=== ajmitch_ is now known as ajmitch === TheRealMue is now known as TheMue === gary_poster is now known as gary_poster|away === mwhudson is now known as zz_mwhudson [03:38] hey, can anyone help with charm-tools ? [03:47] bangorang!!!!!!!!!!!!! [03:48] +1, insightful [03:50] :) [03:50] I thought I remember 13.04 being pushed as the default server launch [03:51] I dont want 12.04 [03:51] webbrandon: 13.04 has been retired: https://wiki.ubuntu.com/Releases [03:51] how about 13.10 [03:51] :) [03:52] kidding [03:54] can't find anything in the code for aws to handle this image-stream configuration setting that sits in the template. I just wanted to launch a current version. [04:14] sarnold: this is useful : juju metadata validate-images -s saucy -r us-east-1 [04:41] Hi -- I'm trying to bootstrap azure, I keep getting 'The affinity group name is empty or was not specified', my storage-account-name is in West US, my location is West US. Is there something I'm missing. I get the same error on Trunk and 1.17 (trusty) [04:45] 1.16 too... must be something I'm doing. [05:01] ya, switching to the east us sorted it. /me ignores for now === CyberJacob|Away is now known as CyberJacob === mthaddon` is now known as mthaddon [09:04] hi, can i launsch an "instance" in openstack without the dashbaord? because the dashboard does not work properly, i.e. the page "instances & volumes" ends up in an "interal server error" when clicking on it. [09:22] I'm using the openstack charms to deploy Havana. I'm setting the mtu on the quantum-gateway via dhcp to 1546 but something seems to be reverting it. Can you think of anything that might account for that ? [09:54] gnuoy, odd [09:54] I can't see anything in the charm that touches the MTU [09:54] https://code.launchpad.net/~openstack-charmers/charms/precise/quantum-gateway/trunk [09:55] jamespage, and none of the neutron services are likely to fiddle with it themselves ? [09:55] So quick question. With Ubuntu 14.04 is Openstack deployment with Juju going to get any better? [09:55] Or any cool new features being added? [09:56] Lord_Set, mailing incremental improvements; ML2 plugin will be the default for neutron [09:56] Awesome [09:57] What about Cisco Nexus/Neutron integration via Juju? Well the Openstack/Neutron side of it [09:57] Lord_Set, some guys from Cisco have been working on that - they did an announce a few weeks ago [09:57] the changes still need to land in the official charms but that is planned [09:57] Yay great news. I missed that somehow. [09:58] mailing == mainly [09:58] gnuoy, I don't think so [09:58] I love my Nexus stuff [09:58] jamespage, ok, thanks for your time. I'm going to redeploy the node in question and see what happens [09:59] gnuoy, ok [09:59] The Nexus API makes VM deployment a lot less of a headache but takes some time to setup and configure properly === alexlist` is now known as alexlist [10:02] MAAS and Juju and soon going to dominate Mirantis. Trying to use Mirantis made me want to cry after a bit. Its deployment model is so rigid and to change or customize anything about deployment via Mirantis just plain hurts. [10:14] Lord_Set, http://blogs.cisco.com/datacenter/availability-of-the-nexus-1000v-cloud-networking-platform-for-openstack-unveiled-at-cisco-live/ [10:26] Thanks James [10:28] That is beyond awesome and I cannot wait! Cisco is doing some amazing things with Openstack and virtual machine networking. [13:11] hello everyone, I was hoping someone could help me debug why a LXC container where I ran juju bootstrap (as per https://juju.ubuntu.com/docs/config-LXC.html) now can not connect to the internet [14:00] jamespage, after redeploying the quantum-gateway I see eth0 has the desired mtu setting. Looking at the other interfaces br0 also has it but the rest don't http://pastebin.ubuntu.com/6970990/ is that ok ? [14:03] gnuoy, yup [14:03] jamespage, excellent, thanks === benji_ is now known as benji === rektide_ is now known as rektide === medberry is now known as Guest5611 === Guest5611 is now known as med___ === frobware is now known as zz_frobware [20:49] A few of us just today started seeing "ERROR empty image-stream in environment configuration" when trying to deploy.. has something changed to cause this to happen? [20:52] cjohnston: heh, just hitting up people on this in #juju-dev [20:52] cjohnston: this on ec2? [20:52] natefinch: ^ makes me wonder if something is down on the simple streams side perhaps? or something like that? [20:52] rick_h__: canonistack [20:52] cjohnston: ok [20:53] rick_h__: https://bugs.launchpad.net/juju-core/+bug/1217397 ? [20:53] <_mup_> Bug #1217397: image-stream configuration option for openstack provider

[20:53] cjohnston: yea, it seemed to be azure only but something today is causing issues with it on other providers [20:54] :-( [20:54] cjohnston: what juju version? [20:54] I've got another person testing on precise, I'm on trusty 1.17.2 [20:54] the newest upload in t [20:54] 1.17.2-0ubuntu4 [20:54] I have 1.17.2-0ubuntu4 and I'm seeing it too [20:54] ok, well at least that bit is common [20:54] natefinch: this is looking like a bigger issue? [20:55] sinzui: ^ [20:55] I've been able to bootstrap and deploy just fine earlier today though (with no upgrade in between) [20:55] plars: yea, why I'm wondering if it's something associated with finding images not juju itself [20:58] rick_h__: that error message is definitely a juju error message, it's trying to validate your configuration.... but for some reason it's barfing on the image-stream value, even though it's supposed to be fine for it to be unset [21:15] plars: on trusty? [21:15] rick_h__: yes [21:16] natefinch: my teammate got it to run using 1.17.1 on precise [21:18] rick_h__: hmmm... I'll bring it up with the rest of the devs. Seems there's something funky going on with 1.17.2 [21:18] natefinch: thanks. Did you want me to file a bug or are you ok from here then? [21:22] rick_h__: bugs are always good. Help prevent things from falling through the cracks [21:26] natefinch: filed https://bugs.launchpad.net/juju-core/+bug/1283246 [21:26] <_mup_> Bug #1283246: cannot deploy on 1.17.2 on trusty: error on image-stream [21:26] plars: cjohnston please add any notes and track & [21:26] thanks rick_h__ [21:30] rick_h__: thanks. I'll make sure it gets looked at, ASAP. [21:31] natefinch: could we maybe get a slightly older version thrown in a ppa somewhere? [21:34] cjohnston: 1.16 should be what's in ppa:juju/stable [21:34] natefinch: don't think it's built for trusty [21:34] or I lied [21:34] and it is [21:34] https://launchpad.net/~juju/+archive/stable?field.series_filter=trusty [21:34] ack [21:35] 1.16 complains about alot of config fields :-/ [21:36] unfortunately, forwards compatibility is not something we've spent any time on === med___ is now known as med_ [22:16] has there been a recent change to the charmstore api? [22:18] lazyPower: how so? [22:18] and by store you mean juju store or charmworld? [22:19] rick_h__: i mean charmworld - my apologies. I keep forgetting there are 2 api's [22:19] lazyPower: what change are you getting. There was a release this week, but trying to recall what api changes might have existed [22:20] i have amulet tests that were working an hour ago - now i'm getting a non-helpful stacktrace bubbling up from charmworldlib. I'm trying to decompose to code to find the endpoint its calling - that way i can rule out a temporary hiccup in charmworldlib and validate its the code [22:21] lazyPower: gotcha, yea nothing changed today in charmworld. I'd have to see more details to help trace. Let me make sure charmworld is reporting healthy [22:21] bah, elasticsearch bit it [22:24] cool so i'm not going crazy then [22:24] I am also encountering an error with juju lazyPower, and rick_h__ [22:24] rick_h__, I have a traceback that looks like this http://pastebin.ubuntu.com/6973456/ [22:25] I can grab logs if you need them. [22:25] mbruzek_: its because the search provider tanked [22:26] lazyPower: mbruzek_ if it's charmworld I'm working with webops on it atm [22:26] thats the one :) Thanks rick_h__ [22:47] Could I launch a t1.micro bootstrap to manage m1 type instances [22:47] webbrandon: thats not recommended at all. since hte controller node runs zookeeper and what not [22:48] you'll run out of memory on the bootstrap node when multiple nodes converge [22:48] juju doesn't use zookeeper [22:48] wait, what? [22:48] juju has used mongodb for 8 months [22:48] i'm just full of misinformation today [22:49] lazyPower: you sending me down the wrong path huh??? [22:49] webbrandon: well, its still not recommended... [22:49] lol [22:49] regardless of my fud [22:50] sinzui: ty for correcting me. I'm thinking of chef-server not juju [22:50] my internal wiring is getting crossed looking at this charm [22:50] webbrandon, juju's boostrap node has been run on 512m, but I think 1G is the minimum. Each deployment may require its own sizes just because the services need more storage or ram [22:51] webbrandon, I often bootstrap with --constraints "mem=1G" the each deploy has --constrains "mem=2G" [22:51] sinzui: thank you [22:52] sinzui: I am thinking of launching t1 because its free to manage a bitcoin pooling service [22:53] unfortunatly i have to run m class instance just to install the software [22:54] webbrandon, alas juju doesn't let you specify an aws/hp instance-type (yet). The constraints only guide the selection of an instance type. I think the the HP rules have issues in fact. 2G on aws gets me a more memory that 2G on Hp [22:55] I understand how to launch the right instance type. === CyberJacob is now known as CyberJacob|Away === medberry is now known as med_ [23:41] lazyPower: should be coming back [23:42] rick_h__: awesome. I'll run a test and see if its cleared up [23:42] thank you for driving this rick_h__ [23:42] lazyPower: heh, last one around on a friday when it goes boom :) [23:42] lucky me! [23:43] We'll be meeting up in vegas soon enough. I'll buy the first beer :) [23:44] thanks rick_h__ [23:44] now you can't say it is a thank-less job. [23:45] mbruzek_: hah, good call