2014/02/21/#juju.txt

=== 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] <davecheney> hey, can anyone help with charm-tools ?
[03:47] <webbrandon> bangorang!!!!!!!!!!!!!
[03:48] <davecheney> +1, insightful
[03:50] <webbrandon> :)
[03:50] <webbrandon> I thought I remember 13.04 being pushed as the default server launch
[03:51] <webbrandon> I dont want 12.04
[03:51] <sarnold> webbrandon: 13.04 has been retired: https://wiki.ubuntu.com/Releases
[03:51] <webbrandon> how about 13.10
[03:51] <webbrandon> :)
[03:52] <webbrandon> kidding
[03:54] <webbrandon> 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] <webbrandon> sarnold: this is useful :  juju metadata validate-images -s saucy -r us-east-1
[04:41] <dpb1`> 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] <dpb1`> 1.16 too... must be something I'm doing.
[05:01] <dpb1`> 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] <tomixxx3> 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] <gnuoy> 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] <jamespage> gnuoy, odd
[09:54] <jamespage> I can't see anything in the charm that touches the MTU
[09:54] <jamespage> https://code.launchpad.net/~openstack-charmers/charms/precise/quantum-gateway/trunk
[09:55] <gnuoy> jamespage, and none of the neutron services are likely to fiddle with it themselves ?
[09:55] <Lord_Set> So quick question. With Ubuntu 14.04 is Openstack deployment with Juju going to get any better?
[09:55] <Lord_Set> Or any cool new features being added?
[09:56] <jamespage> Lord_Set, mailing incremental improvements; ML2 plugin will be the default for neutron
[09:56] <Lord_Set> Awesome
[09:57] <Lord_Set> What about Cisco Nexus/Neutron integration via Juju? Well the Openstack/Neutron side of it
[09:57] <jamespage> Lord_Set, some guys from Cisco have been working on that - they did an announce a few weeks ago
[09:57] <jamespage> the changes still need to land in the official charms but that is planned
[09:57] <Lord_Set> Yay great news. I missed that somehow.
[09:58] <jamespage> mailing == mainly
[09:58] <jamespage> gnuoy, I don't think so
[09:58] <Lord_Set> I love my Nexus stuff
[09:58] <gnuoy> jamespage, ok, thanks for your time. I'm going to redeploy the node in question and see what happens
[09:59] <jamespage> gnuoy, ok
[09:59] <Lord_Set> 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] <Lord_Set> 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] <jamespage> Lord_Set, http://blogs.cisco.com/datacenter/availability-of-the-nexus-1000v-cloud-networking-platform-for-openstack-unveiled-at-cisco-live/
[10:26] <Lord_Set> Thanks James
[10:28] <Lord_Set> That is beyond awesome and I cannot wait! Cisco is doing some amazing things with Openstack and virtual machine networking.
[13:11] <nessita> 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] <gnuoy> 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] <jamespage> gnuoy, yup
[14:03] <gnuoy> 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] <cjohnston> 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] <rick_h__> cjohnston: heh, just hitting up people on this in #juju-dev
[20:52] <rick_h__> cjohnston: this on ec2?
[20:52] <rick_h__> natefinch: ^ makes me wonder if something is down on the simple streams side perhaps? or something like that?
[20:52] <cjohnston> rick_h__: canonistack
[20:52] <rick_h__> cjohnston: ok
[20:53] <cjohnston> rick_h__: https://bugs.launchpad.net/juju-core/+bug/1217397  ?
[20:53] <_mup_> Bug #1217397: image-stream configuration option for openstack provider <config> <openstack-provider> <juju-core:Fix Released by wallyworld> <https://launchpad.net/bugs/1217397>
[20:53] <rick_h__> cjohnston: yea, it seemed to be azure only but something today is causing issues with it on other providers
[20:54] <cjohnston> :-(
[20:54] <rick_h__> cjohnston: what juju version?
[20:54] <rick_h__> I've got another person testing on precise, I'm on trusty 1.17.2
[20:54] <cjohnston> the newest upload in t
[20:54] <cjohnston> 1.17.2-0ubuntu4
[20:54] <plars> I have 1.17.2-0ubuntu4 and I'm seeing it too
[20:54] <rick_h__> ok, well at least that bit is common
[20:54] <rick_h__> natefinch: this is looking like a bigger issue?
[20:55] <rick_h__> sinzui: ^
[20:55] <plars> I've been able to bootstrap and deploy just fine earlier today though (with no upgrade in between)
[20:55] <rick_h__> plars: yea, why I'm wondering if it's something associated with finding images not juju itself
[20:58] <natefinch> 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] <rick_h__> plars: on trusty?
[21:15] <plars> rick_h__: yes
[21:16] <rick_h__> natefinch: my teammate got it to run using 1.17.1 on precise
[21:18] <natefinch> 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] <rick_h__> natefinch: thanks. Did you want me to file a bug or are you ok from here then?
[21:22] <natefinch> rick_h__: bugs are always good.  Help prevent things from falling through the cracks
[21:26] <rick_h__> 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 <juju-core:New> <https://launchpad.net/bugs/1283246>
[21:26] <rick_h__> plars: cjohnston please add any notes and track &
[21:26] <cjohnston> thanks rick_h__
[21:30] <natefinch> rick_h__: thanks.  I'll make sure it gets looked at, ASAP.
[21:31] <cjohnston> natefinch: could we maybe get a slightly older version thrown in a ppa somewhere?
[21:34] <natefinch> cjohnston: 1.16 should be what's in ppa:juju/stable
[21:34] <rick_h__> natefinch: don't think it's built for trusty
[21:34] <rick_h__> or I lied
[21:34] <rick_h__> and it is
[21:34] <rick_h__> https://launchpad.net/~juju/+archive/stable?field.series_filter=trusty
[21:34] <cjohnston> ack
[21:35] <cjohnston> 1.16 complains about alot of config fields :-/
[21:36] <natefinch> unfortunately, forwards compatibility is not something we've spent any time on
=== med___ is now known as med_
[22:16] <lazyPower> has there been a recent change to the charmstore api?
[22:18] <rick_h__> lazyPower: how so?
[22:18] <rick_h__> and by store you mean juju store or charmworld?
[22:19] <lazyPower> rick_h__: i mean charmworld - my apologies. I keep forgetting there are 2 api's
[22:19] <rick_h__> lazyPower: what change are you getting. There was a release this week, but trying to recall what api changes might have existed
[22:20] <lazyPower> 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] <rick_h__> 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] <rick_h__> bah, elasticsearch bit it
[22:24] <lazyPower> cool so i'm not going crazy then
[22:24] <mbruzek_> I am also encountering an error with juju lazyPower, and rick_h__
[22:24] <mbruzek_> rick_h__, I have a traceback that looks like this http://pastebin.ubuntu.com/6973456/
[22:25] <mbruzek_> I can grab logs if you need them.
[22:25] <lazyPower> mbruzek_: its because the search provider tanked
[22:26] <rick_h__> lazyPower: mbruzek_ if it's charmworld I'm working with webops on it atm
[22:26] <lazyPower> thats the one :) Thanks rick_h__
[22:47] <webbrandon> Could I launch a t1.micro bootstrap to manage m1 type instances
[22:47] <lazyPower> webbrandon: thats not recommended at all. since hte controller node runs zookeeper and what not
[22:48] <lazyPower> you'll run out of memory on the bootstrap node when multiple nodes converge
[22:48] <sinzui> juju doesn't use zookeeper
[22:48] <lazyPower> wait, what?
[22:48] <sinzui> juju has used mongodb for 8 months
[22:48] <lazyPower> i'm just full of misinformation today
[22:49] <webbrandon> lazyPower: you sending me down the wrong path huh???
[22:49] <lazyPower> webbrandon: well, its still not recommended...
[22:49] <webbrandon> lol
[22:49] <lazyPower> regardless of my fud
[22:50] <lazyPower> sinzui: ty for correcting me. I'm thinking of chef-server not juju
[22:50] <lazyPower> my internal wiring is getting crossed looking at this charm
[22:50] <sinzui> 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] <sinzui> webbrandon, I often bootstrap with --constraints "mem=1G" the each deploy has --constrains "mem=2G"
[22:51] <webbrandon> sinzui: thank you
[22:52] <webbrandon> sinzui: I am thinking of launching t1 because its free to manage a bitcoin pooling service
[22:53] <webbrandon> unfortunatly i have to run m class instance just to install the software
[22:54] <sinzui> 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] <webbrandon> I understand how to launch the right instance type.
=== CyberJacob is now known as CyberJacob|Away
=== medberry is now known as med_
[23:41] <rick_h__> lazyPower: should be coming back
[23:42] <lazyPower> rick_h__: awesome. I'll run a test and see if its cleared up
[23:42] <lazyPower> thank you for driving this rick_h__
[23:42] <rick_h__> lazyPower: heh, last one around on a friday when it goes boom :)
[23:42] <rick_h__> lucky me!
[23:43] <lazyPower> We'll be meeting up in vegas soon enough. I'll buy the first beer :)
[23:44] <mbruzek_> thanks rick_h__
[23:44] <mbruzek_> now you can't say it is a thank-less job.
[23:45] <rick_h__> mbruzek_: hah, good call