[00:55] <smoser> hey all
[00:55] <smoser> per http://www.dreamhost.com/newsletter/1113.html#a2
[00:55] <smoser> dreamcompute is now "officially in beta"
[00:55] <smoser> so you could sign up at http://www.dreamhost.com/cloud/dreamcompute/
[00:55] <smoser> and presumably get a account and then point juju at it
[00:56] <smoser> (i say presumably becauase i have not gotten response to my signup from earlier this mornign)
[00:56] <thumper> o/ smoser
[01:01] <lazypower> Thats good news!
[01:50] <marcoceppi> smoser: sweet, I'll give it a go as soon as I get creds
[02:00] <hatch> marcoceppi hey wb :)
[02:01] <marcoceppi> o/
[02:01] <hatch> now go review my Ghost charm :P
[02:02]  * hatch snickers and runs away
[02:03] <marcoceppi> hatch: it's not in the queue, go put it in the queue
[02:03] <Luca__> marcoceppi: I am back now. So in regards of network infrastructure, could you point some document that explains how networks are set up using charms?
[02:05] <marcoceppi> hatch: If it's not listed here: http://manage.jujucharms.com/tools/review-queue then I'm not going to review it
[02:05] <hatch> hmm
[02:06] <hatch> marcoceppi so https://juju.ubuntu.com/docs/authors-charm-store.html#submitting is not a complete list of steps then?
[02:06] <marcoceppi> hatch: it is, link to the bug
[02:07] <marcoceppi> I'll see what's missing
[02:07]  * marcoceppi is working on a juju charm submit command to make this process a little more automated
[02:07] <hatch> https://bugs.launchpad.net/charms/+bug/1229377
[02:07] <_mup_> Bug #1229377: Charm needed: Ghost blogging platform <Juju Charms Collection:New> <https://launchpad.net/bugs/1229377>
[02:07] <hatch> I don't see where it says to add it to the review queue....or how for that matter :)
[02:08] <lazypower> hatch: Did you add the "charmers" group to the bug?
[02:08] <marcoceppi> hatch: huh, this should be in the queue. Charmers is watching the bug and it's in a new or fix commited status.
[02:08] <hatch> lazypower yup
[02:08] <marcoceppi> hatch: I'll open a bug and make a note to review this first tomorrow
[02:08] <hatch> marcoceppi cool, so I did everything correctly then?
[02:09] <hatch> I'm not really too concerned I just want to make sure that the documentation is up to date so outsiders don't run into issue
[02:09] <marcoceppi> hatch: looks like it, I'll ping charmworld and figure out why this didn't get scraped
[02:09] <hatch> great thanks
[02:10] <hatch> marcoceppi in that list it says "store error on ghost"
[02:11] <marcoceppi> hatch: that's something different, but equally interesting
[02:11] <hatch> oh :) cool well I was just joking about rushing to get it approved, take your time
[02:12] <marcoceppi> hatch: too late, rush order already processed and your card has been debited $125
[02:12] <hatch> awwww man
[02:13] <marcoceppi> there's a cancellation fee of $200
[02:13] <hatch> man you must be taking business lessons from the government
[02:13] <hatch> haha
[02:17] <hatch> well thanks in advance, lemme know if you need any help with anything
[02:19] <marcoceppi> hatch: hey, I got it to show in the queue
[02:21] <marcoceppi> hatch: you need to assign the bug to someone, preferably yourself. I'll update the docs. Hopefully `juju charm submit` will be easy enough that the majority of instructions can be condensed to that one command
[02:21]  * marcoceppi searches for other charms that may have fallen through the cracks
[03:02] <hatch> marcoceppi ahh interesting thanks
[05:05] <Luca__> Hi there
[05:05] <Luca__> does anybody know how to reuse a machine once issued a juju destroy-machine command?
[05:06] <Luca__> it seems even this machine goes back to ready status in MAAS, once picked up again for a different service provision it just stuck and does not do anything
[05:07] <davecheney> Luca__: that is werid
[05:07] <davecheney> it shoulnd't do that
[05:07] <davecheney> after destroy-machine
[05:07] <davecheney> can you confirm that the machine is no longer listed in juju status ?
[05:09] <Luca__> davecheney: I confirm it is no longer in the juju status
[05:09] <Luca__> my understanding is that it should be reused and go through the OS reinstallation
[05:09] <davecheney> yup
[05:09] <davecheney> the only thing which could prevent that is constraints
[05:10] <davecheney> /var/log/juju/machine-0.log on machine-0 will have the details of whta is going on
[05:10] <Luca__> I am checking now the machine, status is pending, however not doing any installation
[05:10] <Luca__> let me check
[05:12] <Luca__> what should I exactly look for? there is a bunch of logs
[05:13] <Luca__> 2013-12-02 05:02:03 DEBUG juju firewaller.go:346 worker/firewaller: started watching unit rabbitmq-server/0
[05:14] <davecheney> it will say worker/provisioner
[05:17] <Luca__> nothing about worker/provisioner
[05:17] <Luca__> the messages I can find are worker/firewaller
[05:17] <Luca__> 2013-12-02 04:46:28 DEBUG juju firewaller.go:346 worker/firewaller: started watching unit mysql-hacluster/0 2013-12-02 04:46:28 DEBUG juju firewaller.go:346 worker/firewaller: started watching unit mysql-hacluster/1 2013-12-02 05:02:03 DEBUG juju firewaller.go:346 worker/firewaller: started watching unit rabbitmq-server/0 2013-12-02 05:02:03 DEBUG juju firewaller.go:495 worker/firewaller: started watching machine 13 2013-12-02 05:02:03 D
[05:20] <davecheney> Luca__: nothing from the provisioner ?
[05:20] <davecheney> in the whole file ?
[05:21] <Luca__> yes, nothing in the whole file
[05:21] <Luca__> but I could deploy mysql and ceph
[05:22] <Luca__> hold on
[05:22] <Luca__> filtering for provisioner return some
[05:22] <Luca__> 2013-12-02 05:02:03 INFO juju.provisioner provisioner_task.go:239 found machine "13" pending provisioning
[05:23] <Luca__> 2013-12-02 05:02:03 INFO juju.provisioner provisioner_task.go:239 found machine "14" pending provisioning
[05:23] <Luca__> 2013-12-02 05:02:03 DEBUG juju.provisioner provisioner_task.go:303 Stopping instances: [nfqwq.master:/MAAS/api/1.0/nodes/node-de6db216-59d4-11e3-8f27-525400378304/ ac69q.master:/MAAS/api/1.0/nodes/node-ee42a908-59d4-11e3-9087-525400378304/ pwq4h.master:/MAAS/api/1.0/nodes/node-472158c0-59d6-11e3-9087-525400378304/]
[05:23] <Luca__> 2013-12-02 05:02:05 INFO juju.provisioner provisioner_task.go:367 started machine 13 as instance /MAAS/api/1.0/nodes/node-de6db216-59d4-11e3-8f27-525400378304/ with hardware <nil>
[05:24] <Luca__> but does not seem doing anything
[05:27] <davecheney> Luca__: ok, time to check the maas log
[05:27] <davecheney> if maas says that machine is in ready state
[05:28] <davecheney> then it will have a record of why it refused to start a machine
[05:28] <Luca__> from MAAS Web machine was ready
[05:28] <davecheney> i think what has happened is thing happende too fasr
[05:28] <davecheney> 3 seconds between maas stopping the machien then reusing it
[05:29] <davecheney> not a lot of time for all the bookkeeping maas has to do
[05:29] <davecheney> you could try deleting the service, unit and machine
[05:29] <davecheney> or actualy just delet ehte unie
[05:29] <lifeless> davecheney: maas shouldn't have 3s of work to do
[05:29] <davecheney> juju remove-unit $UNIT
[05:29] <Luca__> ok
[05:29] <davecheney> juju destroy-machine 14
[05:29] <davecheney> then watching the log
[05:29] <Luca__> ok
[05:29] <davecheney> waiting til you see maas remove the unit
[05:30] <davecheney> then juju add-unit $SERVICE
[05:30] <Luca__> MAAS logs or JUJU logs?
[05:32] <Luca__> 2013-12-02 05:31:43 DEBUG juju firewaller.go:338 worker/firewaller: stopped watching unit rabbitmq-server/0
[05:32] <Luca__> 2013-12-02 05:31:43 DEBUG juju firewaller.go:338 worker/firewaller: stopped watching unit rabbitmq-server/1
[05:33] <davecheney> the juju machine-0 log for a start
[05:33] <davecheney> but watching the maas logs is recommended as it can be very terse otherwise
[05:36] <Luca__> it returns some errors but not sure whether is related
[05:37] <Luca__> 2013-12-02 05:32:53 DEBUG juju.rpc.jsoncodec codec.go:107 <- {"RequestId":3057,"Type":"Provisioner","Request":"Life","Params":{"Entities":[{"Tag":"machine-14"}]}}
[05:37] <Luca__> 2013-12-02 05:32:53 DEBUG juju.rpc.jsoncodec codec.go:172 -> {"RequestId":3057,"Response":{"Results":[{"Life":"dying","Error":null}]}}
[05:37] <Luca__> 2013-12-02 05:32:53 DEBUG juju.rpc.jsoncodec codec.go:107 <- {"RequestId":3058,"Type":"Provisioner","Request":"InstanceId","Params":{"Entities":[{"Tag":"machine-14"}]}}
[05:38] <Luca__> machine 13 and 14's life is dying
[05:38] <davecheney> ok, those arent' errors
[05:39] <davecheney> have machine 13 and 14 returned to maas ready state ?
[05:39] <Luca__> 2013-12-02 05:32:53 DEBUG juju.rpc.jsoncodec codec.go:172 -> {"RequestId":3064,"Response":{"Results":[{"Error":null,"Result":"/MAAS/api/1.0/nodes/node-de6db216-59d4-11e3-8f27-525400378304/"}]}}
[05:39] <Luca__> nope
[05:39] <davecheney> it's possible they haven't been properly relased back to maaas so now nobody owns them
[05:39] <davecheney> ok, this isn't good
[05:39] <davecheney> maas has lost track of the machine
[05:39] <Luca__> state is still dying
[05:39] <davecheney> which version of juju are you using ?
[05:39] <davecheney> which version of maas are you using ?
[05:40] <Luca__> maas: 1.4+bzr1701+dfsg-0+1718+212~ppa0~ubuntu12.04.1
[05:41] <Luca__> juju-core 1.16.3-0ubuntu1~ubuntu12.04.1~juju1
[05:41] <Luca__> on 12.04.03
[05:41] <davecheney> hmm, i think thta maas install is quite old
[05:43] <davecheney> hmm, actually, error: null is fine
[05:44] <davecheney> i just checked the code
[05:44] <Luca__> ok
[05:44] <davecheney> so, maas gave you back the same machine you deleted
[05:44] <davecheney> what state is maas saying /MAAS/api/1.0/nodes/node-de6db216-59d4-11e3-8f27-525400378304 is in ?
[05:45] <Luca__> life: dying
[05:47] <davecheney> in maas or in juju ?
[05:47] <Luca__> juju status
[05:49] <davecheney> what about maas ?
[05:49] <Luca__> I am looking under /var/log/maas/
[05:52] <Luca__> is there any specific file  I should check? maas.log is empty and others do not seems to have any information related to this node
[05:58] <davecheney> so there is no record that maas assigned that node to match
[05:58] <davecheney> 16:39 < Luca__> 2013-12-02 05:32:53 DEBUG juju.rpc.jsoncodec codec.go:172 -> {"RequestId":3064,"Response":{"Results":[{"Error":null,"Result":"/MAAS/api/1.0/nodes/node-de6db216-59d4-11e3-8f27-525400378304/"}]}}
[05:58] <Luca__> this is what I get from bootstrap node
[05:59] <Luca__> but if I check maas node I cant find any reference to those under /var/log/maas/
[06:02] <davecheney> what is the request that matches that response ?
[06:07] <Luca__> 2013-12-02 05:32:53 DEBUG juju.rpc.jsoncodec codec.go:107 <- {"RequestId":3064,"Type":"Provisioner","Request":"InstanceId","Params":{"Entities":[{"Tag":"machine-13"}]}}
[06:09] <davecheney> ok, that is getting wiereder and weiredr
[06:09] <davecheney> how many machines did you deleted
[06:09] <davecheney> it looks like /MAAS/api/1.0/nodes/node-de6db216-59d4-11e3-8f27-525400378304/ has been assigned as machine/14
[06:12] <Luca__> Yes I agree. I have never been able to reclaim a destroyed machine and what I was doing was replacing the virtual HDD with a new one, going through the discover process in maas and make the machine available once again
[06:21] <Luca__> davecheney: any thought about that?
[06:29] <Luca__> I did delete 2 machines
[06:30] <Luca__> #13 and #14
[06:30] <Luca__> which were the machines were I tried to install rabbitmq
[07:49] <Luca__> getting this 2013-12-02 05:02:15 ERROR juju runner.go:200 worker: fatal "api": agent should be terminated
[07:50] <Luca__> From here machine is in dying staus
[07:50] <makara> hi. Juju is stuck on "juju status". I've destroyed environments, setup .bashrc to handle keys, and then "juju init; juju bootstrap; juju status" without any other customizations
[07:50] <makara> what's going on?
[07:53] <makara> what's the difference between 'juju init' and 'juju generate-config' ?
[08:04] <lazypower> makara: nothing. generate-config      alias for init
[08:05] <lazypower> makara: What do you mean juju is stuck on juju status? Can you provide logs?
[08:12] <makara> i type in the command and it hangs
[08:12] <makara> sorry, I'm too tired to send logs
[08:13] <makara> i was hoping somebody else had the same problem :)
[08:37] <Luca__> Is there anybody that could deploy openstack with charms here?
[09:46] <marcoceppi> makara: are you sourcing your .bashrc?
[09:48] <marcoceppi> makara: run juju bootstrap with --debug and --show-log the paste the output http://paste.ubuntu.com
[09:49] <makara> marcoceppi, by sourcing...i've added environment variables to .bashrc
[09:50] <marcoceppi> makara: cool, when you have a moment run bootstrap with the extra flags and let us know the output
[09:58] <makara> marcoceppi, http://pastebin.com/x1QCGGnD
[10:01] <marcoceppi> makara: give it a few mins and run juju status --debug --show-log
[10:05] <makara> marcoceppi, do I need any special ports open? I'm behind a fairly restrictive firewall
[10:05] <makara> 2013-12-02 10:04:21 DEBUG juju.state open.go:88 connection failed, will retry: dial tcp 54.226.118.252:37017: connection timed out
[10:06] <davecheney> makara: yes, you need 37017 and 27017
[10:06] <marcoceppi> you need access to 37017 or a proxy server outside of your firewall
[10:07] <davecheney> marcoceppi: um, i don't think juju cli can use a proxy
[10:07] <davecheney> although you could deploy the gui
[10:07] <marcoceppi> davecheney: :(
[10:07] <marcoceppi> davecheney: no you cant, he can't reach the bootatrap node
[10:09] <marcoceppi> davecheney: i thought you guys respected proxies. maybe not. could use sshuttle as a proxy i guess. either way we should find and document a work around
[10:09] <davecheney> marcoceppi: it's complicated
[10:09] <davecheney> we respect the apt proxy
[10:09] <davecheney> but the mongo traffic isn't proxyable
[10:09] <davecheney> and that sort of carried through into the api
[10:09]  * davecheney raises a bug
[10:11] <marcoceppi> i wonder if quickstart would fix this. you could get a gui on the bootstrap node directly without using the deploy command
[10:11] <teknico> jamespage, hi, got an MP for the glance charm, any chance having a look? thanks https://code.launchpad.net/~teknico/charms/precise/glance/preload-ubuntu-images/+merge/196341
[10:11] <davecheney> https://bugs.launchpad.net/juju-core/+bug/1256849
[10:11] <_mup_> Bug #1256849: cmd/juju: no support for proxies <juju-core:New> <https://launchpad.net/bugs/1256849>
[10:12] <davecheney> marcoceppi: maybe we should just always deploy the gui on the bootstrap node
[10:12] <marcoceppi> +10000000000000000000000
[10:13] <marcoceppi> or have a flag --without-gui
[10:13] <jamespage> teknico, ah
[10:13] <marcoceppi> for those that dont want it
[10:13]  * davecheney foes to raise another bug
[10:16] <davecheney> https://bugs.launchpad.net/juju-core/+bug/1256852
[10:16] <_mup_> Bug #1256852: cmd/juju: juju should always deploy the gui when bootstrapping <juju-core:New> <https://launchpad.net/bugs/1256852>
[10:16] <jamespage> teknico, ok - we need to discuss how we integrate simplestreams sync with glance; I'm not convinced this is the right way todo it
[10:16] <teknico> jamespage, we do. I'm not convinced either, it couples the two of them too much
[10:17] <jamespage> teknico, ageed
[10:17] <jamespage> I'd envisaged a simplestreams-mirror charm
[10:17] <jamespage> that you deploy (probably in a container)
[10:17] <jamespage> and then relate to keystone
[10:17] <jamespage> so it can register a product-streams endpoint and find out where things like glance are
[10:19] <teknico> a separate charm altogether?
[10:19] <teknico> I was considering just calling the tools/sstream-mirror-glance script, but it's not packaged (yet)
[10:20] <teknico> (that script is in simplestreams)
[10:48] <makara> hi. I've setup a juju control instance so I shouldn't have problems with ports. How can I use a specific SSH key when I deploy with juju?
[10:50] <makara> i also want to install instances into a VPC
[11:01] <marcoceppi> makara: we don't support VPC at this time, and you can specify ssh keys in your environments.yaml file (~/.juju/) using the authorized-keys option
[11:26] <makara> marcoceppi, where can I find a complete list of options for the .yaml?
[11:28] <davecheney> makara: if you do juju init --show
[11:28] <davecheney> then all the option are presented
[11:29] <davecheney> commented out (ie, set to default) in the sample config
[11:30] <marcoceppi> davecheney: yeah, but we don't have /all/ of the options in there
[11:30] <marcoceppi> like authorized-keys
[11:32] <makara> and its not in https://juju.ubuntu.com/docs/config-aws.html
[11:34] <makara> what exactly does bootstrap do?
[11:35] <makara> i've had to create an instance with juju on EC2 to get past company firewall issues. Now I just want to start deploying
[11:38] <makara> can I just bootstrap to the instance with juju installed on it?
[11:39] <makara> my pc > dmz instance with juju > bootstrap > wordpress
[11:40] <makara> is a little verbose
[11:42] <jamespage> teknico, yeah - a separate charm
[11:43] <teknico> jamespage, invoking a simplestreams script would decouple the glance charm from simplestreams internals, would that be enough?
[11:44] <jamespage> teknico, I'm not worried about that so much; but the sync process needs to be able to deal with
[11:44] <jamespage> a) multiple instances of the glance charm in different service units for scale-out/ha
[11:44] <jamespage> b) periodic re-syncing of images
[11:45] <jamespage> c) endpoint registration for things like juju to use
[11:45] <jamespage> if we do it right, we should also be able to sync juju tools into a openstack cloud in the same way
[11:46] <teknico> wow, this expands the scope of the thing quite a bit :-)
[11:46] <jamespage> by splitting it into its own charm, we don't have to worry about scale out (its a async, period process)
[11:46] <jamespage> teknico, I don't really want to have a feature which conflicts/confuses those objectives in the glance charm itself
[11:47] <teknico> yeah, I understand it being a separate and extrinsic concern for the charm
[11:49] <teknico> jamespage, can you expand a little on "register a product-streams endpoint" please?
[11:49] <teknico> (which I believe is what your c) point refers to)
[11:49] <jamespage> teknico, juju uses the keystone service catalog to lookup both image information and juju tool information
[11:50] <teknico> oh right, the service catalog in keystone
[11:50] <jamespage> the metadata for these are normally stored in swift, with the images in glance for ubuntu images, and in swift for juju tools
[11:50] <jamespage> the sync charm is something that communicates with a number of services in openstack, not just glance :-)
[11:51] <teknico> when you say "sync charm is" you mean "will be", right? you're not referring to something that already exists
[11:54] <teknico> jamespage, and yeah, the current MP talks to both keystone and swift via simplestreams
[13:44] <jcastro> hey evilnickveitch
[13:44] <jcastro> http://discourse.ubuntu.com/t/how-to-translate-juju-docs/1284
[13:45] <evilnickveitch> jcastro, ok, cool
[13:53] <jcastro> jamespage, how familiar are you with eclipse virgo?
[13:54] <jamespage> jcastro, still using juno right now
[13:55] <jcastro> hey so they're writing a charm and have it all mostly working, but are having some upstart problems, I was thinking of pinging you when they hit the review queue to have you take a look?
[13:56] <jcastro> It won't be soon
[13:57] <jcastro> https://github.com/glyn/virgo-charm/ if you are bored though. :)
[15:00] <X-warrior> marcoceppi: are u there?
[15:00] <marcoceppi> X-warrior: yup
[15:01] <X-warrior> oh I thought you were one of the creators of elasticsearch charm, but looking on git commit tree you just did the latest commit
[15:30] <yolanda>  hi jamespage, pushed some updates to heat charm
[15:36] <jamespage> yolanda, cool - I'll take a peek later!
[15:36] <yolanda> thx
[16:34] <X-warrior> Is instance-type not working on 1.16.3?
[16:34] <marcoceppi> X-warrior: correct, it hasn't been fully ported from juju 0.7
[16:35] <marcoceppi> X-warrior: https://juju.ubuntu.com/docs/reference-constraints.html
[16:41] <benji> marcoceppi: I'm ready when you are: https://plus.google.com/hangouts/_/76cpit9hbj5c3kr1burqftqmk0?hl=en
[17:31] <lazypower> Buenos Dias everyone
[18:26] <X-warrior> If I deploy something with juju, it will save it on charmcache and in future if I deploy again, it will deploy the same version. Right?
[18:29] <marcoceppi> X-warrior: so long as the environment is alive. If you destroy-environment the cache goes away
[18:29] <X-warrior> marcoceppi: so if I destroy the machine where that one service is deployed
[18:30] <X-warrior> it will still be in cache...
[18:31] <X-warrior> Based on that, I think that upgrade-charm should work even if we don't have any instance with that service running...
[18:31] <X-warrior> if that service exists in cache, it should be updated imo
[18:44] <lazypower> window 3
[20:44] <zzecool> Hello fellas i need your help , i just install juju locally on my pc and when i try to deploy the juju-gui  on the "watch juju status" i see that it creates a second machine that is pending . Is this normal ?
[20:48] <zzecool> anyone ?
[20:52] <marcoceppi> zzecool: yes, that's expected
[20:52] <marcoceppi> zzecool: if you wanted to not have it create another machine, you can tell juju-gui to deploy to the bootstrap node using this line instead:
[20:53] <marcoceppi> juju deploy --to 0 juju-gui
[20:53] <zzecool> ohh thank you :)
[20:53] <marcoceppi> zzecool: since you've "deployed" the juju-gui already, you'll need to first destroy it, before deploying again.
[20:53] <marcoceppi> juju destroy-service juju-gui
[20:53] <marcoceppi> juju deploy --to 0 juju-gui
[20:53] <zzecool> how long does it take to deploy for the first times?
[20:53] <marcoceppi> juju terminate-machine 1
[20:54] <marcoceppi> zzecool: it depends on the cloud provider
[20:54] <zzecool> it locally
[20:54] <marcoceppi> and the charm, some charms can take a long time, others are pretty quick
[20:54] <marcoceppi> zzecool: ah, for local deployments it needs to download the ubuntu cloud image, it's about 250MB so it can take some time
[20:54] <zzecool> i see
[20:54] <zzecool> really thank you
[20:54] <zzecool> :)
[20:54] <marcoceppi> zzecool: also, you can't use deploy --to 0 on local provider, so don't worry about it creating a new machine
[20:55] <zzecool> ohh
[20:55] <marcoceppi> zzecool: the 0 machine (bootstrap node) is technically your computer for the local provider
[20:55] <marcoceppi> you wouldn't want to put the gui directly on you machine!
[20:55] <marcoceppi> since you're not "paying" for the machines in the cloud provider, having an extra machine for the gui won't hurt
[20:56] <marcoceppi> zzecool: what version of juju are you using? (juju version)
[20:56] <zzecool> the latest
[20:56] <zzecool> i used the ubuntu ppa
[20:56] <marcoceppi> zzecool: cool, perfect
[20:57] <zzecool> i think i misunderstood the juju thing
[20:57] <zzecool> so all tha deployment thing doesnt install the actual applications on my pc
[20:57] <zzecool> is it something like a sandbox?
[20:58] <marcoceppi> zzecool: no no, it's using LXC to create contianers that simulate a cloud on your machine
[20:58] <marcoceppi> except, with the local provider, machine 0 is technically your laptop/desktop. The rest of the machines juju deploys are LXC machines running on your laptop/desktop
[20:59] <zzecool> LXC is like virtual machines right?
[20:59] <zzecool> so i guess i cant use machine 0 to deploy things for security reasons ?
[21:00] <zzecool> marcoceppi: Thank you for you help :)
[21:01] <marcoceppi> zzecool: right, only with local provider, machine 0 is off limits
[21:02] <zzecool> i see
[21:02] <marcoceppi> LXC is conceptually the same thing as virtual machines, just faster and lighter weight
[21:03] <zzecool> then my whole concept is a fail . The reason i installed juju was to easily install observium locally to monitor my pc
[21:04] <zzecool> so i guess if i install observium i will monitor the LXC machine (1) and not my real pc...
[21:09] <marcoceppi> zzecool: correct
[21:09] <marcoceppi> zzecool: juju is a service orchestration tool, meant to drive bare metal and cloud machines
[21:09] <marcoceppi> the local provider is a way to develop and hack on charms, what juju deploys, without needing lots of servers or spending money on a cloud
[21:10] <zzecool> It is so nice though , it should be used for local install as well instead of synaptic or apt-get
[21:34] <thumper> marcoceppi: you there?
[21:34] <thumper> marcoceppi: FWIW, you can't deploy units to machine 0 with the local provider
[21:34] <thumper> as the host machines IS machine-0
[21:35] <thumper> zzecool: if this is the first time with the local provider, it is probably syncing the underlying ubuntu cloud images
[21:36] <thumper> marcoceppi: with the local provider, machine-0 isn't able to host units (and should be constrained in the data model)
[21:36] <thumper> if it thinks it can, it is a bug
[21:36] <zzecool> thumper: yeap he allready told this
[21:37] <thumper> zzecool: cool, just flicked through the chatter without reading in depth :)
 zzecool: also, you can't use deploy --to 0 on local provider, so don't worry about it creating a new machine
[21:37] <zzecool> :)
[21:37] <zzecool> thumper: np thanks though
[21:39] <dpb1> If I'm on a juju deployed unit, is there anyway for me to signal that I'm leaving the service and shutting down?
[21:47] <thumper> dpb1: not yet
[21:48] <thumper> dpb1: I'm assuming you mean the unit itself wants to leave?
[21:48] <dpb1> Hey thumper, I just wrote to juju@ as well.  Yes, right.
[21:48] <thumper> what is the use case for this?
[21:49] <thumper> ah...
[21:49] <dpb1> I guess the same use case that shutdown on the command line of an instance presents
[21:49] <thumper> just read the email
[21:49] <dpb1> ok, I think that makes it more clear
[21:49] <thumper> I think we added 'destroy-machine --force'
[21:49] <thumper> which works in this case (I think)
[21:50] <thumper> not sure which release that is coming in though
[21:50] <dpb1> ok... I can check that
[21:50] <dpb1> thx for the pointer.
[21:57] <davecheney> marcoceppi: really ? that is le suck
[22:03] <marcoceppi> davecheney: which is?
[22:07] <davecheney> marcoceppi: the fact we don't document all the options in juju init > sample.yaml
[22:07] <marcoceppi> davecheney: oh, yeah. It is
[22:07] <marcoceppi> we don't document it anywhere
[22:07] <marcoceppi> you have to crawl through the code
[22:08] <davecheney> i'm sure evilnickveitch won't appreciate us dropping that one in his lap
[22:08] <marcoceppi> at least, we don't document all the fringe options
[22:08]  * marcoceppi makes a bug
[22:08] <marcoceppi> I might do it, well I might gather all the config options, I'll let someone smarter figure out what they mean
[22:11] <davecheney> `juju is`
[22:14] <marcoceppi> <3 easter eggs