] <joachimhs> marcoceppi: OK.
[12:55] <marcoceppi> I don't manage or set up JVMs myself, so I can't speak to the limitations of that product, but if you can have multiple JVMs per physical machine, then you can do that in a charm
[12:57] <joachimhs> marcoceppi: OK. You mentioned that container support isn't landed yet.. Is it preferred to run on top of KVM, or another hypervisor or is "juju switch local" a viable option?
[13:01] <marcoceppi> joachimhs: we're talking about two different things. The Local provider works really well, it allows you to spin up a "cloud-like" environment on your local machine using LXC. This is meant as a development tool. container support is the co-location of two or more charms on a single physical machine. The current juju model is once unit of a service per machine. So if you deploy mysql you'll get a single unit of machine on a single
[13:01] <marcoceppi> machine and that's that. You wouldn't nessiarily (and safely) be able to put a unit of mediawiki on that same node. It's designed to be used just for MySQL. with container support coming to do Juju you can deploy a unit of a service in an LXC container on a physical node. Thereby you can employ density on larger hardware by running services in containers on fewer physical machines. So say co-locating a unit of mediawiki and mysql on
[13:01] <marcoceppi> the same physical hardware. Since they're in containers they can continue to "own" the machine but now share that one parent node. This whole container setup will be virtually transparent to you as the end user, you would just tell juju you want to use containers and on what machine to put them
[13:03] <joachimhs> marcoceppi: OK. That makes sense
[13:04] <joachimhs> marcoceppi: Thanks!
[15:24] <vds> trying to debug a hook, how do I land on the target machine so I can execute commands like realtion-set or unit-get?
[15:45] <marcoceppi> vds: use `juju debug-hooks <unit>`
[15:46] <marcoceppi> vds: that will trap hooks prior to execution, so if you have a hook in an error state and want to run it again, launch debug-hooks and then in another window type `juju resolved --retry <unit>`
[15:48] <marcoceppi> then debug-hooks will catch the hook in a new tab in the tmux session, and you can run the hook again by typing "hooks/<hook-name>" or run juju commands since you're now in a hook environment
[15:48] <marcoceppi> vds: https://juju.ubuntu.com/docs/authors-hooks.html
[15:57] <jcastro> hear ye
[15:58] <jcastro> hear ye, the Juju Charm Weekly status is happening in 2 minutes
[15:58] <jcastro> http://ubuntuonair.com if you want to follow along
[15:58] <jcastro> https://plus.google.com/hangouts/_/b340bb6de5946ca99622521920b9c1454b06c6e4?authuser=0&hl=en
[15:58] <jcastro> ^^ if you want to participate in the hangout
[16:00] <jcastro> marcoceppi: evilnickveitch m_3 ^^^
[16:04] <vds> marcoceppi, thanks, it worked
[16:04] <vds> it looks like in the docs (https://juju.ubuntu.com/docs/authors-hooks.html ) juju resolved is not mentioned at all
[16:04] <marcoceppi> jcastro: pad http://pad.ubuntu.com/7mf2jvKXNa
[18:35] <lazyPwork> crap i missed the charm status. Was it on air so i can play catchup?
[18:35] <marcoceppi> lazyPwork: yup, there's a video on the juju lists
[18:36] <marcoceppi> https://lists.ubuntu.com/archives/juju/2013-October/003008.html
[18:37] <lazyPwork> Already on it ty
[19:57] <arosales> marcoceppi, sinzui I put my findings for installing brew on the thread https://github.com/mxcl/homebrew/pull/22772
[19:57] <arosales> sinzui, is 1.14.1 brew formula still failing though?
[20:52] <sinzui> arosales, I think we need to make a new pull request to get 1.14.1 in. The request you commented on was closed.
[21:02] <arosales> sinzui, do you think the 1.14.1 formula will build?
[21:02] <sinzui> I don't know.
[21:04] <sinzui> I would only know  by installing the entire home brew setup, and make the change to see if it builds.
[21:04] <sinzui> arosales, The details of why it failed are empty http://bot.brew.sh/job/Homebrew%20Pull%20Requests/2772/
[21:06] <arosales> sinzui, hmm that just gives me a Status Code: 404 no exception or stacktrace
[22:02] <zradmin> Im a bit further with my quantum issue. I can ping the router ip i create from the quantum-gateway node, and when i run ip netns on it it shows all the ports etc available. If i run it on the nova-compute node however i get a blank list instead and nothing travels over the br-tun interface... has anyone else had a similar problem?