[00:00] hmm. and keeps going down [00:01] ugh.. looks like i hit the 2gb limit on 32bit [00:02] which is odd because the db size is nowhere near that [00:05] hm.. no that wasn't it [00:06] its back [00:06] shazzner_, incidentally thats also now available at http://jujucharms.com [00:06] m_3, thanks for the heads up [00:07] i installed some cheesy monitoring software (amon) recently (also stores in mongodb).. i can't think of any other reason why mongo went bust [00:08] hmmm... right. shazzner_ thanks for the catch! [00:11] hazmat: oh awesome! :) [00:11] and np [00:46] <_mup_> juju/relation-id-option r510 committed by jim.baker@canonical.com [00:46] <_mup_> Relation name is optional for relation-ids [00:47] SpamapS: any news on contest? [00:47] :D [00:52] <_mup_> juju/relation-id-option r511 committed by jim.baker@canonical.com [00:52] <_mup_> Docstrings, PEP8, PyFlakes [00:53] <_mup_> juju/relation-id r495 committed by jim.baker@canonical.com [00:53] <_mup_> Merged trunk [00:54] <_mup_> juju/relation-hook-context r502 committed by jim.baker@canonical.com [00:54] <_mup_> Merged upstream [00:54] bkerensa: initial submissions should be in by now.. I was thinking jcastro would make an announcement on cloud.ubuntu.com [00:55] jcastro: ^^ ? [00:55] :D [00:55] <_mup_> juju/relation-ids-command r505 committed by jim.baker@canonical.com [00:55] <_mup_> Merged upstream [00:55] hopefully mine got in :) [00:55] <_mup_> juju/relation-id-option r512 committed by jim.baker@canonical.com [00:55] <_mup_> Merged upstream [00:58] SpamapS: sure [01:04] <_mup_> juju/relation-id-option r513 committed by jim.baker@canonical.com [01:04] <_mup_> A last bit of PEP8 [01:08] <_mup_> juju/relation-id-option r514 committed by jim.baker@canonical.com [01:08] <_mup_> Removed debugging [01:13] <_mup_> Bug #965845 was filed: Adds support for -r option to specify relation id for relation hook commands. < https://launchpad.net/bugs/965845 > [01:16] so that completes the relation id support - when these branches land, this will enable enumerating relation ids and then use the relation id in any hook - not just relation hooks [01:35] jimbaker, awesome [01:41] I'm trying to figure out how to get files using the filestorage [01:42] what does mean? [01:42] it's a Deferred object [01:42] Deferred is from the twisted module [01:43] is there somethings else I need to do with the object to read the contents? [01:44] ninjix: you are trying to use the juju code from an other python module right? [01:46] yes [01:46] I'm using a main file for testing out the new provider functions [01:48] I want to start interacting with a webdav server [01:49] getting my head around twisted module [01:52] ninjix: this is the best introduction -> http://krondo.com/?p=1209 [01:52] thank you [01:56] avoine: this is a very helpful link [01:56] yeah it's really great === almaisan-away is now known as al-maisan === al-maisan is now known as almaisan-away [06:22] <_mup_> Bug #781949 was filed: Must check certificates for validity < https://launchpad.net/bugs/781949 > === almaisan-away is now known as al-maisan === al-maisan is now known as almaisan-away === zyga is now known as zyga-afk === almaisan-away is now known as al-maisan [09:44] jcastro AlanBell et al: I'm preparing some previews of the Ubuntu Mono font for a request. Does anyone have a really nice juju-status screen they can snapshot? [09:52] not I === zyga-afk is now known as zyga [10:31] AlanBell: any idea who here has a big cluster/cloud? [10:31] sladen: amazon [10:33] * sladen gives lifeless a cookie [10:33] :P [10:37] sladen: maybe just take the simplets charm you casn find and fire a bunch up, should be fairly cheep it just booting screenshot shutting down eh? [10:37] cheap [10:37] plus all the other types blah blah [10:37] * imbrandon rubs eyes === daker_ is now known as daker === zyga is now known as zyga-afk === al-maisan is now known as almaisan-away === hazmat is now known as kapilt [15:06] Is there a flag to specify which Juju configuration to use? [15:06] like a -c? [15:09] marcoceppi: configuration or environment? [15:09] I guess environment, is there a switch to specify a different environment file? [15:09] marcoceppi: environment is 'juju status -emyenv' [15:09] you actually just add them into a single environments.yaml file [15:09] and then choose based on the '-e' flag [15:10] right, okay. Was just wondering if you could use multiple environment files [15:10] there's talk of specifying this with JUJU_ENV too, but I don't know the status of the bug/branch [15:10] no... I've seen a bug for that though [15:10] * marcoceppi is breaking the convention [15:10] I used to keep a bunch and just symlink them to ~/.juju/environments.yaml [15:11] but now I usually just cram them into a single environments.yaml file [15:11] I've got like 14 or so atm [15:11] really use 6 or so regularly [15:11] yeah, I'm to 6 here [15:11] review queue, charmtesting, openstack, charm-dev, etc [15:12] marcoceppi: Bug #922398 [15:12] <_mup_> Bug #922398: cli should accept alternate config file path < https://launchpad.net/bugs/922398 > [15:13] m_3: Cool, thanks, until then I'll just keep pluging away [15:14] Maybe I'll take a stab at patching it tonight [15:14] marcoceppi: cool!~ [15:14] it's suddenly become an important use case for me [15:16] marcoceppi: you could have juju-jitsu manage the env files with a symlink too [15:16] just sayin [15:17] better to put in a MP with the real fix, but maybe add it to jujujitsu until it landss === zyga-afk is now known as zyga [15:21] marcoceppi, any luck with the HP cloud? [15:24] jamespage: no, unfortunately not yet. I'm going to contact their forums see if I can convince them to expose the S3 api [15:24] Tried to use AWS s3 bucket but there's no way to differentiate the auth information for HP cloud and AWS [15:28] marcoceppi, hmm [15:28] guess you never know [15:34] marcoceppi: hows your local omg env one look, i'm curious, not tried anyting local yet [15:34] the wp might work unmodified for drupal, well that part [15:36] aws just droped in git deployments of php apps to beanstalk [15:36] nice [15:37] and have a free teir for RDS now too, sweet, gonna try that oput today and see the reliability of it [15:38] man aws is almost to good to be tru sometimes, then i get the bill at the end of the month [15:38] lol [15:39] i wonder if they stuck rds into the s3cmd like they did the simpledb [15:39] * imbrandon goes to look [15:47] * m_3 loving the sbuild charm... awesome way to capture packaging lore/process [15:48] errr, i missed that one, yea sounds like a good idea, i wondered if someone was gonna capture the buildds brinety etal from debian [15:49] i think that might make a intresting deployable cloud [15:51] m_3: I hadn't thought of it that way.. very interesting [15:54] SpamapS: I'm gonna ask for a really good readme :) [15:56] kees built it for the bugsquash party so everybody could have an instant build env [15:56] Right [15:56] I know that's sort of the point of sbuild itself, but... it's _much_ easier this way imo [15:57] m_3: I have to agree. just having a readme and all the tools you need installed on an instance is useful [15:57] SpamapS: out of curosity from our convo the other day i did go looking a bit to find something else like juju, and i did actualy come accross one, they use alot diff termanology and such but once your past that its really juju charms , the main diff it theirs is 100% wrapped in a hosted web gui [15:57] but thats about the ONLY diff i couldl find [15:57] even deploys to same places etc [15:57] BTW, has anybody tried putting 'placement: local' in an ec2 environment? I didn't realize this was still there.. but it allows you to build a single-node juju environment on ec2. [15:58] SpamapS: nope but i am going to now, i could use that [15:58] imbrandon: hosted web gui.. sounds like a PaaS [15:58] * m_3 rethinks some stuff... [15:58] SpamapS: it is kinda a paas of juju [15:58] well not juju [15:58] but rember when you said we couldnt find anything the same [15:58] i could not either [15:59] cept one [15:59] RightScale [15:59] imbrandon: I dunno if I Meant to say that. There are other things the same, but none that let you share so easily and are fully open source. [15:59] lol [15:59] RightScale is pretty expensive yo. ;) [15:59] ahhh kk [15:59] i'm using the forever free versino for less than 5 nodes [15:59] and deployed a few to aws to toy with it [16:00] its kinda clunky tho i mean nice ideas but not exaclent exacutino i would expect from their hype [16:00] anyhow, was just a passing thing [16:00] :) [16:01] figured i'd scope the landscape for stuff to bring "home", i did gleen a few tidbits heere and there from a few [16:02] but most of these places just plain a dont ork or b suck [16:03] * imbrandon is still waiting for HP to call and let him have a beta env, its the only major player i havent touched yet even revisted gae to make sure they dident open something more than storage up [16:59] imbrandon: I might be able to help with that ;) [16:59] m_3: SpamapS do we have a division of charms for review, or are we just going to review all of them? [17:00] marcoceppi: I was just picking based on cool names :) [17:01] I'm going to download a whole bunch and review on the plane [17:07] marcoceppi: cool!... yeah, there's no harm in multiple reviews for a charm [17:07] appflower looks cool [17:07] I'm playing with their demo [17:11] juju status doesn't show relation status per-unit anymore? [17:15] arg, local provider isn't working anymore [17:16] robbiew: that would be awesom :) [17:17] imbrandon: let me see what I can do [17:18] k [17:19] imbrandon: pm'd you [17:22] marcoceppi: whats broken? [17:23] SpamapS: I'm just getting a perma-pending status [17:23] second re-bootstrap, rebooted between bootstraps [17:23] about 15 mins waiting [17:24] I'm not quite sure where to start diagnosing [17:24] 12.04, btw [17:24] Sorry, the charm is still pending, machine is reported as running [17:31] my ProxmoxLaunchMachine.start_machine() is not executing pass where it yields from the proxmox.acquire_system(). I have try hardcoding the acquire_system() to assign the instance_id a value but juju keeps returning the common/launch.py [17:32] marcoceppi: ps auxfw should show lxc-start with an init running and gettys and stuff.. yes or no? [17:32] ninjix: what are you yielding? @inlineCallbacks can be a bit confusing [17:32] if I assign instance_id in the code without getting the value from acquire_system the rest of the start_machine works [17:33] SpamapS: yeah... did some reactor schooling last night [17:36] SpamapS: in the debugger I can see the instance_id value being assigned in my provider's client class [17:36] SpamapS: then the returnValue(instance_id) is called [17:37] ninjix: I think I'd need to see the code to understand :-P [17:38] SpamapS: sure... how would you like to receive it? [17:38] hrm, i think i'm commiting a juju crime here [17:38] ninjix: bzr branch would be the simplest [17:39] ok [17:39] ninjix: lp:~youruser/juju/yourprovider ;) [17:40] ok so in my "normal" rollout i would include a tools dir with my build scripts that includes like compilers.jar ( closure ) and less.js ( rhino version ) anlong with js.jar rhino etc etc etc [17:40] but i'm thinking thats like packing and would be a no no in a charm [17:40] or am i good [17:40] well a no no unless your bootratraping a new compiler on arm or something :) [17:42] is whoever it was done with the subway juju yet, i wanna actually use it :) lol [17:49] lp:~clayton-kramer/juju/proxmox [17:51] SpamapS: I started hacking on the orchestra provider as my learning example [18:01] i started a full web lemp stack one as mine :) even trying to get a little team for it so its not just me #juju-thestack , heh [18:02] nginx + php + mysql ndb + zend php ce [18:02] memcache etc, whole stack [18:09] imbrandon: so, what I'd recommend is that you put tools you want to have in multiple charms into a package in a PPA... [18:09] ] [18:09] k [18:09] hrm [18:10] how about a github repo then i can submodule it and keep up with upstream thats also on git hub [18:10] free [18:10] :) [18:10] would that work, esp if i like did from a taged release , then its pretty much a package [18:10] without the overhead [18:14] imbrandon: github works fine yes. [18:15] rockin, maybe the more it grows and would become useful to others i might pack it [18:15] imbrandon: the overhead of predictability and efficient encapsulation... yeah.. totally. ;) [18:15] imbrandon: you need to earn about daily build PPA's .. there's no overhead other than having to use bzr. :) [18:15] predictability iof the tag same as git-buildpackage [18:16] i more ment the maint overhead, on me, some of these tols move fast [18:16] like less [18:16] releases 4 or 5 times a month [18:16] and thats one of say 10 [18:17] i mean i guess i could automate a nightly to pull down the latest etc i see what ya mean [18:17] 'but then i loose the predictability [18:17] with the repo sub mosdule i can update with just a git merge command and still ahve fiull control on when it does it [18:17] or if [18:18] imbrandon: I don't really understand what you're trying to accomplish.. but.. as long as you get the binaries in a way that happens the same every time you add-unit .. I don't care. :) [18:18] and i can maintain encapsulation with the tags [18:18] SpamapS: I have my branch up on lp now [18:18] yup yup :) [18:19] ninjix: saw that.. I'm a bit busy at the moment.. but will look at it sometime in the next couple of hours. [18:19] * SpamapS goes afk briefly [18:19] SpamapS: thank you [18:19] yea i'm sure its a little strange as its a internal process to me i'm trying to map but yea [18:19] kk [18:20] that an untill about the 3rd iteration of something i noramlly do it the hard way for some dumb arse reason [18:20] gotta fix that [18:20] :) [18:22] <_mup_> juju/force-upgrade r468 committed by kapil.thangavelu@canonical.com [18:22] <_mup_> cleanup pre merge, fix an upgrade error handling test failure [18:27] <_mup_> juju/trunk r500 committed by kapil.thangavelu@canonical.com [18:27] <_mup_> merge force-upgrade [r=bcsaller] [18:27] <_mup_> As an aid to charm developers, upgrades can be forced which merely puts the charm [18:27] <_mup_> into place on the relevant units. No upgrade-hook is executed if the upgrade is [18:27] <_mup_> forced. [18:40] I'm working on a kusabax imageboard charm [18:41] just thought I'd annouce that, I dunno [18:43] I've yet to see an imageboard that's actively developed and included in the repositories [18:43] kusabax seems like the most active and stable [18:44] shazzner: go ahead and file a bug for it on bugs.launchpad.net/charms when you get a chance... it makes jcastro <3 [18:47] m_3: will do :) [18:48] um, how the heck do I make a launchpad repository for a new charm again? [18:50] wait I think I remember now [18:52] shazzner: just push it from a bzr repo... 'bzr push lp:~/charms/oneiric//trunk' [18:55] m_3: I got 'No such source package kusabax.' [18:56] I've done this all before I just can't remember argh [18:57] shazzner: hmmm... what lp url are you trying to push to? [18:58] m_3: bzr push lp:~shazzner/charms/oneiric/kusabax/trunk [18:59] wow [18:59] * m_3 scratching head [18:59] huh [18:59] my steps: [19:00] bzr init (inside kusaba dir) [19:00] bzr add * [19:00] bzr commit -m "blah" [19:00] bzr push lp:~shazzner/charms/oneiric/kusabax/trunk [19:00] well that's strange... it just worked for me [19:00] bzr init [19:00] touch readme [19:00] bzr add readme [19:01] bzr commit -m'initial revision' [19:01] bzr push lp:~mark-mims/charms/oneiric/kusabax/trunk [19:01] * m_3 totally confused [19:02] could there be some existing .bzr cruft in that directory? [19:02] oh haha, would it matter if I'm on precise? [19:03] hmmm... no, I don't think so (I'm on precise too) [19:04] huh [19:04] it must be though since this just worked: bzr push lp:~shazzner/charms/precise/kusabax/trunk [19:04] shazzner: sorry, limit of my lp know-how... [19:05] m_3: np :) it's kind of confusing haha [19:05] thanks for your help though [19:05] shazzner: sure [19:06] marcoceppi: ping (re wordpress/haproxy-patch) [19:26] ninjix: here? [19:27] SpamapS: yes [19:27] <_mup_> juju/relation-id r496 committed by jim.baker@canonical.com [19:27] <_mup_> relation_ident instead of relation_id; lazily compute relation_name from relation_ident [19:27] ninjix: ok, I have your code now. Help me out, where are you stuck? [19:28] ninjix: instance_id = yield proxmox.acquire_system() [19:28] ninjix: there? [19:28] yes [19:29] ninjix: ProxmoxClient.acquire_system() has @inlineCallbacks but no yields [19:29] SpamapS: :) [19:30] SpamapS: thanks [19:30] * SpamapS really doesn't like the way twisted does this. :-P [19:32] I've heard dev chatter about twisted but this is my first time working with it [19:33] SpamapS: would you say their thinking was twisted? [19:33] ba-da-bump [19:35] baboomtish [19:37] * SpamapS tries the fish [19:42] m_3: don't worry about that hadoop thing, I had jamespage check it ouy [19:48] ahh... this makes sense now. you need the yields so the reactor can pick up where it left off [19:48] jcastro: cool thanks! [19:53] ninjix: precisely! [19:56] * ninjix considers dropping another twisted joke [20:00] ninjix: don't, there may be children present ;) [20:01] SpamapS: can just keep twisting this until it's dry [20:20] <_mup_> juju/relation-id r497 committed by jim.baker@canonical.com [20:20] <_mup_> Fixed remaining tests for review-suggested refactoring [20:31] m_3: what's up? [20:39] SpamapS: I don't see any lxc-start processes [20:50] SpamapS: yahooo. just got a bootstrap to startup :) [20:56] SpamapS: getting this error on bootstrap now: Failed to start job juju-marco-local-file-storage; no output detected [21:35] <_mup_> juju/relation-hook-context r503 committed by jim.baker@canonical.com [21:35] <_mup_> Merged upstream [21:49] SpamapS: fixed the failed to start job [21:49] Still says pending on the charm [22:39] EPIC , i was altready a customer and a huge fan but they got me for good now , http://newrelic.com/developers === kapilt is now known as hazmat