[00:02] ah intersting [00:03] ok, thanks for the input [00:03] (you could just have openstack without juju, if that's more acceptable to your operations folks.. :) [00:03] then you'd only have one juju in play, and that might be easier to wrap heads around [00:03] I want the charms! lol [00:03] we are losing our primary sysadmin, so we are exploring options [00:03] and juju is looknig very good === defunctzombie is now known as defunctzombie_zz [00:55] aimatt: Right, what sarnold described is pretty accurate as far as using Juju to drive both the bare metal and then deployments on the bare metal. I'll dig up a few links for you, but there are video demos of sabdfl having a cluster a machines registered in MaaS, then deploying OpenStack on that in under 15 mins, then deploying services on to the OpenStack environment right after that. It's a little inception-like but a very [00:55] interesting use case [01:10] ahasenack: ping [01:38] aimatt: another nice benefit of juju is the reusability of charms [01:38] you don't need to write a memcached charm [01:38] we have one already === axw_ is now known as axw === CyberJacob|Away is now known as CyberJacob === CyberJacob is now known as CyberJacob|Away === CyberJacob|Away is now known as CyberJacob === defunctzombie_zz is now known as defunctzombie === sputnik1_ is now known as sputnik13 [07:20] I just finished a three part tutorial on how to get going with Node.js / MongoDB on Ubuntu and then deploy it to cloud services. Could I get any interested parties to review it for fixes / ideas on how to improve it? https://github.com/TheMindCompany/mongonode-app/blob/master/tutorial.pdf === defunctzombie is now known as defunctzombie_zz [07:33] Hello. I'm trying to use juju. However I'm having a problem. I did juju-deploy juju-gui [07:34] Then juju added another node as machine #1, however this machine has the agent-state 'not-started'. When I login through SSH to this machine and do sudo restart juju-machine-agent. I get a message that it's restarted, but still no change in the status. Here is the output of juju status: http://pastebin.com/Uv7Zg2qx === defunctzombie_zz is now known as defunctzombie [08:51] What would be the best way to stress test a web application or better yet possibly have Juju auto scale an application. === defunctzombie is now known as defunctzombie_zz === freeflyi1g is now known as freeflying [10:57] Hmm. I have a problem trying to terminate-machine [10:57] I get: $ juju terminate-machine 40 [10:57] error: no machines were destroyed: container info for machine 40 not found [10:57] (I've recently upgraded to 1.11.5-raring-amd64) [10:58] the machine exists and has no services running on it (previously ran destroy-service which worked just fine). [10:58] is this likely to a recent bug? [12:05] drj11: it may be, I've not encountered that error before. [12:22] hi [12:22] can someone merge fix for this bug? https://bugs.launchpad.net/charms/+source/memcached/+bug/1206491 [12:22] <_mup_> Bug #1206491: Memcached service doesn't start [12:30] marcoceppi, thanks [12:30] pavelpachkovskij: np [13:40] and one more https://code.launchpad.net/~charmers/charms/precise/redis-master/trunk/+merge/177828 [13:42] pavelpachkovskij: it says there are 0 lines to merge [13:42] crap [13:42] ah, the merge is backwards [13:43] marcoceppi, I did smth wrong? [13:43] pavelpachkovskij: you're requestion to merge lp:charms/redis-master in to your branch. Resubmit with your branch merging in to lp:charms/redis-master [13:43] marcoceppi, ah [13:43] marcoceppi, one sec [13:44] pavelpachkovskij: FWIW, I just looked at your commit and it LGTM [13:44] surpised that wasn't caught sooner [13:44] https://code.launchpad.net/~pavel-pachkovskij/charms/precise/redis-master/fixed-open-port-issue/+merge/177847 [13:46] marcoceppi, thanks again ) [13:46] np [15:45] I just finished a three part tutorial on how to get going with Node.js / MongoDB on Ubuntu and then deploy it to cloud services. Could I get any interested parties to review it for fixes / ideas on how to improve it? https://github.com/TheMindCompany/mongonode-app/blob/master/tutorial.pdf [15:45] last night not just now :) [15:50] I think I should try and put more detail into using Juju. Around page 10(beginning of juju lesson) I got tired and lazy. [15:51] weblife: I'll take a look in a bit. thanks for the link! [15:54] juju charmers meeting in ~6 minutes! [15:55] weblife: have you seen the node-app charm? [15:55] it automates pages 1-3 of your tutorial, heh [15:57] marcoceppi Thank you [15:57] https://plus.google.com/hangouts/_/b853ceaa631f9a4d4359556e3cbf27ffae3abc81?authuser=0&hl=en [15:57] the hangout will be there for anyone who wants to participate in the meeting [15:58] jcastro I used the node-app charm. But my revision that installs from source and installs express also [15:58] adding an option to the charm to install from source and do express would be a nice contribution! [15:59] jcastro I know. Trying to gain experience so you guys allow me into charmers so I can submit [16:00] oh, you don't need to be in charmers to submit [16:00] you just can start submitting [16:00] charmers is for when you want to review etc, later on when you're experienced with juju [16:00] pavelpachkovskij from CSU Fullerton?? [16:02] hello [16:02] is there a link to charmers meeting already? [16:02] Pad: http://pad.ubuntu.com/7mf2jvKXNa [16:02] We'll be streaming the meeting on http://ubuntuonair.com if anyone wants to follow along [16:02] https://plus.google.com/hangouts/_/b853ceaa631f9a4d4359556e3cbf27ffae3abc81?authuser=0&hl=en [16:02] if you want to participate [16:02] weblife, no [16:03] good :x [16:03] hehe [16:06] yes Brandon Clark is me [16:07] o/ [16:26] experimental kernel messing up my connection :x [16:42] Hmm. I don't think its my kernel anymore, it could be g-hangouts :x [16:42] TRy one more time to listen in [16:44] Or is it over?? haha [16:44] weblife: they're published to youtube so you can always watch post-recording [16:45] I know but want to be able and chime in if I wanted to. Let you guys know other care your doing it also. [16:46] weblife: if you have feedback or questions you can reply to the mailing list email that Jorge sends out every week with the highlights. We're always happy to answer questions and field comments from users! [16:48] marcoceppi Thank you. Your always a help. [16:53] marcoceppi: hey for local provider [16:53] it seems default-series: precise doesn't work for me [16:53] jcastro: yaeh? [16:53] ideas? [16:53] jcastro: define "doesn't work" for you? [16:53] it launches saucy containers [16:53] machine 0 will always reflect the host machine, each additional machien should be the default-series defined [16:54] ah nuts [16:54] so like if I bootstrap [16:54] but do juju deploy --to 0 wordpress [16:54] error: cannot assign unit "wordpress/0" to machine 0: series does not match [16:54] jcastro: you're going to have a bad time [16:54] nod [16:54] ok so test on a real cloud is what you're saying. :) [16:54] * marcoceppi nods [16:55] I hear HP is nice this time of year [16:55] juju switch hpcloud [16:55] :) [16:55] I love this command [16:55] I can just switch from local to cloud in one go [16:55] I still haven't used juju switch [16:55] !! [16:56] you're not doing -eblahblah every time are you? [16:56] jcastro: I am [16:56] old habbits die hard [16:56] I'm terrified of not using switch and accidentally destroying ubuntu-discourse instead of local or something silly [16:56] true dat [16:57] Maybe I'll add a plugin or something that is a safe guard [16:57] like juju safe destroy-environment [16:57] which will just confirm every action you do by telling you current environment, etc [16:57] in pyju there was an env setting you ould set to make it so you could never accidentally blow it away [16:58] IMO juju should always echo the current environment on every command [16:58] "blah blah success/failed on $env" [16:58] Well, juju-core gives you just enough rope. Especially since it doesn't offer an ack mechanism [16:58] jcastro: that doesn't help prevent the issue though, just tells you even sooner that you goof'd [16:58] $PS1 mangling? :) [16:59] I think -v offers that too jcastro "2013-07-31 16:58:45 INFO juju.environs.local environprovider.go:32 opening environment "local-core" [16:59] " [16:59] is the first line typically when using -v [16:59] anywho, -v could use some work on not being so verbose [17:06] marcoceppi: 'juju lock-environment' + 'juju unlock-environment' ? [17:06] marcoceppi: ie. a flag that you can set to help protect you from blowing off your foot? [17:07] http://ec2-54-211-7-242.compute-1.amazonaws.com/?p=4 [17:07] haha man [17:07] that was --to easy [17:08] marcoceppi: wanna siege up that badboy, see how it performs? [17:19] Hmm strange. Juju status says no instances after I bootstrap AWS. Haven't had this happen before. === andreas__ is now known as ahasenack [17:20] weblife destroy's it to try again. [17:22] weblife this happens time to time [17:23] I think possible reason may be when you try to bootstrap right after destroy-environment [17:23] but this is just my theory [17:23] can't delete security groups from ec2 console either, grrr. I'm going to eat them. [17:23] pavel: that is what I did. Makes sense... [17:24] now I know [17:25] weblife: it's also "the cloud" in general. Sometimes weird stuff just happens === defunctzombie_zz is now known as defunctzombie === defunctzombie is now known as defunctzombie_zz [17:30] rick_h: this search result bug where it goes back to the result instead of the readme is killing me [17:31] jcastro: it's landed, there was talk of doing a deploy on wed [17:31] gary_poster: is that deploy on deck for today? ^^ [17:32] rick_h, we can't deploy till benji gets the cache bug landed, and we make a release. At this point I'm guessing tomorrow, rick_h and jcastro. [17:32] gary_poster: k, thanks for the heads up. [17:33] ok guys, so deploy --to is basically awesome [17:33] jcastro: yea, checking our your weekly call video now. I need to test that out and lxc out [17:34] marcoceppi & pavel: I believe in this case it may have been the AWS region. Used west-1 after a little wait the second time and still happened [17:35] worked on east-1 though [17:35] gary_poster: Is "--to" support on the roadmap for gui? [17:36] marcoceppi, yes. we think it will be actually a place where the GUI can give a significantly better experience than the CLI, including automated searching for available machines based on constraints. We will be talking about tech plans next week at the manager sprint and work is slated to begin september [17:37] engineering work, that is [17:37] gary_poster: awesome news, thanks :D [17:37] :-) [17:38] Question!! Why would I get: error: no settings found for "mongonode-app". Here is my config.yaml: http://github.com/TheMindCompany/node-app/blob/master/config.yaml [17:38] It works if I deploy local with --repository [17:39] weblife: what's the deploy command you're using? [17:40] marcoceppi: juju deploy --config config.yaml node-app mongonode-app [17:40] I am operating from ~/ [17:40] weblife: So, when you deploy with --config there needs to be a filed called config.yaml (or whatever is named in the flag) and it needs to be a key value set of options you want to deploy with preceeded by the service name [17:41] Let me give you an example [17:42] weblife: [17:43] http://paste.ubuntu.com/5933505/ [17:43] You don't need to specify a --config flag during deployment, it's only if you want to preseed config values [17:44] weblife: You could just as easily do juju set mongonode-app app-port=80 install_root=/home/ubuntu app_node_env=development [17:44] after deployment [17:46] http://ec2-50-19-27-38.compute-1.amazonaws.com/ [17:46] less than 10 minutes guys [17:46] look at that! [17:47] jcastro social sign returns json error file [17:48] or fb one did [17:48] google one works [17:48] I think for FB I need an API key [17:48] jcastro: these guys claim they can have an instance up and running in 55 seconds.. https://www.digitalocean.com/ [17:48] jcastro I am sure thats it if you haven't registered it [17:48] jcastro: interesting if true ;) I wonder how much effort it'd be to hook up their API with juju.. [17:48] "instance up and running" doesn't count. "service up and running" is what counts. [17:49] sarnold: when we get an SSH provider we'll get digital ocean [17:49] jcastro: sure, but getting an OS going in 55 seconds is a darn sight quicker than my AWS testing so far :) [17:49] true dat [17:49] that's like having the indy 500 be about who can start their car up faster. That doesn't win the race. [17:49] but yeah, 55 seconds is nice! [17:52] jcastro: heh, races can be won or lost on pitcrew performance.. [17:55] arosales: do you have the recordmydesktop snippet handy? [17:55] I am going to do a video on this [17:55] too cool not to [17:56] jcastro, recordmydesktop --freq=22050 --buffer-size 96000 --width 1024 --height 768 --fps 24 -x 6 -y 96 --delay 3 -o ".ogv" === defunctzombie_zz is now known as defunctzombie [17:58] jcastro, the buffer size may not need to be that large but I errored on it being larger just to be safe [17:58] nod [18:00] marcoceppi: Thank you this has had me confused. Now the getting started guide makes sense to me on this. [18:10] marcoceppi: Have you experienced problems when connecting to your node-app deployments on the first connection after startup? It seems like I always get a 504 first try, then I refresh and everything is okay. [18:13] Maybe if I ping the localhost site in the startup script after it loads this won't happen. [18:26] my connection hates me today :x [19:28] jamespage: around? [19:33] weblife: Not that I can recall [19:36] sidnei, he's on holiday [19:37] adam_g: oh. so let me throw this out there and see if anybody cares: https://bugs.launchpad.net/juju-core/+bug/1203816 is about setting up some proxying for the local provider [19:37] <_mup_> Bug #1203816: local provider should support use of local proxy [19:38] i hacked together something that reads MIRROR from /etc/default/lxc, but i wonder if it would be best to run apt-config dump Acquire::http::Proxy instead [19:38] and accordingly, set 'apt_proxy' in cloud-init instead of 'apt_mirror' [19:50] sidnei, seems sensible. i still hope for someday being able to express some subset of cloud-config via environments.yaml [19:50] adam_g: that's an alternative too, and i think it might not be too hard to implement. for all environments i guess? [19:52] arosales: hey so what do you think about a dedicated review day next week? [19:52] we need to get out of this hole [19:56] sidnei, yea, i suppose. [20:01] jcastro, marcoceppi and I are talking about it atm [20:05] heya sidnei [20:06] have you considered becoming a ~charmer and do an occasional review? [20:19] jcastro, all squired away on the ODS submissions? [20:20] we're good [20:20] helped adam power through one of his [20:20] jcastro, thx [20:20] standing by if anyone needs help [20:22] gonna test one more thing before I submit my patch to node-app [20:54] jcastro: i have *wink*. i also need to polish my 3mo old apache/squid/haproxy branches to the latest charm-helpers. [20:55] Does anyone have an example of using nrpe-external-master so that verify I have written a hook properly? [20:57] speaking of which [20:57] sinzui: http://bazaar.launchpad.net/~sidnei/charms/precise/apache2/trunk/view/head:/hooks/hooks.py#L257 it's using the old charm-helpers but i think it only needs the imports to be adjusted [20:57] sidnei, I am past that [20:58] should be good then [20:58] sidnei, I have nagios, juju-gui, and nrep deployed. I joined nagios to juju-gui, and I see it monitored. [20:59] sidnei, I joined nrpe-external-master to juju-gui, I saw the hook fire. I do not see it in nagios [21:00] sidnei, I joined nrpe-external-master to nagios, I do not see the the extra checks in nagios [21:00] sinzui: nrpe-external-master doesn't talk to nagios, it needs be manually rsync off to an existing nagios or some such [21:00] I have set the nagios_master property to the ip of nagios, but I still don't see any extra checks [21:01] wedgwood: ^ you know a bit more about nrpe-external-master and how it's rsynced [21:02] looking [21:03] sinzui: iirc setting the nagios_master property sets up an rsync entry point so that you can rsync from the nagios host, but that needs triggering manually [21:03] sinzui: it's meant only for webops since they manage their nagios outside of juju [21:03] sinzui: where's your hook code? [21:04] wedgwood, https://code.launchpad.net/~sinzui/charms/precise/juju-gui/nagios/+merge/177588 [21:04] ^ I fixed a typo in my local copy, but it is otherwise the same [21:11] Trying to submit a merge into https://code.launchpad.net/~web-brandon/+junk/node-app but it tells me : This branch is not mergeable into lp:~web-brandon/+junk/node-app. Anyone understand why? [21:11] weblife, it is +junk [21:11] personal branches are yours. [21:12] okay I think I get it. [21:12] weblife, merged are supported by shared contexts like a project or a package [21:14] weblife, when a personal branch (+junk is a bad term) is ready to be shared with others for collaboration, you need to push it to a new or existing project. [21:16] weblife, I think you want to push your branch to web-brandon/charms/precise/node-app/install-fix [21:19] sinzui: I don't see where scripts/update-nrpe.py is called [21:21] wedgwood, that script is also symlinked as hooks/nrpe-external-master-relation-changed [21:21] ah [21:26] sinzui: ok, sorry, read scrollback more closely. [21:26] nrpe-external-master doesn't do anything on its own. I think your nagios server will need to initiate the rsync from /var/lib/nagios/export [21:27] that dir contains the nagios configs to be consumed by the server [21:27] wedgwood, np. I am the one truly confused. I don't no how to verify my updated charm truly works as it needs to on prodstack [21:27] bugger /no/know [21:28] This day is too long [21:28] sinzui: if you also implement the nrpe interface then you can use the plain nrpe charm, but it takes a bit more setup. [21:28] hi sinzui [21:28] I understand. I think it's ... I won't say sufficient... indicative to check whether that directory is populated after the hook fires. [21:31] sinzui: that pushed and I was able to merge. Thank you very much [21:31] weblife, great [21:31] hi thumper [21:31] https://code.launchpad.net/~charmers/charms/precise/node-app/trunk/+merge/177943 [21:49] weblife: Thanks for the submission, but you've done it backwards :) You want to merge your branch in to lp:charms/node-app - not the other way around [21:54] lol, what did I do? Let me take a look. [21:55] jcastro, marcoceppi had a good point that folks should probably email the juju list when requesting membership to ~charmers [21:55] weblife: you had them reversed, if you look at the merge proposal it says no diff output [21:55] jcastro, and then charmers and +1/give feedback on the applicant [21:56] I will email the list so folks can comment there since it is somewhat of a policy change. [21:57] haha, whoops. weblife is an obvious newbie to launchpad now... [21:59] weblife: it's okay, I botched probably the first 10-15 merges on launchpad when I really started using it. I'm very accustomed to the Github pull req interface [22:04] jcastro, is there a link to the charm reviewer calendar? [22:04] arosales: see /topic [22:04] jcastro, I don't see a link? [22:04] Review Calendar: http://goo.gl/uK9HD [22:05] jcastro, sorry I didn't scroll :-/ [22:05] thanks [22:08] marcoceppi That seems backwards how they make you do that. Okay gave it another go. [22:09] weblife: much better. I'm not going to be able to get to this today, and the review queue is a bit long, but I'm officially on deck for review next week. So if it isn't reviewed this week, it should be covered by end of next week! [22:14] marcoceppi: No problem, I'm happy with the experience at this point. Needed to learn how launchpad works because I have been wanting to involve myself in projects like this. [22:15] Helped make node-app better and helped update important documentation. Good experience. [22:16] weblife: LP can be daunting at first, but it's actually pretty intuiative once you start using it [22:16] Now its time to focus on my tutorial and some how-to videos. Maybe someone will hire me soon. :) [22:17] marcoceppi: yes it was [22:17] I appreciate all your help over the last few [22:18] weblife: no problem! Happy to help. Thank you for your contributions! [22:21] Issues with JuJu Bootstrapping | http://askubuntu.com/q/327177 [22:24] ahasenack: Are these two problems similar to what you're experiencing right now? http://askubuntu.com/q/318970/41 http://askubuntu.com/q/327177/41 === CyberJacob is now known as CyberJacob|Away === CyberJacob|Away is now known as CyberJacob [22:43] marcoceppi: let me check [22:44] marcoceppi: it is, I don't know to which container/path to upload the json files produced by image-metadata === CyberJacob is now known as CyberJacob|Away