[00:47] are there many charms for trusty yet? [00:48] or does that sort of thing wait until after trusty is out? [00:56] Hey all, is it possible to select a MAAS node when running the first 'juju bootstrap' to setup the juju environment? I have pre-defined nodes and I'd like to make sure it provisions on the right one [01:08] mwhudson: the latter [01:10] davecheney: is the cheat in the mean time branching all the precise repos to $localcharmrepo/trusty or however you do that? [01:11] mwhudson: do you need them to run on trusty? You can just juju deploy precise/mysql for instance? [01:11] mwhudson: correct [01:11] rick_h_: i am mostly interested in architectures that are only supported in trusty :-) [01:12] rick_h_: that is the main rason [01:12] mwhudson: if you download zip of a charm you can drop it on the gui and pick the series to use [01:12] mwhudson: ah gotcha [01:12] well i guess armhf was in precise but i still want to be closer to the bleeding edge [01:12] rick_h_: ok, bit tedious when dealing with a gigantic bundle [01:12] (unless there is a trick for that too) [01:13] mwhudson: definitely, no trick until we get fat bundles and the like [01:13] ah ok === arosales_ is now known as arosales === vladk|offline is now known as vladk === CyberJacob|Away is now known as CyberJacob === vladk is now known as vladk|offline === CyberJacob is now known as CyberJacob|Away === vladk|offline is now known as vladk [08:12] hola juju people ! With juju 1.18.0 i have been told one can tell juju, when using openstack provider, to use a given neutron network for vms (so it does not die when the current tenant as several available networks). [08:12] any idea actually how to tell juju wich net to use ? [08:24] ahhh, most probably the new network setting that is in environment.yaml [08:31] rbasak: morning. the new quickstart package(in your ppa) seems to work very well [09:02] frankban: great, thanks! I'll upload. [09:03] rbasak: cool thank you [09:08] (done) === Kyle____ is now known as Kyle === Kyle is now known as Guest44217 === Guest44217 is now known as Eliz [11:35] I'm having to `sudo pkill mongod` each time I re-bootstrap a local environment. Is this known, or the same issue as bug 1208430 ? [11:35] <_mup_> Bug #1208430: mongodb runs as root user [13:58] i m experiencing problem with juju 1.18.0 tyring to use it behind a proxy. [13:58] https://pastebin.canonical.com/108082/ [13:58] if anyone has an idea, you are welcome === hatch__ is now known as hatch [14:16] melmoth: can you access the maas API directly using HTTP_PROXY envvar and wget say? doesn't seem like a juju related error [14:17] i dont need a proxy to access MAAS, and the proxy had no access to MAAS directly [14:18] (all my boxes, MAAS as well are vm in a hypervisor;hosted on some libvirt network). The proxy is a real proxy that my main hypervisor must use to access external boxes [14:18] so if i try to access the maas box via the proxy, it will not work.. is it what is going on here ? [14:20] melmoth: printenv? [14:21] mgz http://pastebin.com/UsBbLR5u [14:21] (i m using juju directly from the box where maas is installed) [14:26] melmoth: that's odd. what about /home/ubuntu/.profile and /home/ubuntu/.juju-proxy? running juju from inside of the maas... that it's using itself? seems shakey. [14:26] i always did it that way. [14:26] i mean, juju is just a client, it should be ran from...anywhere [14:27] i could fire up yet a new vm just for it, but, i never had to, nor do i see the point [14:27] lets have a look at the profile [14:28] http://pastebin.ubuntu.com/7226462/ [14:29] well, I have no idea where your juju client is getting the proxy setting from then. [14:29] well, from the nevironement most probably, i did set juju set http-proxy [14:29] because i need the new machine to use the proxy [14:29] that was the whole point of the exercice [14:30] but i dont need communication from the juju box or the bootstrap node to MAAS to be proxyed [14:30] hmmm [14:30] may be i should set no-proxy for the maas box ? [14:30] well, you are telling the bootstrap node to use a proxy with that [14:30] affecting the local client when uploading a charm is a little odd [14:31] melmoth: you probably should [14:31] i m just not sure what name to use or wich actual machine it s trying to cennect to [14:32] let s try with 192.168.101.2 [14:34] \o/ [14:34] thanks mgz :-) [14:34] workies? [14:35] looks like it did. At least i have a prompt. it s a subordinate so nothing happened, that s normal. let s continue deployment [14:37] yeah, mysql seems to be under deployment [14:40] ace. === alexlist is now known as alexlist` [16:06] So, I'm new to bzr. If I want to submit a (set of) change(s) for review on charm-helpers, do I just do bzr push lp:~johnsca/charm-helpers/name-for-branch in my local copy? [16:08] arosales, Hey I pushed a fix for: https://bugs.launchpad.net/charms/+source/jenkins/+bug/1272520 -- but it's pushed to a +junk branch so I can submit an MP [16:08] <_mup_> Bug #1272520: Unable to relate to jenkins-slave: hook failed: "master-relation-changed" [16:09] * timrc really doesn't want any responsibilities associated with being a member of ~charmers, but this particular bug has bit us, and thought I would kill it with fire [16:09] cory_fu: yes, then do a bzr lp-propose lp:charm-helpers [16:09] timrc: right on. I'll take a look closer to EOB unless you need it landed immediately. [16:09] and follow marco's directive ^ [16:09] lazyPower, Nah, take your time. Not hugely critical. We are deploying from a branch atm... there is some weirdness around authentication too but I think that's separate [16:10] lazyPower, Right-o. Thanks [16:10] timrc: thanks for the mp. I think I had filled a abug against jenkins . . . [16:11] timrc: https://bugs.launchpad.net/charms/+source/jenkins/+bug/1272520 was the bug I was hitting did you see something similiar? [16:11] <_mup_> Bug #1272520: Unable to relate to jenkins-slave: hook failed: "master-relation-changed" [16:12] arosales, Well I didn't see any logs, so I can't be 100% sure but my master-relation-changed problem was due to the fact that the master-relation-changed hook was expecting the slave-relation-joined hook (run by jenkins-slave) to set 'slaveaddress' which it wasn't [16:12] arosales, so master-relation-changed was assigning an empty string and passing that to hooks/addnode and causing breakage [16:13] timrc: sounds promissing that your branch may resolve this issue [16:14] timrc: thanks for taking the time to contribute the fix back into the juju community :-) === JoshStrobl__ is now known as JoshStrobl [16:43] marcoceppi, Just to confirm my lp-propose should be: [16:43] Source: lp:~timrchavez/charm-helpers/jenkins-slave-fix-slave-relation-joined [16:43] Target: lp:charm-helpers [16:44] should result in* rather [16:44] timrc: is this a fix in charm-helpers or the jenkins charm? [16:45] marcoceppi, it's a fix to a jenkins charm... when I attempted to propose my branch for merge against the actual charm I got a little hate :( [16:45] bzr: ERROR: exceptions.Exception: lp:~timrchavez/charm-helpers/jenkins-slave-fix-slave-relation-joined is not mergeable into lp:charms/jenkins-slave [16:46] timrc: you should push your branch to lp:~timrchavez/charms/precise/jenkins/jenkins-slace-fix-slave-relation-joined [16:46] timrc: you're trying to merge not only across projects, but across distros [16:46] marcoceppi, muhaha [16:47] if I could push to charms/ I would have.. [16:47] timrc: you can [16:47] timrc: it's charms in your namespace [16:47] anyone can push to lp:~user/charms/... [16:47] marcoceppi, Ah, I had some problem... maybe my path was bad [16:48] marcoceppi, Ah, I attempted bzr push lp:~timrchavez/charms/precise/jenkins-slave-fix-slave-relation [16:48] timrc: yeah, since charms is a distro, you have to put a branch name in there and jenkins becomes the project name [16:50] Makes sense. Thanks for working around my obtuseness and helping me :) [16:50] timrc: np! once it's all pushed, you can just do `bzr lp-propose lp:charms/jenkins-slave` [17:03] lazyPower, Okay stumbled my way through :) https://code.launchpad.net/~timrchavez/charms/precise/jenkins-slave/jenkins-slave-fix-slave-relation-joined/+merge/214994 === natefinch is now known as natefinch-afk === roadmr is now known as roadmr_afk === roadmr_afk is now known as roadmr === vladk is now known as vladk|offline [18:24] anyone ever successfully configured sticky sessions on the haproxy charm? [18:32] tvansteenburgh: people in IS probably have [18:34] marcoceppi: thanks, will check there === CyberJacob|Away is now known as CyberJacob [19:03] Is there anyway to start a service on the juju agent machine? Or is there a way to deploy to a specific machine number in juju status? [19:06] Jeffrey_: yes, you can use deploy --to [19:07] marcocappi: Then I get "Added charm "local:percise/rethinkdb-2" to the environment. ERROR cannot assign unit "rethinkdb/0" to machine 0: series does not match" and the machine type of number 0 is "series: percise" [19:08] Jeffrey_: is this a local provoder? [19:08] Yes from a directory on the juju client computer. [19:11] Also if I add a charm it creates a new machine that stays in the pending state. How can I start it? [19:11] Jeffrey_: the bootstrap node isn't a real cloud VM, it's your actual machine [19:11] you'll need to add another machine [19:11] marcoceppi: Ok so the bootstrap node can only be for the agent? [19:11] Jeffrey_: only on local provider [19:12] Jeffrey_: all other real clouds you can deploy --to 0 [19:12] marcoceppi: I'm sorry I meant I am running a MaaS provider. [19:12] I am running from a local repository [19:13] Charm is from a local repository [19:13] Jeffrey_: show me your juju status [19:14] marcoceppi: [19:14] http://pastebin.ubuntu.com/7227684/ [19:17] marcoceppi: Even when I just do a normal deploy the new machine that is allocated for the charm doesn't start. Is there any way for force it? [19:18] Jeffrey_: something else is wrong, it should start up eventually [19:19] marcoceppi: This is the status: http://pastebin.ubuntu.com/7227701/ [19:22] macroceppi: I think I figured it out. === mwhudson- is now known as mwhudson [21:06] I'm thinking about changing some configuration variables for the django charm like: requirements_pip_files [21:06] in your opinion what would be better:? [21:06] 1) having a configuration option with a default but ignoring if the results fail [21:06] 2) a configuration variable with empty default and triggering errors if the file is not there [21:12] avoine: the latter, in my opinion [21:23] `juju deploy local:trusty/haproxy --repository=.` [21:23] is the --repository=. needed there? [21:24] i'm hacking on haproxy and my code changes aren't getting deployed to new units [21:24] not using --repository and wondering if i need to [21:25] marcoceppi: ^ [21:25] tvansteenburgh: pretty sure you always need to specify a repository [21:25] or else? [21:26] maybe it's pulling from lp everytime and i haven't noticed [21:26] no, it wont do that [21:27] yeah i dunno what it's doing then [21:36] tvansteenburgh: you can try with the --debug flag [21:36] see what's happening [21:36] well --repository=. fixed my immediate prob, so i'm moving on [22:25] mbruzek: marcoceppi does charm-help-test work for you? [22:25] on charm-tools 1.2.10 its not giving me the help [22:26] sorry I meant to state, charm-help test [22:26] charm help test, also isn't working for me. [22:27] no I get a usage when I try those commands [22:27] http://paste.ubuntu.com/7228396/ [22:27] I'll file a bug [22:28] mbruzek: thanks for confirming I was hoping to look at what the valid outputs are for charm tests [22:28] arosales, does "charm test" not work for you? It appears that the help is broken for me [22:29] mbruzek: 'charm test' does work for me [22:29] mbruzek: just that 'charm help test' does not or as the usage suggests 'charm-help test' also doesn't work [22:29] sure [22:29] That is a bug [22:32] avoine: charm test -h [22:32] should give you output [22:32] https://bugs.launchpad.net/charm-tools/+bug/1305337 [22:32] <_mup_> Bug #1305337: charm help test does not work on 1.2.10 [22:32] marcoceppi that does work [22:33] perhaps it just the usage that needs updating then [22:33] arosales: charm help doesn't ever work? [22:33] i guess it could [22:33] it didn't work for me in the commands I tried it with, list, test, get, etc. [22:35] cory_fu: fyi http://paste.ubuntu.com/7228419/ [22:35] which I finally got to give me from charm test -h [22:35] cory_fu: pass, fail, skip, and timout, so were hitting timeout