=== freeflying_away is now known as freeflying [00:14] lazyPower: good to hear [00:14] lazyPower: I'm giving a juju demo this afternoon to [00:15] I'm excited to get started. The learning curve to just get started was an hour vs the 3 months i've been looking at and working off and on with competing frameworks. [00:16] \o/ [00:16] Honestly, i'm surprised you dont throw that down as one of the benefits. "Get started in about an hour - no long learning curve, no extra cirricular setup required if your PC supports LXC. JUJU - its made from magic" [00:18] :) [00:21] I suppose its now obvious that I program for a marketing firm, looking at the marketing angle first - not the raw and lean [00:22] awesomeness thats infront of me [00:25] Where are you located Thumper? If you don't mind me asking. [00:25] lazyPower: new zealand [00:25] east australia [00:26] davecheney: :P [00:26] seems legit [00:26] australia is the west island [00:26] shit, that means I live in alice sprints [00:26] springs [00:28] is this a bad thing? [00:29] lazyPower: do like living in a desert ? [00:30] I've lived in west texas, it wasn't all bad. === gary_poster is now known as gary_poster|away [01:22] Do i need to destroy my local environment after every reboot or am I not doing something to cleanly shut down my local juju environment? [01:23] lazyPower: what do you want do to ? [01:23] destroy your enviroment, or reboot your computer ? [01:23] Well, I reboot the computer with the juju environment active out of ignorance [01:24] now i cant seem to get the nodes to come back online, juju status hangs with no output. [01:24] lazyPower: sounds like a bug [01:24] it would be nice if rebooting wouldn't destroy all your work [01:24] agreed [01:24] I'll file it === julianwa_ is now known as julian === julian is now known as juliawa === juliawa is now known as julianwa === defunctzombie_zz is now known as defunctzombie [03:39] everyone gone for the day? [03:39] * davecheney waves [03:40] hey davecheny :), you wouldn't happen to know anything about th quntum charm would you? [03:40] quantum charm, sorry [03:41] zradmin: probably not much [03:41] what is your question [03:42] i almost have openstack deployed via juju... but my instances aren't able to receive any traffic [03:43] davecheney: I get alot of these messages in the syslog Sep 30 17:38:03 m7q49 dnsmasq-dhcp[2371]: DHCP packet received on qbr480533b5-f8 which has no address [03:45] zradmin: given i know nothing about the quantum charm [03:45] i'm still going to say that is unrelated to the charm [03:46] all the charm does it just do apt-get install and twiddle some config files [03:46] the rest is going to the be product itself [03:47] yeah... I've inspected the configs plenty of times and the physical interfaces are all active so that should be fine [03:47] zradmin: my best guess of that a dhcp request packet arrived on qbr480533b5-f8 [03:47] for a host that dnsmasq has not been configured to service [03:48] the odd thing is I can see it create the ports etc in the syslog as I create/destroy instances [03:51] zradmin: is the lan isolated ? [03:51] remember, i know nothign about quantum [03:52] yeah its isolated, I may dig a bit deeper into dnsmasq to see if i can manually confirm the port has been created [03:53] so that helped to point me in a new direction :) also its good to vent about it a little so thanks! [03:55] win/win [03:59] zradmin, by instances you mean vms in openstack? [04:01] you might able to get a little farther with neutron admin cli [04:08] hazmat: im using quantum atm [04:09] zradmin, quantum == neutron fwiw [04:09] oic [04:09] rename due to trademark [04:09] anyone around able to help me with local environment key issue ? [04:09] hazmat: nice [04:10] I get this error - Permission denied (publickey,password) when trying to 'juju debug-log' after running 'sudo juju bootstrap' [04:11] I have ssh keys in ~/.ssh/id_rsa, id_rsa.pub [04:11] omgponies: what happens when you do juju ssh 0 [04:11] same error [04:11] i'm not 100% sure if debug log works on the local provider atm [04:12] omgponies: are you using MAAS as well? [04:12] i have a passphrase on the key ... but I've used ssh-add to cache the phrase [04:12] nah, but I do use EC2 with no issues [04:12] just so slloooow [04:13] omgponies: what version of juju are you on? [04:14] omgponies: i suspect we're not passing the right flags when we fork ssh [04:14] 1.14.1-raring-amd64 [04:14] you could try changing /etc/ssh/ssh_config to always pass ForwardAgent: yes [04:16] doesn't seem to help === defunctzombie is now known as defunctzombie_zz [04:17] omgponies: would you be able to raise a bug please [04:17] hazmat: is that in havana btw? i can't seem to find any nuetron console. 12.04 on grizzly [04:17] something like 'juju ssh/debug-log does not work on local provider when passphrase protected ssh key is in used' [04:17] will do ... in juju-core @ launchpad ? [04:19] yup [04:22] zradmin, oh.. yeah.. it havana [04:23] hazmat: is that stable at all yet or still in testing? [04:24] zradmin, dunno.. i haven't used it myself outside of a lab [04:24] zradmin, but the cli here should apply to quantum on grizzly http://docs.openstack.org/user-guide-admin/content/neutron_client_commands.html [04:26] hazmat: yeah its pretty much the same as quantums... it looks like everything creates itself properly for the vms, but i can't novnc/ssh to them because nova says it has no public address === defunctzombie_zz is now known as defunctzombie [04:32] zradmin, so typical setup gets a private tenant network and then a floating ip network for ingress.. maybe try create a floating ip and attach to vm [04:32] er. for public access [04:33] and by typical i mean the charm setup [04:34] hazmat: yeah i've done that as well, eth0 and eth1 (public) are both there but I had to manually create br-ex for some reason. Internal should have still worked as the vswitch creates an interface and attaches to br-int which is supposed to go through br-tun in a gre setup right? I never see any traffic crossing br-tun though [04:35] jamespage, ping ^ === defunctzombie is now known as defunctzombie_zz [04:43] hazmat: so jamespage is who i have to hunt down? [04:44] zradmin, or adam_g.. although negronjl might be able to help as well [04:45] * hazmat heads to sleep === defunctzombie_zz is now known as defunctzombie === defunctzombie is now known as defunctzombie_zz === CyberJacob|Away is now known as CyberJacob === freeflying is now known as freeflying_away === gary_poster|away is now known as gary_poster [12:40] hazmat, ping re https://code.launchpad.net/~james-page/juju-deployer/fixup-to-for-strings/+merge/188296 [12:40] I also note a few other fixes for juju-deployer - if we want to push another point release this week would be good for saucy [12:41] jamespage, cool [12:41] hazmat,scratching my head over how to make the --to work well with nova-compute/ceph/swift all on the same nodes [12:42] I'll ping you an email to explain [12:42] jamespage, thanks.. i noted some fixes/bugs from last week as well. === TheRealMue is now known as TheMue [12:50] How do Juju compare with Docker.io ? [12:51] hazmat, email in your inboc [12:51] x [12:53] Do Juju require that your servers have KVM, or LXC installed and setup? [12:53] jamespage, got a moment for g+? [12:57] hazmat, 20mins? [12:57] jamespage, sounds good [12:57] just need to catch lunch before it becomes late afternoon [13:04] If I start up a charm.. Will I be able to SSH into it and make ad-hoc changes? (I know this isn't recommended, but is it still possible?) === freeflying_away is now known as freeflying [13:18] GMT morning everyone. [13:29] hazmat, good now? [13:38] o/ lazyPwork glad your demo went well [13:47] jamespage, doh.. yeah.. now. was analyzing a cts issue [13:59] marcoceppi: another MP for charm-tools: https://code.launchpad.net/~adeuring/charm-tools/check-maintainer-branch-owner/+merge/188586 === hatch_ is now known as hatch [14:20] adeuring: thank you! [14:23] heelo I'm trying to configure my charm to use haproxy, I've added the hook to my charm but I keep getting: [14:24] $ juju add-relation my-service:website haproxy:reverseproxy [14:24] error: service "my-service" has no "website" relation [14:26] vds: pastebin your metadata.yaml to paste.ubuntu.com please [14:28] marcoceppi, http://paste.ubuntu.com/6179761/ [14:30] vds: you have an indentation error with nrpe-external-master, that and the two lines below it need to be indented by to line up with mongo [14:30] otherwise, your metadata.yaml looks fine and should connect to haproxy without issue [14:31] marcoceppi, thanks! [14:31] vds: adjust that, destroy-environment, re-bootstrap (for good measure) then deploy/relate again [14:31] it should work [14:31] if not, let us know [14:35] marcoceppi: Same. I've generated quite a bit of interest. I feel that I'm going to have a shift in management methodology coming once I have a functional setup in AWS to show them that juju orchestrated. [14:36] marcoceppi: have you come up with a solid way to not lose your progress in the local dev environment between reboots? I've had to completely wipe my local environment and start over every time I reboot due to kernel updates. [14:43] marcoceppi: I'm not sure what the plans are wrt amulet, but I wonder about a configurable test charm for testing relation that works a bit like https://gist.github.com/wavydavy/6779561 [14:44] bloodearnest: that's covered in amulet, I'll give you an example [14:44] lazyPwork: my local environment survives reboots, is this with the vagrant image? [14:44] marcoceppi: sweet, thought it might be [14:47] lazyPwork: what version of juju? we fixed the local reboot thing a while back [14:47] also hi! [14:50] bloodearnest: https://gist.github.com/marcoceppi/6779616 [14:50] jcastro: latest backport from raring [14:50] i'm not infront of my PC with juju installed, i can do some further information aggregation this evening if you're up for helping me get you the data you need [14:51] well, getting launchpad the data the juju team would need [14:51] i sometimes forget you're only superman Jorge :) [14:53] marcoceppi: thanks, that looks good. Some questions though: [14:53] bloodearnest: fire away [14:53] 1. which charm are you testing wordpress or msql? [14:55] 2. can I intercept/change what mysql:db relation supplies to wordpress:db? so I can test for how the wordpress charm handles misconfiguration for example [14:55] bloodearnest: in this example it doesn't actually matter, but this was copied from the wordpress test I was writing. Amulet automatically deploys from the store, unless JUJU_TEST_CHARM environment variable is set, if it matches the charm it's adding (wordpress) it uses ../../ (file is assumed to be in charm/tests) to do the deploy [14:55] bloodearnest: no, not at this time. It only listens, but that can certainly be a feature added [14:55] bloodearnest: something for a 1.1 release, just need to get 1.0 out first :) [14:56] marcoceppi: right, so that first answer makes sense. I think that makes sense in the wordpress setting, as a wordpress deployment will always have a mysql charm, it don't work any other way [14:57] marcoceppi: but haproxy, or squid, it's much more general. [14:58] it could be an arbitary charm supporting a http interface [14:58] sure [14:58] marcoceppi: cool [15:00] marcoceppi: my ideal world, I would only have to deploy a single charm-under-test, and all the rest of the relation add/removes can be faked under my control :) [15:04] bloodearnest: Right, that's something I'm also interested in, since that would be like light-weight testing. Something like spin up an LXC with the charm mounted in a psuedo-deployed state, then feed events to a psuedo hook-environment and allow for fast light weight tests. This is a little more blunt. It's using Juju Deployer to throw this up in an environment and actually hash it out [15:07] marcoceppi: right, which you'd definitely want to do to check out things like wordpresses memcached integration. And wordpress can't do anything without a db, so you'll probably always need that === kentb-out is now known as kentb === freeflying is now known as freeflying_away === hatch_ is now known as hatch [15:18] Hola.. I m deploying cinder (grizzly) on some machine, and strangely enough, the installation proceed without error but once the charm is installed, there is no cinder-volume created at all [15:18] no error in juju logs, nor in juju status [15:18] any idea what to do ? [15:19] the cinder charm has been deployed witrh the following config http://pastebin.ubuntu.com/6179969/ === freeflying_away is now known as freeflying === defunctzombie_zz is now known as defunctzombie === defunctzombie is now known as defunctzombie_zz [15:50] Hey -- has anyone tried to open up a firewall for juju-core bootstrap? Specifically downloading to tools from the public bucket? Is there any gotchas? [17:02] a [17:05] US? [17:16] question, did i get credit for this? http://bazaar.launchpad.net/~charmers/juju-core/docs/revision/137 [17:24] I'm so rustled. We just got the mandate from above to move to Lync. There goes my hubot installation until I spin up a darknet. [17:24] marcoceppi: charm powered darknet? check. [18:25] lazyPwork: what did you have your hubot doing === jose- is now known as jose === _mup__ is now known as _mup_ === adam_g` is now known as adam_g === defunctzombie_zz is now known as defunctzombie [20:04] Juju-Gui Charm API did not respond | http://askubuntu.com/q/352427 === elmo__ is now known as elmo [20:10] rektide: He hooks into janky for jenkins communication, and he executes a few of our internal jobs like kicking off backups === defunctzombie is now known as defunctzombie_zz === defunctzombie_zz is now known as defunctzombie === freeflying is now known as freeflying_away === thumper is now known as thumper-afk === medberry is now known as med_ === CyberJacob is now known as CyberJacob|Away