=== thumper is now known as thumper-afk [01:18] marcoceppi: you got it [01:19] lazyPower: I cowboy'd [01:19] i see this [01:19] i was baking a cake.. sorry. :P [01:19] omgcake [01:19] lazyPower: I do have another plugin for your review [01:19] red velvet with cream cheese icing. My inner fat boy is happy [01:19] my outer fatboy is cursing me [01:20] and I want to set up some travis testing to do like basic "Are all files executable, do they respond to --help and --description" etc [01:20] I'm not allowed to bake anymore, because I keep eating all the things I bake [01:20] marcoceppi: ping me when you have the open MR, and i'll pull + review [01:37] lazyPower: Hey I made a Juju plugin and did a pull request (https://github.com/juju/plugins/pull/3), mind reviewing it and merging it upstream? [01:38] JoshStrobl: ack, on it [01:40] lazyPower: Thanks [01:41] Anyways, going to bed now. Talk to you guys later! === thumper-afk is now known as thumper [01:50] lazyPower: uh, it's not executable, probably shouldn't have been merged yet [01:50] marcoceppi: i was going to fix it and push a patch [01:51] no bueno? [01:51] ah [01:51] i'm catching my repository back up, i cloned upstream on my laptop instead of my fork [01:51] so, doing the shuffle :) [01:51] * marcoceppi makes travis.yml file for linting [01:51] np np [01:58] marcoceppi: open PR, i wont merge it :P [01:59] I'm still making the travis file. Trying to figure out how to test that all files are executable, and respond to --description and --help [02:01] marcoceppi: you can have the travis.yml call a bash file right? [02:02] lazyPower: I think so [02:02] lazyPower: in the mean time [02:02] https://github.com/juju/plugins/pull/5 [02:03] bah, I haven't added my prettyprint plugin yet [02:05] yep [02:05] i was about to say, its giving me an invalid option for prettyprint [02:05] lazyPower: yeah, let me add pprint [02:05] but also fix it so it doesn't suck [02:06] well there goes my idea for a plugin :P [02:06] i was going to parse juju status to condense it down to a more compact format [02:09] marcoceppi: idea for the repository - an ansilliary script to enable/disable some of these plugins. As the repository grows this is going to get bloated [02:09] lazyPower: if the repo grows I'll build the plugin repository site [02:09] so you can just run `juju plugin install ` [02:10] bueno [02:10] that would be a nice problem to have [02:10] * marcoceppi already registered jujuplugins.com [02:10] yeah, i recall all of that talk [02:10] jorge and hatch talked you out of going full monty out the gate [02:16] yeah, I was all ready to go [02:16] marcoceppi: brb, moving my quassel-core === lazypower_ is now known as lazypower [02:36] sweet [02:36] that was painless === lazyPower_ is now known as lazyPower [02:37] marcoceppi: if you ever need to do that, stop the quassel-core daemon, and copy the .conf and .sqlite db (or pgsql if you're using that), dont copy the pem, it'll break your ssl connections. [02:37] lazyPower: cool === freeflying_away is now known as freeflying === mwhudson is now known as zz_mwhudson === timrc is now known as timrc-afk === CyberJacob|Away is now known as CyberJacob === freeflying is now known as freeflying_away === freeflying_away is now known as freeflying [09:11] marcoceppi, mysql-server-5.6 is now in trusty btw [09:44] how do you maintain the charmed software up-to-date? say a new point release is out, no changes to the charm need to be made and users want to upgrade, is there a hook you use to install updates? [12:24] if the documentation says that "Internet access will be restricted from the testing host", what does that mean? does that mean that a charm trying to get the sources of the software might fail or that the test code cannot access the outside? [13:02] jamespage: awesomeee [13:02] marcoceppi, it needs some polish but its functional [13:03] cargill: that means that we also test how well the charm works offline, but most testing environments have no restrictions === timrc-afk is now known as timrc === freeflying is now known as freeflying_away === natefinch is now known as natefinch-afk === amol_ is now known as amol [15:29] I seem to be hitting something that looks very much like https://bugs.launchpad.net/juju-core/+bug/1275282 which is supposed to be fixed, but I'm on 1.17.2-0ubuntu3 [15:29] <_mup_> Bug #1275282: juju bootstrap fails on HP Cloud unless tools are already uploaded [15:30] according to the bug, it was fixed in 1.17.1 though [15:31] plars: what version of juju are you running, `juju version`? [15:31] marcoceppi: 1.17.2-trusty-amd64 [15:31] plars: are you abel to bootstrap? and if not could you pastebin `juju boostrap --debug --show-log`? [15:32] marcoceppi: no, bootstrap doesn't work. I have output from it but not with those options, let me do that... give me a moment [15:50] marcoceppi: ok, with some bits scrubbed for security and some for brevity: http://paste.ubuntu.com/6931844/ [15:51] plars: do you have a custom image stream for this environment? [15:51] marcoceppi: no, I don't think so [15:52] it seems like it's failing to get https://streams.canonical.com/juju/tools/releases/juju-1.17.0-precise-amd64.tgz but that address is resolvable [15:52] I also have no idea why it's picking 1.17.0 instead of 1.17.2 tools [15:53] plars: if you open your environments.yaml, do you see anything with image-metatdata or tools-url for your hp cloud environment? [15:54] image-metadata-url: https://region-a.geo-1.objects.hpcloudsvc.com/v1/11289530460295/images [15:54] marcoceppi: yes [15:54] plars: ah, okay, comment that line out of your environments.yaml [15:55] plars: then `rm -f ~/.juju/environments/.jenv` [15:55] marcoceppi: which I think I just got from a wiki, and I thought it was necessary to make it work [15:55] I'll try it without though [15:55] plars: then bootstrap again with --debug --show-log, etc [15:55] see if it works now [15:55] it was needed for 1.17.0, but since 1.17.1 streams are fixed [15:55] ah, ok [15:57] marcoceppi: I'm trying it now, but I already saw this fly by: [15:57] 2014-02-14 15:56:33 INFO juju.environs.bootstrap bootstrap.go:58 picked newest version: 1.17.0 [15:58] marcoceppi: doesn't work at all without that line: [15:59] 2014-02-14 15:56:36 ERROR juju.cmd supercommand.go:294 cannot start bootstrap instance: index file has no data for cloud {region-a.geo-1 https://region-a.geo-1.identity.hpcloudsvc.com:35357/v2.0/} not found [15:59] sinzui: ^^ [16:00] I have never seen that error, but I understand that HP cloud is very ill for new users and their new regions === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha [17:03] Wow, Ubuntu moving to systemd. This is exciting news. [17:06] Well, at least improving support for it (according to Shuttleworth's blog). Anyways, back to work I go. [17:08] sinzui: any ideas how to get around it? [17:09] sinzui: I think when I previously tried bootstrapping and deploying things from saucy it worked, but now on trusty it does not [17:15] plars, which 1.17.2 and 1.16.6 are working fine in region: az-3.region-a.geo-1 (looking at recent test runs) [17:16] sinzui: I *am* running 1.17.2 [17:16] 1.17.2-0ubuntu3 specifically [17:17] plars, I just bootstrapped on Hp from trusty with that version [17:18] plars, so I suspect the region is the issue, and that does relate to what I have head about sick regions [17:20] sinzui: I just specify region-a.geo-1 in my environments.yaml, so I guess I need the az-3 there also? [17:20] nope [17:20] same error [17:20] ERROR bootstrap failed: cannot start bootstrap instance: index file has no data for cloud {region-a.geo-1 https://region-a.geo-1.identity.hpcloudsvc.com:35357/v2.0/} not found [17:22] plars, I definitely authenticated to that URL when using [17:22] region: az-3.region-a.geo-1 [17:22] plars, are you a new HP Cloud user? as of this year? [17:22] sinzui: yes [17:22] sinzui: I've had it bootstrapped before when I was on saucy though [17:22] sinzui: and even had things deployed to it [17:25] plars, this is the config I use to bootstrap a few minutes ago: http://pastebin.ubuntu.com/ [17:26] sinzui: that's just pastbin alone [17:29] plars, sorry, [17:29] plars, http://pastebin.ubuntu.com/6932350/ [17:29] sinzui: ah, you have a tools-url [17:30] I do. that makes download of tools faster, but I don't think it matters for authentication [17:31] plars, for 1.17.x, the key was rename tools-metadata-url. I use the old name because I test on 1.16 too [17:33] sinzui: ok, but before I removed my image-metadata-url I was not getting this error [17:33] sinzui: now I put it back and I get yet a different one [17:34] hmm. [17:34] just out of curiosity, let me try again with 1.16.3-0ubuntu0.13.10.1 (saucy) [17:34] yes, bootstrap worked fine there (same config) [17:37] plars, what was image-metadata-url set too? [17:38] sinzui: https://region-a.geo-1.objects.hpcloudsvc.com/v1/11289530460295/images [17:38] sinzui: according to marcoceppi that used to be needed but not now [17:39] now even if I put it back, I get a totally different error on trusty [17:39] I am not sure the data there is sane....https://region-a.geo-1.objects.hpcloudsvc.com/v1/11289530460295/images/streams/v1/com.ubuntu.cloud:released:imagemetadata.json [17:40] ah, because of the region change [17:40] ok, changing my reation back to just region-a.geo-1, I am getting back to the original failure [17:42] plars, It has less info than http://cloud-images.ubuntu.com/releases/streams/v1/com.ubuntu.cloud:released:hpcloud.json [17:44] I wonder if setting image-metadata-url: https://cloud-images.ubuntu.com/releases [17:44] works [17:44] plars, ^ [17:48] looking at the debug log during boot I can see my image-metadata-url is implicitly http://cloud-images.ubuntu.com/releases === JoshStrobl is now known as JoshStrobl-afk [17:58] sinzui: and with that url for image-metadata-url, I wind up right back at: [17:58] ERROR bootstrap failed: cannot start bootstrap instance: index file has no data for cloud {region-a.geo-1 https://region-a.geo-1.identity.hpcloudsvc.com:35357/v2.0/} not found [17:59] sinzui: I don't understand the internals of it, but I'm guessing there's some index file it's looking for there? [17:59] plar, that is the webservice that does authentication to uses hp-cloud [18:04] plars, file a bug at bugs.launchpad.net/juju-core . I need to pass this to developers, and most have reached their EoW [18:05] plars, they probably want you to "bootstrap --debug" to capture more details about what juju is doing === natefinch-afk is now known as natefinch [19:11] any plans to make juju more PaaS aware? [19:21] viperZ28: can you be a bit more specific in what you're asking? [19:24] I was using local and I just started to look at this yesterday so I am newbie, but I was wondering once JuJu is setup any number of developers can have access to the same environment? [19:26] viperZ28: well, you can allow developers to access the same cloud deployment [19:27] viperZ28: juju is unique in that it's poth paas and not paas at the same time [19:28] lazyPower: let me gather my thoughts on this and come back. I am currently using Stackato (CloudFoundry) so it is a shift in thinking. I agree about it being a PaaS and not a PaaS. [19:28] I was amazed how easy it was to deploy a RabbitMQ clustered environment but I later had issues as some of the nodes crashed. [19:29] viperZ28: you can use juju to deploy cloud foundry :) but also deploy it's own apps. You can even use juju to deploy openstack on bare metal then deploy cloudfoundry on top of the openstack that was deployed by juju using juju [19:29] it's a very unique tool in the ecosytem of tools to use [19:30] Are there plans for vsphere integration? [19:30] and docker [19:40] viperZ28: we're working on docker integration [19:41] Hi when will it possible to start to play with juju/trusty ? [19:42] chris38home_: You can already use juju on trusty. [19:51] yep, but there was no charms for trusty, in charm-gui ? [19:51] in juju-gui [19:52] no juju-gui indeed [19:52] chris38home_: the charms themselves may need some modification before they work on trusty [19:53] chris38home_: I believe you can easily copy them all into a repo of your own, change the targeted version, aim juju to your local repository, and start testing.. [19:53] chris38home_: we've got an effort to add testing to all charms so we can make sure they work on trusty before promoting them [19:53] marcoceppi: nice :) [19:54] You'll probaby see a lot of charms land for trusty a few weeks leanding up to the trusty release [19:54] nice [20:12] sinzui, is manual provider part of qa? [20:13] hazmat, no [20:13] may in the next two weeks though [20:30] I have a few units that show "life: dying" since yesterday, anyway to force restart juju or the services? [20:33] sinzui, would be good, i'm seeing a few regressions on trunk and its being used for prod work fwiw [20:34] viperZ28, version? [20:34] juju version [20:34] 1.16.6-precise-amd64 [20:34] also all my instance show "instance-state: missing", I am using local [20:40] viperZ28: there has been quite a bit of effort into the 1.17 series of juju, which is in ppa:juju/devel right now. I ran into some issues with local prior to the upgrade of -dev series. [20:40] ^ in regards to the local provider. [20:41] thanks, I will take a look. I had heard the code was forked at some point, one python and one go, is there any truth to this? [20:43] That happened a while ago. Juju is mostly Golang now. I think it was the 1.15 series but dont quote me on that. I'll have to dig to find that version where it changed. [20:44] lazyPower: 1.0 [20:44] I believe [20:46] ah, so much further behind us than I was thinking [20:46] good looking out rick_h_ *hat tip* [20:47] lazyPower: yeah, last python release was 0.7, first golang was 1.0 [20:47] viperZ28: I think you can run juju terminate-machine --force to get rid of the dying but not dead yet [20:48] marcoceppi: thanks [20:50] how can I start all over? I want to switch to the dev branch [20:50] viperZ28: run sudo juju destroy-environment local [20:50] thanks! [20:50] viperZ28: then sudo add-apt-repository ppa:juju/devel; sudo apt-get update; sudo apt-get install juju-core [20:50] then bootstrap and roll out [20:51] https://www.irccloud.com/pastebin/SpHP3MwK [20:51] viperZ28: that means your destroy-environment didn't clean up very well [20:51] viperZ28: here are instructions on how to resolve that: [20:53] http://askubuntu.com/a/403619/41 [20:55] thanks again [20:59] hazmat: ping [21:16] where can I find the logs? [21:17] viperZ28: for local provider, they're in ~/.juju/local/log [21:20] I tried to install juju-gui, which worked before [21:20] 2014-02-14 21:18:33 WARNING juju.worker.instanceupdater updater.go:231 cannot get instance info for instance "": no instances found [21:20] https://www.irccloud.com/pastebin/blkSC54r [21:21] https://www.irccloud.com/pastebin/nSv5sJxD [21:25] seems container is out of sync [21:31] marcoceppi: need to add 'sudo rm -rf /var/lib/lxc' to the cleanup process (the link you gave me earlier) https://bugs.launchpad.net/juju-core/+bug/1227145 [21:31] <_mup_> Bug #1227145: Juju isn't cleaning up destroyed LXC containers

[21:31] viperZ28: it's just for 1.16, it's been fixed in 1.17 [21:32] juju version [21:32] 1.17.2-precise-amd64 [21:32] those were probably still there from my initial 1.16 install [21:35] viperZ28: likely [21:36] destroying "/var/lib/lxc" is not a good idea either :-) Directory has to be there [21:39] looks like juju (1.17.2) add-machine lxc has a bug, using br0 instead of lxcbr0 in /etc/lxc/auto/juju-machine-0-lxc-0.conf [21:40] chris38home_: you'll want to open a bug about that asap so it gets fixed by the next stable [21:41] though /etc/lxc/default.conf has the good value (on a precise host) [21:52] can someone help me understand relations? [21:52] viperZ28: sure, what doesn't make sense? [21:53] I pushed a tomcat and a mysql charm, I would like to push an app to the tomcat instance and have it bound to the mysql [21:53] is that possible? [21:53] or do charms have to implicitly define a relation point? [21:54] viperZ28: charms have to explicitly define each relation/interface they connect to [21:55] so if I can find/build a tomcat charm that defines a mysql relation then how would the credentials appear to the app? [21:57] viperZ28: so the mysql interface says that if you /require/ a relation with the mysql interface whatever service provides the mysql interface will send those credentials [21:57] this pretty much answers my question http://manage.jujucharms.com/~robert-ayres/precise/tomcat [21:57] marcoceppi : https://bugs.launchpad.net/juju-core/+bug/1280461 [21:57] viperZ28: so in db-relation-changed you can run `relation-get` to get the various keys [21:57] chris38home_: ta, the QA team will triage and prompt you for any additional details [22:00] so this is sort of the paas/not a paas. By default you don't get wiring for various services but it seems you can build those yourself to handle that. [22:00] viperZ28: well charms define what they can talk to [22:00] it's more like legos [22:01] so there are a few charms that provide the MySQL interface, there's a MySQL charm, a percona charm, etc [22:01] how can I inspect a charm? [22:01] viperZ28: you can view it in the charm browser, or download it with charm tools [22:02] http://manage.jujucharms.com/~robert-ayres/precise/tomcat [22:06] is it possible to view service settings via command line? [22:08] viperZ28: yes, juju get [22:08] marcoceppi: sweet! [22:08] thanks [22:13] it would be nice if it handled NATing [23:42] * chris38home_ seeing that he can do juju add-machine kvm:0, is this curently working ? [23:58] chris38home_: eh, it probably works? Its a new feature [23:58] chris38home_: you can also do juju deploy --to kvm:0 (or lxc:0)