[00:14] hi guys [00:14] Allo Valduare [00:14] looking into this maas and juju stuff tonight :) [00:14] wishing I had spare hardware to test on [00:15] I've you've got a single beefy machine you can try out MAAS in avirtual machine [00:15] that I dont [00:15] I run amde350 proc board [00:16] does juju and these charms play nicely with arm too? [00:16] I dont have an answer for that... [00:17] But I can get back to you tomorrow [00:18] looking forward to the day when I could use things like mk902 for small server farm heh [00:19] Ideally thats just compilation time magic and some PPA's [00:19] build an arm compiler farm, and start porting packages. [00:19] there's already some ARM support in the ubuntu archives [00:20] beyond my skillset [00:22] Valduare: yes, charms work on arm systems [00:23] davecheney: I was just talking to one of the guys in #maas about arm devices for maas as well [00:23] looks like things are getting pretty close to a guy like me being able to set something up heh [00:23] Valduare: there aren't a lot in the realm that mortals ca nafford [00:23] but there are server options like hp's moonshot [00:24] and the calexeda machines [rip] [00:24] davecheney: I use a gigabyte ga-e350n board with 16 gigs of ram hosting 14 vm’s right now [00:24] maas requires that the machines can netboot and have an LOM [00:24] a LOM [00:24] LOM? [00:24] but you can always use the manual provider to enroll machines that you've already setup [00:25] LOM == iLO == IPMI == Lights out management [00:25] ah [00:25] IBM call them service processors [00:25] dell call's them DRACs [00:25] with these arm units using half watt idle can maas be configured to not try and shut them down? [00:25] Valduare: skip maas [00:25] use the manual privcer [00:26] provider [00:36] davecheney: whats the benefit to skipping maas? Less to worry about being ported or.... [00:36] my knowledge of arm in the realm of ubuntu is sadly limited. I'll add it ot my required reading list. [00:39] I suspect its because maas is mostly useful if you want to turn machines off and on :) [00:39] if they're just going to be on all the time, why bother? [00:40] isnt it also for if you want a baremetal server for something ? [00:40] and then roll it back “into the cloud" [00:41] Valduare: /s/cloud/butt/ [00:41] maas lets you treat raw tin like virtual machines [00:42] so you can 'checkout', or 'reserve' a raw hardware which is running *exactly* the same cloud image that you find on any of our CPC partners [00:42] (certified public cloud) [00:42] I get the purpose of MAAS as a whole, i was just curious about skipping it as the provisioner. [00:42] I'm actually pining over a VMAAS setup for some reason [00:42] (mostly because I want to learn more about maas) [00:43] lazyPower: we skip the privisioner because there is no maas provider for little arm boards [00:43] rpi, bbb, pandaboard etc don't have a management nic [00:43] so we can't remote power them or remote netboot them [00:43] at the moment [00:43] most of the parts exist if you're prepared to work at it [00:43] for instance the arm builders that ubuntu uses are remotely powered [00:44] ahhh ok - i've been in talks with marco about it, and the dual nic was a blocker for my suggested rpi setup [00:44] and you can use uboot (grub for arm device) [00:44] makes sense [00:44] to netboot [00:44] but as a function of time == money [00:44] this would never be more than a side project [00:44] honestly [00:45] it's easier to just use the ssh provider [00:45] and skip maas [00:45] I'm considering picking up a few BBB boards and building a micro cluster out of them for some in home apps, and power down my beefy quad core to save on power. [00:45] it doesn't take much to run python code [00:45] lazyPower: look into mk902 [00:46] lazyPower: meh, just rent a digital ocean server for 5 bucks a month [00:46] much more for your money than BBB [00:46] davecheney: these are apps that are specific to my house, and not suited for the cloud. I dont want to run my garage door opener on DO :P [00:46] lazyPower: don't you trust the cloud ? [00:46] not with the keys to my house, nope nope [00:46] Valduare: /s/cloud/butt/ [00:46] With my RSS Reader and blog, you bet. [00:48] davecheney: lol [00:50] lazyPower: the BBB is a great first choice [00:51] little bit faster than the Rpi [00:51] and you get a little bit of built in flash [00:51] Yea, the 1g of on board memory is what akes me happy [00:51] it's 512mb [00:51] theres a 1g bbb board i thought [00:51] ORLY ? [00:51] * davecheney googles [00:51] https://www.tigal.com//product.asp?pid=3406 [00:52] yup, 512mb [00:52] gah, nope, says 512 in the read out [00:52] these things look nice compared to the pandaboard es I've got: http://cubox-i.com/table/ [00:52] BBB will run at 1Ghz with enough current, or 550mhz with USB 500ma [00:52] cheaper -and- faster... [00:52] sarnold: oh no [00:52] not the cubox i [00:53] the biuld quality makes dollar store kids toys look solid [00:53] i have one [00:53] oh no! [00:53] sigh [00:53] it's almost like $55 isn't a good price for a computer :) [00:53] http://www.hardkernel.com/main/products/prdt_info.php?g_code=G137510300620 [00:53] consider the udoo board [00:54] they support ubuntu now [00:54] identical hardware (iMX.6) to the cubox [00:54] guys guys… have you seen the mk902 specs? [00:54] A10/A20 I think [00:55] haha Valduare, you sound like a broken record. Are you a share holder with the mk902? [00:55] no [00:55] Valduare: looks nice [00:55] just cant wait to get one ordered :P [00:56] granted most of the community around the mk devices are home theatre box types [00:56] but the hardware runs ubuntu quite nicely [00:56] Valduare: i was thinking of the older mk800 series devices which were HDMI stick type things [00:56] I have a bunch of mk808 sticks [00:57] just returned an mk802 cozyswan clone to amazon, deciding wether to put the money towards an mk902 or ram upgrade for my laptop [00:58] hmm... $110 pc... [01:07] sarnold: sold, i want one of these cuboxi's [01:07] actually, ~ 5 of them to be exact. [01:07] haha [01:08] lazyPower: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=741127 [01:09] their current kernel [01:09] in fact all the iMX6 kernels are OOOOOOOOOOOOOOOOOOOLD [01:09] 3.0.36 [01:09] ow [01:09] yeah [01:09] bourns! [01:10] bummer :| [01:11] you know, i dont think buliding that would be all that crazy, and cloud storage is cheap... [01:11] however, just about every update ever run would cause it to break stuff [01:11] nevermind, i'm wishful thinking again [01:12] * sarnold giggles, "butt storage" [01:12] Vatt is this s/cloud/butt/ thing that seems to have taken off. Did i miss a warthogs post or something? [01:13] lazyPower: welcome to the internet [01:13] lazyPower: https://addons.mozilla.org/en-US/firefox/addon/cloud-to-butt-plus/ [01:13] https://chrome.google.com/webstore/detail/cloud-to-butt-plus/apmlngnhgbnjpajelfkmabhkfapgnoai?hl=en [01:13] JYNX! [01:13] diversity! firefox -and- chrome are covered :) === CyberJacob is now known as CyberJacob|Away [01:14] http://i.imgur.com/XEJveUI.png [02:24] so these charms [02:24] how are they made [02:30] Valduare: charms are a collecton of shell scripts [02:30] in a tarball [02:31] with names like [02:31] hooks/install [02:31] hooks/config-changed [02:32] so all the normal commands i’d do to provision a servr just wraped up in a shell ? [02:32] Valduare: exactly [02:32] so not something you’d do for if your just provisioning 2 or 3 of a certain webapp etc? :P [02:33] Valduare: sure [02:33] but the thing is [02:33] you can use other peoples charms [02:33] in fact [02:33] you shold [02:33] charms encapsulate the best practice of configuring a particular service [02:33] need mysql [02:33] deploy the mysql charm [02:33] need rabbit mq [02:33] deploy the rabbit mq charm [02:33] need an openstack install [02:34] deploy the openstack charms, or all in one as a bundle [02:34] are these charms checked over and scrutinized? [02:34] Valduare: of course [02:35] https://juju.ubuntu.com/docs/authors-charm-policy.html [02:38] interesint [02:38] is there place I can search what charms are available [02:38] looking for a canvas server specifically - instructure learning management system [02:38] Valduare: sure, https://jujucharms.com/ [02:39] i don't think we have had a canvas charm contributed [02:39] darn [02:39] its pretty fancy lms [02:46] Valduare: actually - to build on davecheney's answer, you're not just dropping shell scripts. We actually encourage users to use configuration management tools if you're already familiar with them [02:47] if you look at the rails charm, it uses chef. The puppet master server, puppet master. I'm getting fairly handy with ansible in a few of my newer iterations on charms [02:47] s/, puppet master/puppet/ [02:47] i’ve never done anything more than follow a blog post guide for setting up what Im looking to do heh [02:48] welp, you're 80% of the way there. Encapsulate teh logics into idempotent bash scripts, and segregate the logic into the proper hooks and you've got a charm. === vladk|offline is now known as vladk === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha === thumper is now known as thumper-cooking [04:33] hey! in the case I want to deploy rabbitmq-server, do I need to deploy ceph-client or hacluster? [04:35] jose: not unless you need those services [04:35] davecheney: cool, thanks~ [04:35] !* [04:35] I'm in the middle of an attempt to write a Reddit charm - it's more hard than what I initially thought [04:36] reddit ? [04:37] yep, what powers reddit.com, it's open source :) [04:37] https://github.com/reddit/reddit/ [04:37] cool :) [04:37] jose: right, are you deploying reddit with juju ? [04:37] (I didn't know about it until a while ago) [04:37] davecheney: not yet, I'm in the process of writing the charm still [04:38] but that would be the final purpose, yes [04:38] cool [04:38] you've come to the right channel then [04:38] oh, oh, are you a charmer? [04:40] some think i am, i think they are mistaken [04:40] mind doing a review of a charm? [04:41] jose: sure I can take a quick look [04:41] awesome :) [04:41] if you want a proper review, why not consider submitting it to the store [04:41] https://bugs.launchpad.net/charms/+bug/1199052 and https://code.launchpad.net/~jose/charms/precise/mailman/trunk [04:41] <_mup_> Bug #1199052: New charm: mailman [04:41] well, I already have, but it's still on the queue [04:41] 11 days as of today [05:01] Error in creating juju bootstrap | http://askubuntu.com/q/438851 [05:05] davecheney: did you get to take a peek? [05:10] jose: i had a little look [05:10] it's late in my work day [05:10] oh, I thought you were down under [05:10] * jose has to stop making assumptions [05:10] jose: i am [05:11] well, anyways :) === vladk is now known as vladk|offline [06:51] How to restart OpenStack services after restart/logout? | http://askubuntu.com/q/438871 === CyberJacob|Away is now known as CyberJacob === mrpink is now known as misterpink === vladk|offline is now known as vladk [08:35] marcoceppi: hi [08:36] marcoceppi: i am facing problem in charm relation [08:38] marcoceppi: i am setting private ip and server port in relation set in relation-joined of provider charm but cant get it in requere charm relation joined/changed using relation get [08:38] i am facing problem in charm relation [08:38] i am setting private ip and server port in relation set in relation-joined of provider charm but cant get it in require charm relation joined/changed using relation-get [08:59] themonk: mind if I take a look at the code? [08:59] is it in LP somewhere? [10:13] if there are any charmers around with some time I need a review of: [10:13] https://code.launchpad.net/~james-page/charm-helpers/neutron-packages/+merge/198518 [10:13] https://code.launchpad.net/~openstack-charmers/charm-helpers/active-active/+merge/211285 [10:13] and [10:13] https://code.launchpad.net/~openstack-charmers/charm-helpers/icehouse/+merge/211934 [10:14] aside from active-active, the other two branchs have been in use in OpenStack CI for the last week [10:16] nope - they are all in use in the CI environment === _zchander_ is now known as zchander [10:45] jamespage: reviewing now [10:45] marcoceppi, thanks [10:52] jamespage: icehouse branch gives me conflicts, it apparently does something different than the first neutron-packages branch [10:52] marcoceppi, most likely yes [10:52] I can resolve the diference really quickly if you tell me which to go with [10:52] jamespage: http://paste.ubuntu.com/7150639/ [10:52] I'm guessing MERGE-SOURCE is the way to go [10:53] marcoceppi, that is odd - I thought that icehouse had that branch merged [10:53] marcoceppi, but yes MERGE_SOURCE is good [10:53] fwiw, it's the only conflict I got [10:53] jamespage: ack, will resolve [10:53] that has the change to accomodate the 3.13 kernel [10:53] marcoceppi, apologues - these have stacked up a bit [10:54] jamespage: np, seems like a lot of good work though [10:54] marcoceppi, more to come [10:54] ssl-everywhere man! [10:54] jamespage: ping me if you need more charm-helpers merges this week [10:54] marcoceppi, will do [10:58] marcoceppi, if you fancy another quicky - https://code.launchpad.net/~james-page/charms/precise/rabbitmq-server/source-take-two/+merge/212576 [10:58] that should unblock mbruzek's test failures for his tests branch [10:58] marcoceppi, just read that back [10:58] lol [10:59] marcoceppi, I have a full charm-helpers redux backed up so need to get that and mbruzek's stuff landed [10:59] jamespage: so source is no longer immutable for rabbitmq-server charm? [11:00] marcoceppi, you can change it on the fly! [11:00] \o/ jamespage thanks for that, appreciate it will review after the icehouse branch [11:00] marcoceppi, of course for the majority of time, this will not do much [11:00] unless a new rabbitmq is avaliable. [11:00] ack [11:01] I'll just drop that little note in the readme when I review it [11:02] marcoceppi, ah - good idea [11:02] I can update that now [11:08] marcoceppi, notes added to README [11:08] jamespage: Oh, brilliant, thank you! [11:10] lazyPower: hey, can you put a few charms in the top of your queue for reviewing today? [11:21] ping lazyPower [11:32] I'm a bit confused, what happen if I put phpmyadmin service on the same machine as wordpress? [11:32] some times it works some times it doesn't [11:32] checking for the open port is not a requirement for the charm? [11:38] overm1nd: it's typically /never/ a good idea to use the --to flag without isolation [11:38] charms are created with the idea that they have the entire machine at their control [11:38] ok [11:39] what you mean for isolation? [11:39] case in point, by default WordPress uses nginx, php5-fpm and phpmyadmin uses apache2 and mod-php5 [11:39] overm1nd: using either --to lxc: or --to kvm: [11:40] where it will create a new LXC or KVM instance on the machine, so --to lxc:1 will create a new LXC container on machine 1 [11:40] so it will isolate the service [11:41] overm1nd: yeah, in a container/vm instead of directly placing it on the machine [11:41] understand [11:41] if you're going for density, that's the way to go [11:41] I have some early feedback from yesterday about juju [11:42] I was expecting more "stability" in the sense of coerence between states [11:42] overm1nd: how so? [11:43] after one day of juju deploy I see that some time if you push a service with resolved [11:43] it works [11:43] even if the juju resolved before was not [11:43] overm1nd: so you're having charms in an error state after a "long running deployment"? [11:43] mmm no [11:44] but is I change a little the order of commands [11:44] I see different results [11:44] well, that /shouldn't/ happen. charms are designed to be idempotent [11:45] yes I was expecting it [11:45] however, some charms are a bit old and outdated and may not work like we expect. Which charms are you having issues with? [11:45] I have the feeling that not all charms are destroying everithing on the machine [11:46] I experimented with haproxy worpress phpmyadmin [11:46] juju-gui and mysql [11:46] I opened a bug for wp [11:47] overm1nd: ah, I see that bug, thanks [11:47] overm1nd: destroy service doesn't actually clean up the machine [11:47] again I'm not a guru so take it with care [11:47] but the install hoook isn't being robust enough [11:48] overm1nd: typically, you would terminate-machine to remove the machine, then deploy on a clean machine [11:48] in italiy we say "pinze" :P [11:48] ok noted [11:48] but that's a bug in wp as well as the hook isn't really being idempotent, so I'll address that [11:48] some charms does not have any info about the address to use them [11:49] overm1nd: how so? [11:49] like phpmyadmin, munin [11:49] like the url to access to the interface [11:49] All charms can query their address information with `unit-get private-address`, unit-get public-address` [11:50] seems obviuos but I had to dig in the config on the node machine to understant the url [11:50] I didn't know [11:51] another note: why the juju-gui does not say the number of the machine? [11:51] I think its a really useful info [11:51] overm1nd: becuase you're looking at a service overview, not a machine view [11:51] overm1nd: there's working being done to show a machine view of the deployment [11:52] right [11:53] my next step is to start writing a charm [11:53] I want to have revive-server (ex openx) [11:53] scalable [11:54] awesome! [11:55] but I still to understand juju more [11:55] I was not able to use haproxy [11:55] as a frontend for more than 1 service [11:55] maybe I didn't getr the point [11:56] I'm used to have an I density service server by my own [11:56] juju seems to waste a lot of resources :P [11:57] high* density* [11:57] overm1nd: you can, but you have to configure haproxy a little more to do multiple services [11:58] overm1nd: well, juju was designed with high-scale cloud deployments in mind, so if you're looking to manage a smaller pool of resources you have to bend a few things in juju [11:58] is ok to deploy with juju and than configure the machine itself by hand? [11:58] overm1nd: you typically wouldn't do that [11:58] what do you need to configure outside of juju? [11:59] for example [11:59] how can I create a subdomani on apache server and deploy a simple php app? [12:00] I think juju is for doing way bigger deployment than what I need [12:01] but I woiuld like to use it at least as scalable backend of frontend for other services [12:01] overm1nd: deploying a simple php app would be one of two ways [12:01] * overm1nd listening [12:02] either build a subordinate charm to the apache2 charm, or build a standalone charm that installs apache2 and php and your app [12:02] i don't know uet what is a subordinate charm [12:02] yet* [12:02] * overm1nd reading docs [12:02] overm1nd: subordinate charm is a charm that co-exists with another charm on a deployment, it's a way to extend fucntionality that complements a service [12:03] sounds what I need [12:03] * marcoceppi nods [12:04] last question I promise [12:04] so, you can make a charm that deploys on top of the apache2 charm, which will put your app in, say, /var/www/ creates a new virtual host in /etc/apache2/sites-available, enables that vhost, etc [12:04] overm1nd: you can ask as many questions as you need! [12:04] this is a type of charm I will try first to write [12:05] the question is: if a service is stuck in pending [12:05] what can I do? resolve do nothing [12:05] because is not an error state [12:05] and I can not debug using putty :P [12:05] overm1nd: pending likely means the machine is still starting, juju failed to install on the machine, or something else that's bad [12:05] overm1nd: typically, if a machine is pending more than 10 mins something went wrong [12:06] and juju isn't actually set up on that machine [12:07] I think I will try to debug from a wirtual machine [12:07] but is the juju env portable? [12:08] I mean the machine I use to give commands [12:09] if I want to manage the cluster from another pc can I move juju? [12:14] overm1nd: yes [12:14] because from putty mtux does not work [12:15] or I am not lucky :P [12:23] thank you very much for your suggestions marcoceppi I will try them [12:38] marcoceppi: sure [12:38] zchander: pong [12:39] lazyPower: sd-agent, importio if you would please <3 [12:39] marcoceppi: dd-agent you mean? [12:40] lazyPower: nope, sd-agent [12:40] ah ok, see it === Ursinha is now known as Ursinha-afk [13:16] can someone remind me which apt-repo to add to get juju-devel? [13:17] marcoceppi told me i get 3 months of dumb questions for free, so i'm starting right away [13:17] tvansteenburgh: I was kidding about the three months ;) [13:18] tvansteenburgh: and it's not a dumb question! ppa:juju/devel [13:18] thanks :D [13:19] wesleymason: great work on the server density charm. [13:19] haha, he's callin you out marcoceppi [13:19] lazyPower: cheers, I noticed you open an account, assumed you were testing :) [13:19] i like this guy already [13:19] wesleymason: theres just one thing stopping it from going to the charmstore. I left notes on the bug [13:19] can anyone find the rails scalable example bundle? [13:19] its literally a one liner 30 second fix. [13:20] lazyPower: awesome, cheers, I'll talk a look in a bit [13:20] man, awesome, lazyPower, someone added ceph support to owncloud [13:20] jcastro: it appears to be missing. I'm only getting the rails-single. [13:20] jcastro: that was zchander :) [13:20] zchander, nice! [13:21] our community is bringing the heat with these new MP's in the Q [13:21] wesleymason, I think you have my email right? I can send you a juju shirt for your charm, just jet me along your address and size. [13:21] zchander, you too! my address is jorge@ubuntu.com [13:22] lazyPower, importio is still waiting too === Ursinha-afk is now known as Ursinha [13:22] jcastro: next in the Q [13:22] man, a good week in submissions! [13:22] jcastro: w00t, thanks! [13:23] lazyPower, ping me when you promulgate server density, I'd like to blog it [13:23] jcastro: pending a one liner from wesleymason, then give me a minute thirty and you'll have it. [13:24] so draft away senor [13:24] sinzui: Hi, if you have a chance, let me know what I missed with https://bugs.launchpad.net/bugs/1297201 that it's incomplete. [13:24] <_mup_> Bug #1297201: Can not bootstrap after upgrade to 1.17.6 [13:24] noodles775, apt-cache policy juju-local [13:25] lazyPower: quick question because I couldn't find a very solid answer in the the interface docs, if I change the requires def to "serverdensity" on line 15, it using the juju-info interface is still fine, correct? [13:25] correct [13:25] \o/ [13:25] the relationship name should *never* be juju-info, as thats reserved [13:26] but the interface is fine being juju-info, that just means you haven't defined a specific interface for the relationship and to use the default. [13:26] noodles775, do you get your packages from the devel ppa or ubuntu. Mayne ubuntu has a bad package. I cannot help with that, but I can let someone know [13:26] lazyPower: makes sense, thanks [13:26] wesleymason: there's also a "local-monitors" interface, that charms like nagios and munin are using [13:26] wesleymason: suggestion on improvement for the docs? I know the relationship stuff can get a bit confusing - since I've been staring at them for 3 months, i feel like i've got a decent grasp on them and your feedback is pure gold. [13:26] o/ wesleymason [13:27] yeah you're supposed to be able to swap in serverdensity for any one of those tools right? [13:27] where charms will tell the subordinate what services to actually monitor, not sure if that's relevant for the SD charm or not, but thought you might want to know [13:27] aha, I was looking for something like that, might be more appropriate [13:27] wesleymason: it could definetly be a post-store revision [13:27] sinzui: Ah - I see. I'd updated juju-core, but it didn't (and couldn't know to) update juju-local. I should have done it the other way around. Thanks. [13:28] wesleymason: as is this works, I can dig up the local-monitors docs for you if you'd like [13:28] marcoceppi: yeah, I'll have a think on the best one later [13:28] noodles775, well I don't think you should need to. upgrade should have noticed there was a newer juju-local [13:28] lazyPower: I'll put my thoughts together tonight and mail you something [13:28] wesleymason: cool, I'll open a bug on lp when this is promulgated to track it's progress, etc [13:28] wesleymason: ta! I appreciate the feedback [13:35] lazyPower: pushed some changes [13:36] lazyPower: Today I made some additional changes for the ownCloud charm and the Merge Proposal did work :D [13:36] Thanks for your information [13:36] wesleymason: excellent. One thing for you to be aware of, if you go the route of local-monitors you will want ot *Add* the relationship instead of changing this one. [13:36] that way you maintain compat with existing deployments. [13:36] zchander: awesome! [13:36] lazyPower: ah, noted [13:36] glad its sorted :) [13:37] wesleymason: thanks for the quick turn around. Promulgating now [13:37] Anyone familiar with File_sender? (https://www.assembla.com/spaces/file_sender/wiki) [13:38] btw - I gave an unconf talk at PHP NE Conference 2014 the other week on Juju, wrote it up about an hour before giving it, and gave a live demo of deploying a PHP app as a service on the local provider that all worked like a charm: http://1stvamp.github.io/phpne-2014-juju-talk/ [13:39] nice! [13:40] jcastro: read wesleymason's comment here, this will get you excited ^ [13:45] zchander I appreciate your work on owncloud but why not seafile? [13:45] seems less popular but more mature [13:46] I don't know file_sender [13:46] overm1nd: you could add it :) [13:47] ehehe too much stuff to do on juju for know :P [13:54] overm1nd: It was something I found via Google and it was already a charm. ;) It made my life a bit easier :D [13:55] ok, I wanted only to suggest you better alternatives [13:56] I digged in to the sync world recently [13:59] is owncloud still an old version in the charm? we should fix that [14:00] lazyPower, hey before you promulgate serverdensity check the README, the usage section needs some markdown love for the commands [14:00] jcastro: already promulgated, and it renders correct in uberwriter [14:01] https://jujucharms.com/sidebar/search/~wesmason/precise/serverdensity-24/?text=serverdensity#readme [14:01] huh [14:01] but not in the store? [14:01] weak sauce [14:01] i cant wait until our docs are in markdown [14:01] i'll just run the readme's through that and get a 1:1 comparison [14:02] apparently my tools support more flavors of markdown than you can shake a stick at [14:02] we should specifically do github-markdown [14:02] or whatever the defacto standard is [14:03] GHFM is fairly defacto, for better or worse [14:04] jcastro: the md docs I'm working on are doing GFM, with a few extras [14:06] jcastro: it looks like charmworld has a bit of a back log? check again in about 10 mins [14:06] yeah that was my plan [14:06] rails/example-complex is missing too [14:06] I told rick about it [14:12] jcastro: I have included the 6.0.2 version in my ‘version’ [14:13] nice! [14:16] jose: ok i will === CyberJacob is now known as CyberJacob|Away [14:43] guys is there an up-to-date juju windows client? on the site is 1.16.6 [14:47] overm1nd: 1.16.6 is the lastest stable juju client [14:48] we don't build devel windows clients [14:48] ah ok [14:48] nor do we put devel releases of juju client in homebrew [14:48] so is the enviroment file grammare change a bit since last rel? [14:49] I get no registered provider for manual [14:49] error [14:50] when I try to import my juju conf from another machine [14:54] overm1nd: yeah, the provider name changes from null to manual in 1.16 -> 1.17 [14:54] ok thx [14:59] wesleymason: jcastro: serverdensity is in charm store, fyi https://jujucharms.com/precise/serverdensity/ [15:08] \o/ [15:12] wesleymason: nice work [15:13] bloodearnest: ta [15:15] lazyPower, hey what was our TLDR for the sample rails app? [15:16] jcastro: I don't think we have one atm. I haven't had a chance to circle back on it === hatch__ is now known as hatch [15:47] hazmat: question about deployer using the -o flag. If a charm doesn't have that key that I'm overriding will it fail or will it just press on? [15:47] I'm assuming the latter, but just want to confirm [15:50] rbasak: we just released a new version of juju-quickstart: https://pypi.python.org/pypi/juju-quickstart [16:04] frankban: thanks! We're in feature freeze now though. We need an exception I guess? [16:05] rbasak: I think we already asked for an exception, rick_h_ ^^^ [16:05] rbasak: yes, there's an existing exception, sec I'll get the link [16:06] I see https://bugs.launchpad.net/ubuntu/+source/juju-quickstart/+bug/1282630 [16:06] https://bugs.launchpad.net/ubuntu/+source/juju-quickstart/+bug/1282630 [16:06] <_mup_> Bug #1282630: [FFE] remove sudo support [16:06] <_mup_> Bug #1282630: [FFE] remove sudo support [16:06] yep [16:06] That's to drop sudo support though, not for a new upstream release. [16:06] it ended up juju released an interim release that has our support and we don't need to wait for 1.18 [16:06] I don't see a problem getting an FFe for a new upstream release, but we probably need to modify that existing one or something. [16:07] rbasak: sorry, this release drops the sudo support since 1.17.6 has the fixes we needed that were originally targeted at 1.18 [16:07] And explain what exactly changed, etc. [16:07] rick_h_: please could you document that in the bug? Then we'll need a release team ack. [16:07] rbasak: ok will do [16:07] frankban: can you catagog the fix in there and mention the retarget from 1.18 to 1.17.6 [16:08] rick_h_: sure [16:08] frankban: a summary of everything that changed from the previous release, please. [16:09] rbasak: will do [16:15] hey guys, so mbruzek and I are modifying a charm for a local deployment [16:16] but we are still seeing the old hook get run. [16:16] how do we increment the charm version? juju always wants to deploy the charm without our local changes [16:16] we've tried `juju deploy -u` [16:16] delete charm cache if you want to be sure ? [16:16] do we just bump `revision` up manually? [16:17] the -u is what clears the cache afaict [16:19] jcastro: I always have to destroy my environment and bootstrap again when doing that, if you have a clue it'd be nice to know [16:20] jcastro: what version juju? [16:20] 1.17 or 1.16? [16:20] yeah we've done that, I am suspecting our JUJU_REPOSITORY is pointing to the wrong place [16:20] 1.17 [16:21] oh, bah, I'm using 1.16 [16:21] jcastro: then -u is being depricated and not needed [16:21] jcastro: are you switching from charmstore to local? [16:21] yeah [16:21] jcastro: or has this always been local? [16:21] jcastro: have you used --switch yet? [16:21] no [16:21] it's always been local [16:22] jcastro: ah, okay [16:22] so basically, we are editing the charm on disk [16:22] jcastro: what does juju status show the charm revision at? [16:22] 5 [16:22] it's supposed to be 6 [16:22] jcastro: is revision file set to 6? [16:22] it is now [16:22] it was not before [16:22] jcastro: try doing upgrade-charm now [16:22] if that doesn't work then your JUJU_REPOSITORY might not be set properly, but juju should have complained about that [16:24] found the issue [16:25] we were deploying from the pristine upstream charm, and not from our local repository where we made the changes [16:25] jose, juju upgrade-charm normally is what you do [16:25] jcastro: even if there's no upgrade-charm hook? [16:27] I'll let you know next time, already started over [16:27] jose: who am I meeting with re: open week presentation in :30? [16:28] lazyPower: that's in a month still [16:28] wat [16:28] :P [16:28] 22 April at 18 UTC [16:28] i thought it was today! [16:28] WOW [16:28] clearly I'm on-the-ball [16:28] we like to plan things with time :) [16:29] and i see on the table on the wiki page, i was anticipating this being shoot from the hip [16:29] awesome [16:29] I'll still get to all of you guys who will be presenting to let you know who your contact during the event will be [16:30] I haven't got my university schedule yet, so I can't promise anything [16:30] jose: welp the fact its not today buys me tons of time. Roger dodger === tvansteenburgh is now known as tvansteenburgh-l [16:31] maybe some time to promulgate the mailman charm? :P [16:31] jk [16:31] Your review will be before I EOD :) [16:31] you've waited patiently [16:31] that has merit in my book [16:31] \o/ [16:32] well, I'll continue looking into this reddit charm, install hook fails :P === tvansteenburgh-l is now known as tvanlunchburgh === vladk is now known as vladk|offline [16:48] marcoceppi, you rock - thanks for the reviews and merges today! [16:48] marcoceppi, found the problem [16:48] you won't believe it [16:48] https://github.com/juju/docs/pull/34/files [16:49] It's forgettable enough that I felt the need to write it down [16:49] jcastro: wat [16:49] yeah [16:49] our hooks were all files [16:49] because scp follows links, it doesn't copy them over [16:49] so any charm you copy over with scp that has symlinks for hooks = ouch [16:49] jcastro: interesting [16:49] you need to either rsync or tar-then-scp [16:50] rsync ftw [16:58] cobbler is ruining my life === tvanlunchburgh is now known as tvansteenburgh [17:42] Pagekite charm submitted! [17:51] interesting solution [17:51] bookmarking pagekite for later reference === vladk|offline is now known as vladk === jono is now known as Guest59567 [18:36] should I worry about this [18:37] juju bootstrap WARNING no tools available, attempting to retrieve from https://juju-dist.s3.amazonaws.com/ WARNING picked arbitrary tools &{"1.16.6-precise-amd64" "http://192.168.4.1/MAAS/api/1.0/files/?key=55dd93f8-b44c-11e3-ac66-7054d2ab7b42&op=get_by_key" "60a008598b13effa46ee50fa13bbe9601ad1018bbb5d9eda379f117e14f145ec" %!q(int64=4666968)} [18:51] Fishy__: no, that's just a warning === zchander is now known as zchander_work === zchander_ is now known as zchander === CyberJacob|Away is now known as CyberJacob [19:12] * zchander won't be scripting/hacking/copying/coding tonight. Just relaxing on the couch with a PS3 controller in my hands ;) === Ursinha is now known as Ursinha-afk [19:23] Do I need a certain # of nodes enlisted before I can use juju maas? [19:23] I have 1 now and status fails === Ursinha-afk is now known as Ursinha [19:30] Fishy: you need probably two nodes [19:30] one for bootstrap [19:30] one for a service [19:33] bummer [19:34] can I make local LXC nodes and have them enlist with maas? [19:34] running out of space to put laptops on my desk [19:34] + i dont want to waste hardware nodes for bootstrap [19:34] want them LXCs i would think [19:36] Fishy: not lxc, but qemu/kvm, yes [19:42] virtual box? [19:44] Fishy: no, not really [19:44] virtual box and MAAS don't work together very well [19:44] bah [19:45] virtualbox's main selling point seemed to be that you got something slightly better than vnc for graphics... [19:45] yah I can just point and click admin it in 30 seconds [19:45] you don't really care about graphics in this case, right? might as well pick the more stable / easier / lighter-weight kvm.. [19:45] wwitzel3, you're doing maas with virtualbox aren't you? [19:45] in the past week I have already learned openVZ and LXC [19:45] Fishy: there's a GUI tool for KVM if you want to try that [19:46] trying to avoid a 3rd VM provider ;) [19:46] OpenVZ isn't very good because it relies on an older kernel [19:46] ya we want to get rid of it [19:46] Virtuozzo, their commercial product, is much better [19:46] but some old stuff is on it [19:46] there's virt-manager or somethign like that, which is a GUI to qemu/kvm [19:46] it's like virtualbox, only better ;) [19:47] qemu/kvm at least, and it's compatibile with MAAS [19:47] ok, installing [19:48] qemu and kvm are the same thing? [19:48] well, for now yes [19:48] qemu is a generic machine emulator, kvm is the underlying provisioner [19:49] so, MAAS can talk to QEMU which can talk to KVM, and virt-manager is the GUI to create KVM machines [19:49] KVM virtual machines are isolated like virtualbox machines, so you have that full level of virtualization [19:49] it's just a chain of tools to do stuff [19:51] so i need to make a bridge [19:51] heavy [19:52] nat wont do pxe [19:52] boo [19:52] Fishy: it's not that heavy [19:52] QEMU/KVM think of it as one tool [19:53] it's your virtualizer === vladk is now known as vladk|offline [21:24] maas keeps shutting off my servers (that are enlisted but not yet in use) [21:24] how rude [21:29] Fishy: that's by design :) [21:30] and the idea is to use WOL when they are needed? [21:30] WOL or whatever power management tool, IMPI, AMT, QEMU, etc [21:31] heres to hoping one of them works ;) [21:31] went up from 1 node to 3 nodes, juju status still never returns [21:34] ok after 10 mins says [21:34] Error details: no reachable servers [21:35] which seems to mean the WOL or whatever doesn't work [21:37] Fishy: does the machine in MAAS work with WOL? [21:38] 1 laptop and 2 kvm VMs [21:38] see no reason why it wouldnt [21:38] enabled in laptop bios [21:44] marcoceppi: hey, it's good if I submit patches to the proof errors that are in the review queue? [21:45] jose: totally! [21:45] cool then, working on a couple atm :) === tris- is now known as tris [21:45] will file bugs too [21:46] jose: \o/ [21:53] jose: YES!!! [21:53] that would be amazing [21:54] lazyPower: well, I have some spare time before starting university and I'm a bit tired of checking and checking the reddit charm without results, so I'll do that instead :) [21:54] about the store errors, I can't do much - I can give you a clue that the postfix and mailman ones are because I changed my username [21:58] is there a way to get the remote unit's IP address during a relation-joined event without specifically passing it as a relation property? [21:59] snewpy: I believe `relation-get private-address` gives you that info [22:01] awesome, thx! [22:06] lazyPower, marcoceppi: tells me that I don't have permissions to make an MP? [22:07] jose: how are you trying to propose merge? [22:07] marcoceppi: with an MP [22:07] or should it be done via a bug? (also filed one) [22:07] jose: that clears it up perfectly! [22:08] huh? [22:08] jose: so, what are the targets for the MP [22:08] I'm trying to figure out why it's failing [22:08] lp:charms/assaultcube [22:08] and lp:~jose/charms/precise/assaultcube/1297529-fix is the proposed branch [22:09] huh, jose can we jump on a hangout so I can see your screen? [22:09] sure! [22:09] jose: one sec [22:40] marcoceppi: I see a problem on the charm, does config.yaml allow a name that's 2 characters long? [22:40] or there needs to be a minimum? [22:40] jose: no [22:40] what's the minimum? [22:40] needs to be more than two characters, that's a yaml limitation [22:40] 3 [22:40] ok, thanks [22:42] hey jose, i see you have the mailman charm listed as depends on: lp:~jose/charms/precise/postfix/trunk [22:42] but there's no relationship for this dependency - was it planned but not implemented? [22:42] lazyPower: oh, that's not the case anymore [22:43] correct [22:43] ack. I'll disregard [22:43] thanks :) === CyberJacob is now known as CyberJacob|Away [23:42] hey guys, what should I do with a charm which has no maintainer? set the maintainer to charmers? [23:51] jose: file a bug for needs maintainer against the charm [23:52] ok, in the meanwhile I'll fix the other stuff