[00:29] SpamapS why do you need to use getall? [00:29] SpamapS, its just not going to scale, so if there are use cases... [00:29] we should come up with some alternatives [00:30] dpb_, yeah.. its been in the ppa for a while [00:30] hazmat: to look at every charm [00:30] hazmat: same problem exists in Ubuntu [00:31] SpamapS, define look at every charm? what sort of things do you want to look for? [00:31] hazmat: right now, searching for all charms w/o maintainers [00:31] SpamapS, i'm just trying to think if we can setup some exposed interfaces against the charm farm to do m/r [00:31] hazmat: and generating stats to help assign them :) [00:32] hazmat: also we have to check them all out for charmtester [00:32] SpamapS, charmtester does them in minimal sets [00:32] hazmat: well we don't.. we could lazy-branch them whenever a test wants a different charm [00:32] er.. charmrunner does [00:32] explicit tests are allowed to ask for any charm [00:32] and its expected that it will be from the same place as the testing charm, not the charm store [00:32] better to do that lazy, or just get the store version [00:33] yeah, NACK to the store version, but ACK to just grab it as needed [00:37] hazmat: there are also times where we just want to dig through them and find out who is doing what in charms. Those types of analysis problems are problematic in any project at scale. [00:38] SpamapS, right.. which is why i'm thinking about a map/reduce interface for doing them [00:38] hazmat: ubuntu could benefit if you keep it bzr-pure. [00:39] hmm [00:40] that might be nice, but depends on the interface.. effectively it should just look like run these commands/scripts against this directory [00:40] where directory is charm or deb [00:40] and distribute to a cluster of servers [00:40] m_3 / jcastro / SpamapS / hazmat ( as you all were present in 80% of those few i got work items from ) yall mind helping me find all/any of mine, I seem to be listed as I should be but with nadda items from ANY tracks and i know thats just flat wrong ( and or poke me to a better solution as I only generally rember the ideas of them not the specifics of any really ) [00:40] and not tonight, like over the next days etc [00:41] pwease [00:41] hazmat: charm getall, btw, is only really slow because it connects and disconnects from launchpad's SSH about 5 times per branch [00:42] imbrandon: none of the BP's have been approved just yet [00:42] imbrandon: so you won't see them on status.ubuntu.com [00:42] ohhhh whew, i thought i was screwed [00:42] rockin, ok perfect i'm just premature here [00:42] ok not the first time, i can deal [00:42] SpamapS, sure, but effectively that's already been done by the charm browser.. its keeping up to date copies of the entire charm universe to support hook browsing [00:43] SpamapS, just thinking it would be nice to leverage that and distribute across multiple servers [00:43] hazmat: eeexxxcccellent smithers [00:43] hazmat: sounds like the makings of a 'charm-mirror' tool :) [00:43] ( and yea that was totaly a terrible joke I seriously applogise now for those that read it as intended ) [00:43] SpamapS, perhaps.. not really needed in this context, gluster or ceph would give multi-server access [00:44] SpamapS, getall is a mirror tool ;-) [00:44] * imbrandon gets back to some other crap code due in the morning that is not so sexy but is paying some electric bills in a few months or something along those lines [00:44] yea just needs a mirror alias in jitsu [00:44] :) [00:44] but why? [00:45] hazmat: I know, a crappy craptastic one [00:45] we can create a tool for the analysis [00:45] SpamapS, what sort of analytics are you looking for? [00:45] basically just grep with a file filter? [00:45] hazmat: grep [00:45] yep [00:46] hazmat: no real reason, from me at least except i can see it being only slight help via surgar syntax when using a local offline only maas or something [00:49] Running offline is an absolute requirement for real production work. [00:50] Many places will be using juju to bring up the whole infra.. perhaps before it is connected to the internets [00:50] SpamapS: btw i got to eat my cacke and keep it too, as i kept the new iPad AND just after i dropped of that macbook at the local non-prof that refurbs em for battered woman and a few other seemed very worhy causes and giving my 32 year old kid brother the other MBP in kindness instead of selling it on craigslist ( he's trained better than me tho as he had ubuntu on it before i left his driveway lol ) when i got home my boy john came over and for a [00:52] so now mini with ubuntu as main and osx cloned in a vm of my old install , the new mini 10 with unity but only 1gb ram and atom N270 proc but very much caopable for mobile work and no others here to take myt time maintaining un-needed [00:52] :)2~ [00:53] oh and not spending a extra dime AND have the ipad :) so yea good monday so far [00:55] now if i could get the ipad to read my mind and share nice with ubuntu on the mini custody of the attached blue tooth keyboard when i say tripple tabbed or something , then i think i would have a nerdgasm not having to lift a hand to use both [00:57] ipkvm client softeware is available on the ipad so i bet its doable is i look up the api's and compile a lil app and can even sign it still hell yea, i know what i'm gonna do next weekend [01:01] imbrandon: Hey does the newer mini w/ multi-monitor work in Linux? [01:01] imbrandon: I had heard tell that the thunderbolt stuff didn't work [01:01] yea no it works perfect [01:02] imbrandon: and I need a desktop machine so I can drive two bigger monitors. [01:02] full res and all [01:02] yea i bout the cheapest new new new model , the 599 or what ever it was [01:02] with a instore 50 off [01:02] and then i took the 50 instalnt off and bout5h 8gb of ram [01:03] and it can take 16gb totaly but not upgraded to that yet, and it comes with both the hdmi --> dvi and the display port to dvi dongles both and both work perfect out the box in linux [01:04] and has room for a second hard drive or first ssd that will be my next purchase and its very veyr simple to add in now with the bottom that opens with no tools [01:05] the ONLY down side so far i found is like most apple things i run out of usb quickly and have 2 powered hubs on my desk [01:05] and the only other tiny thing is if you get the "non server" e.g. normal os and only one drive confuigured option [01:06] you have to buy a apple only sata cable to add the second drive later like i wanna do [01:06] that they only sell online, no picking up at the apple store :( [01:06] its cheap compareitavly tho at like 12$ or something, just the hassle [01:08] everything else its by far the quickest mac desktop betweem the imacs and other mini's ive owned and that 3rd party usb to dvi was only thing including magic trackpad gestures that dident work out of the box in 12.4 [01:08] 12.04 [01:09] I may actually go with a system76 box tho [01:09] i run 2 vm's with 2gb each ram and photoshop and chrome with tons of tabs plus mail.app and other minor crap like adium with no gitter etc [01:09] I want to have more muscle than the mini [01:09] well its becomming blastphmey but i do atleaste like the idea of being able to run basre metal osx if wanted [01:10] yeah I couldn't care less about OS X anymore [01:10] and will still likely make sure i have the option to do so for a while to come until the withdrals quit after detox :) [01:11] but mostly to compile to iOS and Mac App Store.app [01:11] and those i'm using html5 + js + adobe phonegap more and more [01:11] so wont even need it for that if i jumped head first there [01:12] would only have photoshop to cling to that i'm betting a month with gimp 2.8 and inkscape could fix up [01:12] not a year ago, but now possibly [01:12] anyway, time for family stuff [01:13] cheers yea, i promosd my self to get this code done tonight too [01:13] so afk on irc for now anyhow [01:13] l8tr [01:14] SpamapS: i'll seend you all the lsproc lspci etc etc etc if you do consider it more later [01:14] fyi [01:15] oh and know the damn thing runs HOT all the time, i bought a desk fan i keep running 24/7 pointed at mine [01:15] probably a dealbreaker for most [01:15] like orignial MBP 15inch hot [01:17] imbrandon: http://i.imgur.com/CS7HQ.jpg <-- sexy [01:19] bk yea its nice hardware [01:20] i do give them that and ubunu is the closest thing to a Apple Hardware / OS X marrage that is the real magic right beside iOS and OS X intergration that really makes the apple an apple [01:21] on a pure hardware level they are identical chipsets and all as any XPS series dell, just cheaper and no windows tax :) [01:23] ( and yes I said the Dell was the expensive one, on avg 100$ when compared 3 months or ago but neither company has had major releases yet since so it should still be right on ) [01:24] and HP and Lenovo were both nearly the same price as the Apples but not fair to compate as they only matched specs and used cheaper chipsets and other hardware that was not as nice or required an addon [01:24] :) [01:25] bkerensa: Y U NO SEND ME A MOUSE YET [01:25] I went to UDS and all I got was a free System76 to play with :D [01:25] oh yeah let me e-mail them and then I will fedex after I snap some photos and do a mediocre review [01:25] :D [01:27] bkerensa: you got a system76?! [01:27] np, i'll even do the proper review too if you want, i'm fairly certain I have as much readership :) [01:28] bkerensa: it was a wired one you mentioned correct ? [01:29] actually IDK .me crack whip on himself , u no get distracted with shineys === fenris is now known as Guest25535 === Guest25535 is now known as ejat [02:10] there's somethign broken with lp's aliases [04:05] bkerensa: I actually don't find System76 hardware very attractive at all.. I wish they would spend a little more time on the case design. :-P [04:06] To me they look like Toshibas from 2004 [04:08] I may treat myself to one of these though: https://www.system76.com/desktops/model/leox3 ... mostly because the water cooling makes it quiet. :) === almaisan-away is now known as al-maisan === al-maisan is now known as almaisan-away [11:21] does anyone know if I can set up a default environment in juju so I dont have to keep appending -e blah to all my commands?# [11:33] alias juju='juju -e' in your .bashrc > [11:33] not sure where the > came from ignore that bit! === almaisan-away is now known as al-maisan [11:59] victorp, put 'default: ' at the top of ~/.juju/environment.yaml [11:59] jamespage, Thanks! [12:01] jamespage, worked like a... well... charm :) [12:02] victorp, I remember discussion about being able to use an environment variable as well but I don't think thats been implemented yet [12:02] victorp, np [12:18] <_mup_> juju/trunk r537 committed by kapil.thangavelu@canonical.com [12:18] <_mup_> [merge] dont-proxy-https, update local provider apt-proxy config to passthrough for https repos. [a=davidpbritton][r=hazmat][f=993034] [13:34] hazmat: buenas mornings! [13:39] jcastro, Buenos días [13:42] hazmat: I can haz queue? [13:43] marcoceppi: wordpress charm question [13:43] when you backport the fixes you made for OMG into the official charm [13:44] how do we handle upgrades? Is it just a juju upgrade wordpress or do we need to do anything special? [13:45] jcastro, its up.. http://jujucharms.com/review-queue [13:45] \o/ [13:45] hazmat: want me to file bugs on the other details or are you still WIP? [13:46] (needs a link somewhere on the homepage, needs "8 days old" column, etc. [13:46] jcastro, pls file bugs [13:46] jcastro, i've moved on to other items at the moment [13:46] this is awesome, enough for us to get working [13:46] thanks dude! [13:47] jcastro: if I install LXC again, will cgroups break my ability to suspend again? [13:47] no clue [13:47] suspend works with my LXC [13:47] :( [13:48] Someone (smoser maybe?) told me at UDS that my issue most likely came from a known bug in cgroups [13:48] though the other day it filled my disk with machine logs because I left it running for like a week [13:48] mhall119, your issue was cgroups, yes. [13:48] do not install cgroups [13:48] install cgroup-lite [13:48] its the dep for lxc afaicr (cgroup-lite) [13:49] ah, ok [13:49] Does anyone know of a way around the S3 dependency in juju? [13:50] jcastro: where is the "Get LXC setup and start deploying to it with juju" tutorial? [13:50] I have a swift install that does not have compatibility enabled and want to use it with openstack. [13:54] mhall119: https://juju.ubuntu.com/docs/getting-started.html#configuring-a-local-environment [13:54] I have a WI to update that [13:55] jcastro: what's with all the java dependencies? [13:57] spidersddd, we'll have a better story for openstack providers for alpha testing by the end of the week [13:58] mhall119, zookeeper [13:58] mhall119, its pretty minimal deps for a java server [13:59] libjline-java, liblog4j1.2-java (>> 1.2.15-8), libnetty-java, libxerces2-java [13:59] although transtitive blows that up to near a dozen i think [14:07] amithkk: hey, i don't think endeavor can handle twobottux anymore, for some reason it's swapping really bad and is almost unresponsive. so... idk whats going on, i'll let you know once i've figured it out. [14:07] do I need to generate random strings for control-bucket and admin-secret when using LXC? [14:08] mhall119: yes [14:08] can they be random, or do they have to match something specific? [14:08] m_3, ping [14:08] m_3, i think your migration script broke all the oneiric charms [14:12] mhall119: mine are random [14:12] I just paste in the example and jumble the numbers [14:13] m_3: negronjl: SpamapS: proposed charm process mail is on the list! [14:13] jcastro: docs aren't very clear on this [14:14] I know [14:14] I have a WI to fix up the entire LXC part [14:14] mhall119: it's in lp:juju/docs if you want to toss in some quick fixes [14:14] (busy with charm process crap right now) [14:14] are there specific lengths to these keys? === jcastro changed the topic of #juju to: Reviewer: everyone || Review Queue: http://jujucharms.com/review-queue || Charms at http://jujucharms.com || Want to write a charm? http://juju.ubuntu.com/Charms || OSX client: http://jujutools.github.com/" === jcastro changed the topic of #juju to: Reviewer: ~charmers || Review Queue: http://jujucharms.com/review-queue || Charms at http://jujucharms.com || Want to write a charm? http://juju.ubuntu.com/Charms || OSX client: http://jujutools.github.com/ [14:16] jcastro: When the charm is updated I'll make sure it doesn't break older charms === al-maisan is now known as almaisan-away [14:41] Well this is a weird BZR error: http://paste.ubuntu.com/1000957/ [14:47] what's your local config? [14:48] looks like a slightly borked locations rule or similar. [14:48] http://paste.ubuntu.com/1000974/ [14:48] @pilot in [14:48] hmm, guess that doesn't work [14:49] marcoceppi: as in, run `bzr config` in that dir [14:50] mgz: http://paste.ubuntu.com/1000977/ [14:52] hm, is server side. [14:52] \o/ [14:55] jcastro: i dont think you it in here, must be ina query and then also say its name there too [14:56] http://jujucharms.com/review-queue [14:56] in case anyone missed it. :) [14:56] and there is a fair amount of manual setup on the ical server ( gmail ) too if you hadent done that , its not automated in daniels process i was going to try and script it later what i had time [14:56] nice, me adds it as the first bookmark on the always showing bar [15:00] btw did anyone find out what red tape that needs to be gone through to get this GPV3 or so its listed as on LP code ? I still cant figure out the politics involved but its like 4 weeks now and I still dont even know whom at canonical would actualy need to bless it for hazmat or whomever still has the physical bits, and actully I'm bussy now anyhow but still gonna say it so maybe it wont be another 4 ? [15:01] it's in progress [15:01] I'd say find something else to do for now. :-/ [15:02] i have,, for the last month and will continue, but soon enough i wont and wont ask again i'll just rip the css and then make a clone fork [15:02] marcoceppi, that's related to the mail i just sent [15:02] marcoceppi, all the official oneiric charms got broke [15:02] er.. branches are broken [15:02] hazmat: ah, I'm at like inbox 800 still. Need to catch up [15:03] jono let the news out [15:05] wow ok maybe not, bad title to blog post /me is going back to bed before foot goes into mouth 3rd time in an hour [15:07] juju bootstrap isn't working for me: http://paste.ubuntu.com/1001009/ [15:08] imbrandon: we have the queue now [15:08] ? [15:08] see topic [15:09] i did , i'm looking them over to review some now, just wonder what you were meanign if more than that [15:09] :) [15:09] jcastro: http://paste.ubuntu.com/1001009/ when you can [15:09] jimbaker`: don't forget to add your show to juju.ubuntu.com/Events pls [15:10] mhall119: never seen that one before [15:10] hazmat: which list did you send that mail about broken charms to? [15:10] did you log out after you installed all the LXC stuff? [15:10] mgz, there's one on both lists [15:11] jcastro: yeah, rebooted [15:11] separate threads though [15:11] jcastro: it was doing this before reboot too [15:11] I was hoping reboot would fix it [15:11] hmm not sure. [15:12] okay, I see the reply on the juju list. [15:12] imbrandon: any ideas on mhall119's problems? [15:13] sorry was detached , ummm [15:13] is that lxc , hrm looks like not rebooted after the package was installed [15:14] or logged out, not sure a full reboot is needed [15:14] jcastro: imbrandon: http://paste.ubuntu.com/1001017/ is my ifconfig output [15:14] imbrandon: I rebooted already [15:14] ok try to just destroy the envirnment and bootstrap it a second time clean [15:15] that clears up a zk error, not sure if it will help here [15:15] do I need to run bootstrap as root? [15:15] i think localy yes as sudo [15:15] but not on anything else [15:15] eg aws or osapi [15:16] heh, yeah, sudo juju bootstrap goes a lot better [15:16] jcastro: ^^ would be nice to have in the docs [15:16] wait what? [15:16] since when? [15:16] yea, i forgot that was an issue in the charm room 205 sometimes, that probably needs to be fixed ( iirc jamespage is on it ) or updated in docs that lxc containers need sudo prefix for juju [15:17] for the bootstrap or everything? [15:17] everything, well most everything [15:17] cuz really bootstrap bgut then perms are wrong for it all [15:17] so you gotta keep using it [15:17] hrm - you should not run any juju commands via sudo [15:17] it was brought up as a bug [15:18] imbrandon: you should never use juju with sudo [15:18] jamespage: only on lxc or bootstrapping dont work [15:18] marcoceppi: yes i know this [15:18] i said the same thing [15:18] There's a merge on the docs branch which has a fix in the documentation so you won't need sudo [15:18] imbrandon, no - not with lxc - should just work [15:18] It's related to the issue with libvirtd group not being added to the current users group [15:19] jamespage: right , we are all saying the same thing just diffrent ways [15:19] great [15:19] you SHOULD NOT NEED TO, but atm some do due to that nic stuff adding to grp you was fixingt [15:19] or agreed to look into [15:20] after this meeting I'll review the incoming doc request [15:20] yea , not sure why a reboot did not fix him then though [15:20] imbrandon, https://bugs.launchpad.net/juju/+bug/997667 I think [15:20] <_mup_> Bug #997667: local provider makes assumptions about libvirtd default network setup < https://launchpad.net/bugs/997667 > [15:20] yup yup thats the one [15:20] marvellous! [15:20] imbrandon: the install sometimes doesn't even add the user to the group, so no matter how many times you reboot it won't work. [15:21] and yea we're all on the same page, just all had it worded in our own ways diffently :) [15:21] marcoceppi: ahh [15:21] You'll need to usermod -aG libvirtd [15:21] jcastro: <3 [15:21] imbrandon, marcoceppi: and https://bugs.launchpad.net/juju/+bug/997213 [15:21] <_mup_> Bug #997213: Unhelpful error message when doing a local bootstrap if user is not in libvirtd group < https://launchpad.net/bugs/997213 > [15:21] :) [15:22] mhall119: ^^ see what marcoceppi said to run, that should remove the sudo req for lxc then you should be "normal" at that point [15:23] * marcoceppi cant' wait for the document merge [15:23] might need to log out.in after ? [15:23] imbrandon: you can just run `newgrp libvirtd` for that terminal session [15:23] http://askubuntu.com/a/65360/41 [15:24] That question has everything we learned from UDS charm school, and the pending docs merge has all that information updated in the Getting Started section for LXC [15:24] ok after this meeting we can merge it [15:25] cool [15:25] you're in ~charmers you should be able to just merge [15:25] jcastro: I proposed the merge :) [15:25] brandon is in ~charmers now too [15:25] imbrandon: there you go, your first review queue item. :) [15:25] get that badboy in there! [15:28] heh trying man, got a phone call i just got off of, a code put to github going, and a 101 temprature [15:28] can i start the week over >? [15:30] marcoceppi, nice doc in askubuntu on lxc, the only thing i saw is that there's an inconsistency in describing using a local charm repo, vs apparently going against cs [15:30] oh that queue has the merge preposals and such too [15:30] nice [15:30] imbrandon: take care of brandon first [15:31] oh i am, i'm litrally in bed today, got a zpack from the doc yesterday [15:31] but i get too stir crazy w/o laptop [15:32] ok how is the new process work, i see you pulled the new-charm tags off, is that not needed anymore ? [15:34] hazmat: hmm, so I added ~charmers to a few bugs [15:34] and i'm thinking the needs-review status means it needs review from the whom submitted it not the ~charmers right ? we;re looking/re-looking at the fix-commited and confirmed ones correct ? [15:34] and removed the tag [15:34] and they got removed from the queue! [15:34] marcoceppi: so I just need to add myself to libvirtd group? [15:35] jcastro, the queue is based on tags and merge proposals [15:35] specifically the 'new-charm' tag for new charms, and merge proposals for changes to existing official charms [15:35] hazmat: anything ~charmers is subscribed to is the new process [15:35] 'new-charm' is going away, that's why we're using charmers [15:36] jcastro, news to me [15:36] ? we mentioned it during the call [15:36] jcastro, ic, i'll reply on list, but i've got some other meetings to get to [15:36] atm [15:37] that's why we're reusing charmers, because that's how ubuntu does it [15:37] gotcha [15:37] with subscribing a group to get something in the queue, not a tag. K, talk to you on the list. [15:39] ok gotta get my head round this error, detached from irc a bit [15:40] 'morning all [15:44] yikes [15:44] 7 MPs for docs! [15:44] maybe we should add juju/docs to the ~charmers review queue [15:44] since we can work on those [15:48] jcastro, noted [15:48] negronjl, g'morning [15:48] hazmat: pong [15:48] hazmat: filing [15:48] m_3, greetings [15:48] hazmat: yo :) [15:48] hazmat: from now on I'll file on LP so we don't mix up anything [15:48] hazmat: been moving... fun fun [15:48] m_3, no doubt [15:49] m_3, so bit of a fire drill. but afaics the migration script that was run killed all the oneiric charm branches [15:49] jcastro, awesome, thanks [15:49] m_3, i wanted to review the logic of that script, but also we need to figure out how to get them back [15:49] hazmat: hmmm... lemme look [15:50] m_3, there's messages on both lists regarding this [15:50] hazmat: should be easy enough to add back those versions under oneiric branches [15:50] m_3, with history? [15:51] hazmat: sure... none of that was really killed [15:52] m_3, great i was worried [15:52] jcastro: I tried deploying mysql and wordpress to my LXC, but agent-state for bothis still "pending" [15:52] it takes a while the first time [15:52] it has to DL like 300mb of stuff [15:52] it's been 20 minutes [15:52] oh... [15:53] the branches were renamed by the script... realize they should've been copied instead. perhaps we shouldn't use the `branch-distro` tool from lp to do this in the future [15:53] mhall119: ps auxfw .. you should see lxc-create or lxc-clone or a chroot doing stuff [15:54] SpamapS: I see a couple lxc-start [15:54] but no -create or -clone [15:55] ah, because they appear to be started now :) [15:55] sweet, it's working! [15:56] juju destroy-environment takes down all the instances, right? [15:56] yup. [16:00] jimbaker`: thanks, I'll update that inconsistency [16:01] ok, so now how do I get a charm I'm writing deployed? Do I need to copy it somewhere? [16:01] mhall119: you'll want to create a local repository [16:02] docs? [16:02] mhall119: let me find them [16:02] juju deploy --repository= ~/wherever local:charm-name [16:02] is a repository just a directory, or something special? [16:02] it needs to be /precise/charmname [16:02] ok [16:04] mhall119: it's pretty straight forward. Create like a charms folder, put the series name inside of it (~/charms/precise) then each charm is it's own directory inside of the series folder (~/charms/precise/foo) to deploy just `juju deploy --repository ~/charms local:foo` [16:05] do I need to bootstrap after I destroy-environment? [16:05] mhall119: yup [16:07] imbrandon: you don't need new-charm anymore [16:07] I can't find local deployment details in the doc anywhere [16:08] jcastro: i put it back until hazmat fixed the bug so it would not get overlooked [16:09] I got you bro [16:09] one sec [16:09] ok, deploying summit test charm, where do I see the output/logs? [16:09] ty ty [16:09] juju status and juju debug-log are the startrs [16:10] does it have to download the 300mb every time I bootstrap and deploy? === jcastro changed the topic of #juju to: "Reviewer: ~charmers || Review Queue: http://goo.gl/pnUaL || Charms at http://jujucharms.com || Want to write a charm? http://juju.ubuntu.com/Charms || OSX client: http://jujutools.github.com/" [16:10] no [16:10] it caches for you [16:10] it's just the first time sucks [16:11] ok, cool [16:11] ok that review queue is Good Enough for now [16:11] that looks like the 15 charms I removed from the real queue. :-/ [16:11] yea [16:11] I told you to not go so fast, it can wait. [16:11] lol [16:11] I know I know Clint, I couldn't help myself [16:12] jcastro: mira, for the review-queue, should we still track new-charm tag as well as the bugs assigned to charmers? It appears that by just subscribing charmers and removing the new-charm tag, we broke the review-queue....Am I missing something ? [16:12] negronjl: bug filed, hazmat is on it [16:12] negronjl: right [16:12] jcastro: give me the bug number so I can fix the CLI version of it as well [16:13] https://bugs.launchpad.net/charmworld/+bug/1002976 [16:13] <_mup_> Bug #1002976: Review queue needs to account for ~charmers < https://launchpad.net/bugs/1002976 > [16:13] jcastro: thx [16:13] hey when you are up to date, can you add a link to your tool in the Reviewers section? https://juju.ubuntu.com/CharmsProposedProcess [16:14] dang, sorry about the queue fellas, it was looking so wicked this morning and then I got ahead of myself [16:14] oh well, at least we found a bug! [16:15] I'm on the CLI fixes now [16:15] jcastro: We'll have it all fixed soon :) [16:15] oh hey [16:15] I have another one [16:15] what's the cli project on lp? [16:15] I'll just affects you [16:15] jcastro: it's charm-tools [16:17] negronjl: this one too: https://bugs.launchpad.net/charm-tools/+bug/1002977 [16:17] <_mup_> Bug #1002977: Review queue should list lp:/juju/docs < https://launchpad.net/bugs/1002977 > [16:17] jcastro: got it. working on it so the CLI gets that [16:18] ok I'll affects charm-tools if I file a bug that needs parity in the CLI tool from now [16:18] jcastro: ok [16:18] dist-upgrading ... I may or may not be back in a minute :) [16:19] :) [16:20] negronjl: when he's all done we need to set up the graphs too [16:24] robbiew: Y U NOT SEND ANY CANONICAL PPL [16:25] http://cloudslam.org/cloud/conference-program [16:25] b/c it's called "Cloud Slam" [16:25] lol [16:25] heh its the oldest one in the country tho intel everyone else is there :) [16:26] if we sent people to every cloud conference, we'd have no one to work on ubuntu server/cloud ;) [16:26] woulda been nice to have juju int the Cloud PaaS DeathMatch deploying some PaaS :) [16:27] true i was thinkin that was a big one tho, really guess not [16:27] i mean it is but diffrent big [16:28] more like the Ubuntu Cloud Summit, infact looks like most of the same attendees [16:28] * negronjl is back [16:51] jamespage: nice work on a promulgation! [16:52] up to 81 official folks! [16:52] jcastro, a little tardy.... [16:52] but a nice charm - thanks nathwill! [16:53] thank you for your help in getting it to good quality, jamespage :) [16:54] nathwill, np - you're welcome! [17:00] the waiting is the hardest part.. [17:01] take it on faith [17:02] jcastro : Bug 997339 … u remove the new-charm mean its being review ? [17:02] <_mup_> Bug #997339: Charm needed : Openbravo < https://launchpad.net/bugs/997339 > [17:04] ejat: see the list, subscribing ~charmers is the new review process [17:04] I just updated the bug [17:05] so it's still in the queue [17:05] nathwill: how many charms do you have in the store now? [17:06] jcastro: did you make sure charm-tools and jujucharms.com also respect that new process before changing it? [17:06] ok noted [17:07] jcastro, just the one for now, got another one pending. (precise/varnish). still contemplating what rocking apps i'm familiar with that aren't already charmed, before i start my next one :) [17:08] nathwill: have a look at all the existing stuff.. there's a need to make them better/morescalable/etc. [17:09] SpamapS: ok, i can do something like that... are we not worried about stepping on people's toes, then? [17:09] nathwill: definitely not! If you have an improvement, just make a merge proposal [17:09] indeed [17:09] SpamapS: okey doke :) [17:09] owncloud could use an update if you're looking for an easy one [17:10] big new upstream release, would be nice to have it shiny in the store. [17:10] jcastro, i love the owncloud, so i'll definitely take a gander tonite. [17:10] <3 [17:11] make sure upgrade-charm goes smoothly :) [17:11] kk. looks like there's some room to add the db-relation-joined hook as well... [17:16] marcoceppi: don't forget to self document your maas adventures! [17:18] I need a review on https://code.launchpad.net/~negronjl/charm-tools/review-queue-charmers-group-juju-docs/+merge/106861 [17:18] marcoceppi, SpamapS, jcastro, : ^^ [17:18] on it! [17:19] nuts from the branch name I thought that was docs, not code [17:19] aujuju: What is the best way to use the mysql charm in Juju with dynamic database credientials? [17:20] jcastro: This is the fix so the CLI review-queue get bugs from charmers and the MPs from lp:juju/docs as well [17:20] negronjl: dang, I don't see how I can unassign myself from the review [17:20] jcastro: someone else can claim it [17:21] s/can/has to/ [17:21] clint will save us [17:21] lol [17:22] jcastro: or... you can actually review this :) [17:25] <--- codes as much as he can speak spanish [17:26] ahhh ... SpamapS will save us then ... or marcoceppi :) [17:30] jcastro, so what do you want to show up in the review queue re bugs? [17:30] hazmat: I modified the CLI review queue ... it should probably look like it ( but of course prettier ). [17:30] hazmat: The updated code for the cli is here: https://code.launchpad.net/~negronjl/charm-tools/review-queue-charmers-group-juju-docs [17:31] negronjl, i'm looking at it.. but that doesn't look right to me [17:31] negronjl, its still relying on tags for bugs [17:31] hazmat: I updated the code in the above link [17:31] negronjl, also use python-dateutil [17:31] negronjl, dateutil.parser.parse(some_date_string) == python_datetime [17:31] hazmat: anything ~charmers is subscribed to [17:32] that is an open bug [17:32] ah.. its getting both [17:32] negronjl, ic it now [17:32] hazmat: :) [17:33] negronjl, itertools.chain(bugs.entries, charmer_bugs.entries) [17:33] one block then.. cause its pretty ;-) [17:33] hazmat: hey .. i never said my code was pretty xD [17:34] hazmat: talking with lp folks about the oneiric charms... think I understand what's going on with it now... long-term fix involves bugs #1003016 #1003017 and #991980 ... short-term fix is still in the works [17:34] <_mup_> Bug #1003016: branch-distro stacks on unique_name of new branches, not allowing renames < https://launchpad.net/bugs/1003016 > [17:34] <_mup_> Bug #1003017: Allow changing the new name of the branches produced by branch-distro < https://launchpad.net/bugs/1003017 > [17:34] <_mup_> Bug #991980: Oneiric official branches are all locked < https://launchpad.net/bugs/991980 > [17:34] hazmat: That's actually a good idea so, I'll change that for the nex iteration [17:34] whoa [17:34] I'm impressed byu _mup_ there [17:35] yeah... surprised that worked [17:35] m_3, okay that's the restore fix.. but afaics we shouldn't be renaming in anycase.. [17:35] for the next release [17:35] m_3, that doesn't lok promising [17:35] re restore fix [17:35] m_3, what's the short term fix discussion? [17:36] hazmat: long-term it's the branch-distro tool we used. it wasn't actually renaming, it was creating new branches based on the old [17:36] m_3, but in the process it altered the old branches? [17:36] hazmat: it just created them with branch names we can't use... so we had to rename to /trunk [17:37] jcastro: see I got my peeps writing charms for you now (nathwill) :P [17:37] \o/ [17:37] bkerensa: are you sure you're not his peep? [17:37] lol [17:37] hazmat: yes, as I understand, it changes what those branches are stacked on top of [17:39] m_3, but it shouldn't it be stacking the copy not the original [17:39] the precise branch copy should be stacked on oneiric [17:39] hazmat: not sure... I'll dig through the actual branch-distro tool itself... don't really understand what it's for... upgrading series, but the branches have very specific names [17:39] not the other way around [17:39] m_3, yeah... i just re-read your email now with proper context [17:40] perhaps we need to just do our charm promotion from scratch and not use branch-distro at all [17:40] but its still unclear why it modified the oneiric branches at all. [17:40] SpamapS: let me know when you guys need more reviewers ;P [17:40] agree [17:40] the stacked on branch is for the new target to pull revs from the old [17:40] but we have the old pulling from the new [17:40] which suggests something is wrong with how its being done [17:41] m_3, are the scripts you used for this available? [17:41] hazmat: lp ran `branch-distro` when clint told them to [17:41] then I ran the script that's in the email for renaming [17:42] hazmat: the renaming broke the stacking [17:42] hazmat: but it's a good question to ask why the old oneiric branches get suddenly stacked on top of the new precise one [17:44] hazmat: I believe the reason for the stacking from precise <- oneiric is because the distro model has the oneiric branch *frozen* [17:44] m_3, so none of these scripts are nesc imo.. we can just script bzr to push the branches to the new location and promugate them [17:44] SpamapS, but that's not the case with something like an lts [17:44] hazmat: its always the case [17:44] SpamapS, frozen is fine.. but not accessible i find very hard to believe [17:44] release pocket does not change [17:44] * m_3 hanging with antonio... one sec [17:44] and it works for crap [17:45] lp:ubuntu/oneiric/anything should be "the newest upload" [17:45] not "the one we froze" [17:45] SpamapS, that's what tags are for ;-) [17:45] instead we have lp:ubuntu/oneiric-updates/anything [17:45] SpamapS, wait your saying you can't access the ubuntu/oneiric/anything?... or is that workflow in combo with the rename that killed everything [17:46] hazmat: I'm going to see if we can just write our own fork of branch-distro that doesn't do this [17:46] hazmat: you can read it, but you can never change it [17:46] SpamapS, where does that script live incidentally? [17:46] hazmat: lp:launchpad [17:46] k [17:46] SpamapS, its not a concern atm, we've got some time, but yeah.. that we could come up with a juju charm specific one that matches our use case better [17:47] hazmat: scripts/branch-distro.py [17:47] because their definitely not frozen for juju [17:47] hazmat: but it calls some lib/... stuff [17:47] hazmat: I just think we should use tags, as you say, and updates should be able to go into the branch [17:47] * hazmat knows better than to branch lp:launchpad [17:47] negronjl: sorry, mid-maas adventure. I can take a look in an hour or so if it hasn't been touched yet [17:47] marcoceppi: thx [17:47] 228M /home/clint/src/launchpad/.bzr [17:48] there's actually 88M of code.. wtf?! [17:48] SpamapS, for the distro, the problem with tags is that their per branch, where as the distro wants to create an illusion of a group tag.. and individual tags still means the branch can be overwritten.. [17:48] SpamapS, lp is ginormous [17:49] Ahh there's a ridiculous amount of test data [17:51] hazmat: on to an unrelated issue. Did you have a branch to enable proposed? [17:51] hazmat: we need that for the SRU [17:51] hazmat, it stacks old-on-new otherwise every six months a link (and hence a round-trip) would be added to the common case [17:53] it's not so much that the old version is frozen [17:54] james_w, ah.. its an optimization [17:54] charms could probably get away without any stacking [17:54] indeed [17:54] ubuntu can't as it is *large* [17:54] their really independent [17:54] so we stack to reduce the disk usage [17:54] which charms remain small and relatively few that shouldn't be an issue [17:55] but if bug 1003016 is fixed then stacking shouldn't really affect you [17:55] <_mup_> Bug #1003016: branch-distro stacks on unique_name of new branches, not allowing renames < https://launchpad.net/bugs/1003016 > [17:55] james_w, well that's true for the most part.. but some charms have binaries or src in them.. and the number will grow larger. [17:56] right now a full checkout of all charms runs near ~1G afaicr [17:56] right, and Ubuntu is something like 200G, so it will be a few years until not stacking charms would get close to the usage of Ubuntu [17:56] james_w, indeed [17:57] but I don't see a problem with keeping the stacking [17:57] it's not the stacking that makes lp:charms/oneiric/* immutable [17:57] as long as its done per your bug with numeric ids, it shouldn't be an issue [17:57] yeah [17:57] james_w, what makes it immutable? [17:58] * hazmat bites the bullet and checkouts lp [17:59] hazmat, there's a check in LP that ties mutability to the status of the seres [17:59] series [17:59] but having said that https://launchpad.net/charms/+series says that both oneiric and precise are "Active Development", so perhaps it only allows one to be mutable [18:00] james_w, i dunno that we've really tested it [18:01] so close [18:01] james_w, yeah.. the ppa oneiric charms are still mutable [18:01] http://jujucharms.com/~yellow/oneiric/buildbot-slave has a recent commit [18:02] hazmat, but not the official? [18:02] james_w, the official ones are all unavailble [18:02] james_w, they can't be checkout [18:02] james_w: ultimately, I think we want to not repeat the mistake of having pockets for updates/security in lp:charms/* .. we just want a branch, which is the latest one that everyone should have. [18:02] because of the stacking issue [18:02] hazmat, right [18:03] but I wonder if they are immutable too [18:03] SpamapS, yeah, as long as you don't have pockets you shouldn't have pocket branches for sure [18:03] james_w: well, IMO ubuntu shouldn't have pocket branches either. [18:03] :-P [18:04] well, IMO it needs them as long as it has pockets [18:04] we could get rid of pockets [18:04] m_3: Are you still reviewing https://bugs.launchpad.net/charms/+bug/898714 ? [18:04] <_mup_> Bug #898714: Charm Needed: The Locker Project < https://launchpad.net/bugs/898714 > [18:04] * negronjl is going through the review_queue now [18:04] The archive needs them, because apt needs them.. but for branches, you really only want one mainline of "what should be shipped to users right now" [18:05] Anyway, we've diverged [18:05] The fix seems to be to not use branch-distro as-is and fork its behavior somewhat for charms.. [18:06] don't make the branches immutable.. don't rename them to $series, and that will eliminate the problems we have [18:08] * m_3 back [18:09] negronjl: I grabbed that the night of the contest, but then we decided to use a different triage process [18:09] m_3: is it up for grabs then ? [18:09] m_3: so I can review it [18:09] negronjl: sure [18:09] m_3: ok. I am reviewing it now. [18:09] negronjl: lemme know if I have to do something to the bug to release it [18:09] m_3: I've almost got a working summit charm [18:10] without puppet [18:10] mhall119: cool! [18:10] m_3: ok [18:10] mhall119: lemme know when it's ready to deploy and we can swap it out for linux plumbers [18:11] mhall119: linux plumbers needs a little love anyways... the environment's pretty stale [18:11] SpamapS, I don't think branch-distro makes them immutable [18:11] m_3: this might take some more testing before it's ready for real-world use [18:12] james_w: its entirely possible that I thought they were immutable, but they were actually just stacked improperly [18:12] but the good news is, 90% of the charm is being auto-generated from the django setup [18:12] which means we'llbe able to do the same for any django site [18:13] SpamapS, sounds possible, but they may well be immutable [18:13] django-openid-auth is a pain though, how were you installing 0.4 ? [18:13] SpamapS, if they are then I don't think it's because of branch-distro, I think it's just a piece of logic in Launchpad somewhere [18:17] mhall119: ah, that's right... I remember you were working on a generator for django charms [18:17] mhall119: I'll have to look... one sec [18:21] james_w: ahh, like "newest active dev series wins" ? [18:21] SpamapS, possibly [18:22] the rule used to be "if you can upload the package to that pocket you can push to the branch" [18:22] I'm not sure how that has changed [18:22] mhall119: I do a pip install from requirements.txt first and then install django packages for django openid and python openid to catch anything missing [18:22] and I'm not sure what the rules are for that when there are two active development releases, as Ubuntu never does that [18:22] mhall119, like the idea of the django charm generator - i wonder if it can be genericized to just use some juju cleverness, but best step is to get this sort of stuff working first [18:23] jimbaker`, what juju cleverness would that be? [18:23] james_w: perhaps more confusing is whether or not ~charmers == ~ubuntu-core-dev [18:23] james_w, as in config settings or some of the relation orchestration support [18:23] jimbaker`: IMO juju needs 'extends: other-charm-name' [18:23] jimbaker`: if juju had single inheritance, I'd be a very happy camper. :) [18:24] jimbaker`: what kind of juju cleverness? [18:24] SpamapS, i think that's a charm splice sort of thing for now [18:24] jimbaker`: charm splice is more like mixins.. I want real inheritance. [18:24] so you can have a django app that inherits from the django abstract charm. [18:24] jimbaker`: I'm already looking at things like settings.py, setup.py, and requirements.txt to generate custom charms [18:24] mhall119, without looking at your charm, rather impossible to say specifically [18:26] SpamapS, how would you use charm inheritance on top of splicing? i believe it makes sense, just want to understand better what we are losing [18:26] jimbaker`: splicing is a hack.. just look at how it does relations between spliced charms.. bad bad bad idea IMO [18:26] jimbaker`: you can bzr branch lp:~mhall119/+junk/django_juju if you're interested [18:27] mhall119, thanks! [18:27] jimbaker`: simply add 'django_juju' to your settings.py, and run 'python manage.py charm' [18:27] jimbaker`: inheritance would just allow us to have abstract charms for things like django and nodejs [18:28] SpamapS, seems very reasonable to me. i do wonder if there's an intermediate step that's less hacky but doesn't require this getting into core [18:29] SpamapS, which i find to be very interesting questions! [18:29] jimbaker` the reason it needs to be in core is I want it to be easy to use. [18:29] jimbaker`: I don't want an intermediate "build" step [18:29] I think a subordinate for a django app would likely work reasonably well too [18:29] jimbaker`: I want to be able to say 'extends: mysql' and have juju look in the current namespace, and then maybe also in the default namespace [18:29] james_w: yeah thats the current way to do this [18:29] Anyone have any experience bootstrapping MAAS? [18:30] but Its clunky IMO [18:30] ok, lunch time [18:30] yikes... yeah, I'd better eat too [18:31] we'll chat about how to immediately fix the oneiric charms... I'm not happy that it might require us to rename (break) the precise ones as a step in the process [18:37] mhall119, looks interesting. my first reaction is, it still looks like it could be done with config settings. but i need to delve more into it. regardless, your efforts would not be wasted, just a matter of changing when things are done [18:39] Getting the following error when trying to bootstrap MAAS: http://paste.ubuntu.com/1001423/ [18:44] marcoceppi: looks like dns stuff... maybe check that you can dig stuff using your maas server for dns? dunno [18:45] m_3: I'm using an IP for the MAAS master [18:45] I just realized I haven't "commissioned" the machines [18:45] trying that first [18:48] marcoceppi: lol [18:48] marcoceppi: now you just need to trip over the ethernet cable ;) [18:48] negronjl: they're all virtual machines :) [18:53] SpamapS, ideally we'd have generic charms for those things imo [18:53] SpamapS, even more ideally we just heroku packs ;-) [18:55] which are all mit licensed [18:59] hazmat: but the point is that I want a charm for *my* service, not some generic django thing [18:59] jimbaker`: maybe, I'm not too familiar with configs [19:00] SpamapS, but then its just a matter of configuration of the generic django thingy [19:00] hazmat: django can generically talk to databases, but only my app knows what context each db connection is in, or how I also talk to rabbitmq [19:00] hazmat: no, unless you have dynamic relationships.. I need explicit relations for each thing I want to relate to [19:00] ok, I think I have it working on a local sqlite db [19:01] SpamapS, there are fairly standard settings for all those things, and yes the generic needs to support all of the possible rels, and inject into std settings [19:03] hazmat: lets say I start using neo4j .. django might not support that. Or two instances of the database, one for highly transactional database stuff, and one as a readonly slave of public data that I need... I doubt there are standard settings that can handle this. Plus when I relate to the databases.. how do I give that context without separately named relations? [19:03] I need to make it convert bzr+ssh branch urls to their cooresponding http ones [19:06] \o/ it works! [19:08] jcastro: you need to do some heavy juju evangelizing up here in July [19:08] ;) [19:08] * bkerensa is getting bogged down by the "Puppet is Win" crowd [19:09] juju works awesome with puppet. :) [19:09] bkerensa: indeed. [19:09] jcastro: so, just when we got a "big list" .. you go and move the way to calculate the big list. :-/ [19:10] mhall@mhall-laptop:~/projects/juju-local/mhall-local$ juju add-relation postgresql summit [19:10] No matching endpoints [19:10] yeah dude, I'm sorry. :( [19:10] :( [19:10] jcastro: fixing now in charm-tools [19:10] SpamapS: hazmat is going to rescue me [19:10] heh, who cares about the web version? ;) [19:10] I thought juan just fixed it [19:10] what am I doing wrong with this relation? [19:10] oh if he did then fantastic :) [19:10] hey mira, what will you backport policy be for charm tools in 12.04? [19:10] I want the text one too [19:10] but I don't want a PPA [19:11] jcastro: charm-tools is a great candidate for precise-backports [19:11] just have to do the work of requesting it [19:11] jcastro: yeah well the OSU LUG folk Pro-Puppet and think its futile to use Juju at all [19:11] m_3: I needed to specify a port for the URL :\ [19:11] :P also they seem to have beef with the juju documentation [19:11] said its hard to find info [19:11] we all have beef with juju docs. [19:12] jcastro: I kind of like that we're all subscribed to the new charm bugs now. Getting email about them is actually going to make people look at them more I think. [19:12] everyone is working on them [19:12] The docs aren't that bad IMO [19:12] I mean, there's a getting started [19:12] it works reasonably well [19:12] SpamapS: indeed, I feel stupid that we didn't do it like this in the first place [19:13] Most people get stuck on the "you need an amazon account" [19:13] I hope to make the docs super this cycle [19:13] but oh well, there was something nice about the wild wild west of the last cycle, heh [19:13] jcastro: http://paste.ubuntu.com/1001510/ [19:13] I need help [19:13] bkerensa: and you dont have to convert everyone :) let them be puppet fans, competition is good [19:14] They don't compete! [19:14] just mention once in a while that Wint^Juju is comming [19:14] SpamapS: I fixed the review-queue subcommand in charm-tools but need, review ( https://code.launchpad.net/~negronjl/charm-tools/review-queue-charmers-group-juju-docs/+merge/106861 ) [19:14] If I said "Screw dpkg, apt is the better tool" you would make fun of me for comparing things the wrong way [19:15] jcastro: I will where my PuppetLabs shirt to your talk :P [19:15] mhall119: never saw that one before. [19:15] wear* [19:15] jcastro: if you insisted at the freaklevel i do when i get excited aka your typical lug, yea i'd try then let them learn the hard way [19:15] mhall119: does summit use postgresql? [19:15] bkerensa: I'll be working on puppet modules that will use SpamapS puppetmaster and puppet client charms ... you will be able to deploy your puppet modules with Juju and _without_ having to change the way you code your puppet modules [19:15] marcoceppi: yeah [19:16] mhall119: I don't know where the summit repo is, can you pastebin each metadata.yaml file? [19:16] ah, my summit metadata.yaml had: [19:17] requires: [19:17] db: [19:17] interface: sqlite3 [19:17] do I have to re-deploy summit to use interface: pgsql? [19:17] imbrandon: do you want two smartfish mice? [19:17] man that error sucks [19:17] are they the same ? [19:17] negronjl: I dont use puppet.... I just use their monies :P [19:17] negronjl: review in progress.... [19:18] bkerensa: ... just giving you something to tell the puppet minions :) [19:18] SpamapS: th [19:18] SpamapS: thx [19:18] :) [19:19] negronjl: btw thanks for fixing the tabs :) [19:19] SpamapS: lol ... np [19:19] negronjl: though I'd rather have seen those fixed as a separate change.. its hard to seew hat you changed [19:20] SpamapS: I hate them as much as the next guy ... was just trying a new editor ... [19:20] been there, done that, got those scars [19:20] imbrandon: I have a fullsize and mini still in the box in my swag closet [19:20] ahhh kk i guess [19:20] SpamapS: I guess my thoughts on trying _not_ to complicate things with multiple MPs and more process actually made things more complicated ... :/ [19:20] negronjl: merge away [19:20] imbrandon: Im also getting some of those sexy mice soon though [19:20] SpamapS: thx [19:20] :D [19:20] negronjl: meh, I just did the diff w/ -b [19:21] mhall119: yeah, if it knows how to use pgsql then add another relation db: pgsql [19:21] that way you can keep the sqlite3 there [19:21] after you update the metadata run an upgrade-charm [19:22] and in the README you'll want to say "We use sqlite and postgres, but we recommend using $whichever" so people know how to deploy it [19:22] jcastro: +1 [19:23] man README should be a README, can juju charms grow a proper doc thats not 1000000000 miles long dude to it being best to put everything in there [19:23] marcoceppi: ok [19:24] With the "use a local db" vs. "use a remote db" [19:24] jcastro: that's going to be project-specific [19:24] I'd like to think that we make the charms default to using a remote db, and make a setting to use the local db [19:24] It causes a lot of problems if you default to local db [19:24] its neat for demos.. [19:24] SpamapS: I have it set to use the local sqlite until a db relation is added [19:24] SpamapS: I'd like to think that charms can interchange local to remote [19:24] django doesn't like running without a db [19:24] but ultimately its hard to migrate away from the local db to the remote db.. but its not hard to just "turn on" local db before any relations are made. [19:25] mhall119: yeah, I find that problematic [19:25] mhall119: so defer django until you have the remote db or have the setting to use the local db. [19:25] marcoceppi: sure, but you're left with this orphaned local db... who knows what state it was in.. [19:26] SpamapS: wouldn't the charm clean it up? [19:26] marcoceppi: I would hope not! [19:26] data is precious [19:26] SpamapS: it's using remote now! [19:26] dump to a consumable format, clean the database [19:26] save the data, but remove the moving parts [19:26] marcoceppi: right, but think of the instance where you set things up 90% of the way, then realized "wait we have to scale" .. you don't want the relationship to delete your old DB [19:27] you can remove the unit and the data will stay i dont see why moveing to remote would any more likely trigger a cleanup [19:27] Oh ok you meant something way more sane than I thought ;) [19:27] SpamapS: no, it should push the data out to the remote db! [19:27] Another thing to consider, maybe sqlite should be a subordinate [19:27] Then the workflow is the same, it just ends up local to the unit [19:28] SpamapS: i dunno, if its absolutely required as part of the code on a django app [19:28] why [19:28] do I have to expose the postgresql before my summit unit can see it? [19:28] e..g that sub is not good elsewhere is it ? [19:28] mhall119: no [19:28] or is there an internal private IP it can connect to? [19:28] imbrandon: actually it would be useful for openstack [19:29] imbrandon: nova also defaults to using sqlite, which causes problems because things register with the sqlite db, then the database is changed to mysql.. and the sqlite db registrations are not migrated [19:29] hrm well i intentially left the sqlite in the druapl charm, as its more like a flatfile fallback [19:29] not a use case [19:29] mhall119: the relation settings will provide the internal hostname for the hook to use [19:29] mhall119: the address that the pgsql charm returns is private and should be freely accessible from the summit unit [19:30] I have db_host=`relation-get host` in my db-relation-changed hook on the summit charm [19:30] SpamapS: no thats the apps problem as far as data integrity , every app already needs to deal with that as is [19:30] which gives me 192.168.122.168 [19:30] mhall119: right [19:30] File "/usr/lib/python2.7/dist-packages/psycopg2/__init__.py", line 179, in connect [19:30] connection_factory=connection_factory, async=async) [19:30] but I get that [19:30] psycopg2.OperationalError: could not connect to server: Connection refused [19:31] Is the server running on host "192.168.122.168" and accepting [19:31] TCP/IP connections on port 5432? [19:32] SpamapS: but yea , honestly doesnt that sound like a bug in Nova ( corner case rarely thought of on the dev side , but still one none the less ) [19:33] btw , drupal does the same and has the same issue if you dont check for it and most modules do [19:33] if they uise db direct [19:35] imbrandon: err, no. Its a bug in the charm. [19:35] imbrandon: the charm should migrate the data if its going to change the database target [19:35] i want cross environment relation hooks , please please [19:35] imbrandon: nova is doing exactly what it should. Writing data into the database it is configured for. [19:36] imbrandon: as is keystone, and everything else. [19:36] imbrandon: you can do cross env with a proxy charm [19:36] yea i considered that [19:36] Its not even a horrible idea :) [19:36] leave a 'juju open-tunnel' running so its not slow [19:36] nope, would samnitise it too [19:38] god why do i get so cranky/irritated at every little thing when i am sick, ugh, even irritated at that itself [19:38] bleh [19:39] back in a bit , gonna try to eat something an see how that goes [19:42] SpamapS, its a 95/5 rule re django generic ;-) [19:44] SpamapS: you got some time to show me how to make mysql create a db or really table is fine and use the MEMORY backend and then run a quick grant and schema creation + dataset on it, well thats assuming i dont just need to do it externally etc, kinda hoping its something i dont konw about internally that will do it so i dfont have to manage it seperately [19:44] if its just "run blah.sql" on server start then i can do that already [19:45] hazmat: but what happens when you cross from 95 to 5 and you're using the generic django charm? Fork it? [19:46] hazmat: because ultimately, those 5 percenters are the most successful users. [19:46] SpamapS: every django site will most likely have it's own charm [19:46] yeah that sucks [19:46] I agree, but I think we can fix that [19:46] exactly [19:46] with ineheritance [19:46] yeah, but that's because Django is a framework, not a service [19:46] not asking for it now [19:46] just saying, thats how you fix that [19:46] every django site would still have it's own child-charm [19:46] SpamapS, fair enough [19:47] mhall119, 95% of them shouldn't need it [19:47] in my experience, 100% would [19:47] they need a requirements.txt for deps, and db settings injection, and the rest is just what? [19:47] either that or a custom config so large that it may as well be a custom charm [19:48] mhall119, the charm is going to pull from git/vcs.. the user already has their specific settings configured [19:48] mhall119, what else do they really need? [19:48] hazmat: environment setup, db bootstrapping, etc [19:48] I have to agree that everybody *starting out* will just need a db and requirements [19:48] but even that case is *so* much smoother with inheritance [19:49] mhall119, ignoring django admin commands, which can be handled in a number of different ways.. environment setup.. is just config of settings.py typically [19:49] If I could write a django-myapp charm that just extends: django and drops a requirements.txt file in the right place.. bazinga, thats hot [19:49] SpamapS: you can make a sub do that :) [19:49] not really [19:49] Indeed, but thats clunky :) [19:50] hazmat: sure, it wouldent have the implioed checking inheritance would but would surely drop the half of the app you have in the sub anywhere including the riught place [19:50] hazmat: for summit we have other branches we need to pull in [19:50] mhall119: yea thats broken [19:50] I can write subs that drop files places. Thats the point of subs. Whether you *want* people to use them that way, thats how they'll be used. [19:51] mhall119, summit is a bit of a custom job i'd say [19:51] LTP is the same [19:51] this is the code I use if your curious: https://code.launchpad.net/~patrick-hetu/+junk/python-django [19:51] what django app isn't custom? :) [19:51] mhall119: i saw what chris did their, and hrmmmm yea bad juju [19:51] LIke if you don't need a custom app, use a CMS [19:51] heck, most of ISD's projects have various post-install things that need to be done [19:51] also checkout the *_site in my account to see how I deploy apps [19:52] then you have the question of is it using django 1.2 or a later version [19:52] is it using South or not [19:52] you dont HAVE to , i cleaned it up for him [19:52] YES! [19:52] SpamapS, mhall119 okay.. you've convinced me.. although the version and extensions can be done generically i think.. there are all sorts of things that aren't. [19:52] probably using celery will reqiure some other customizations [19:53] mhall119, not really.. celery config is pretty standard [19:53] no special db setup for it? [19:53] mhall119, no more so than for the django's rdbms db [19:53] ok [19:53] so maybe its 80%/20% ;-) [19:54] and btw , mentioning that ISD does something in a certain way pertaining to web app engeering is not a real heavy argument with me :) [19:54] * hazmat nods [19:55] imbrandon: it is to me [19:55] hazmat: I think 80/20 is definitely realistic in terms of one-db-simple-setup/customized-in-crazy-ways [19:55] mhall119: did you get your connection to pgsql worked out? you need to disambiguate the endpoint `juju add-relation summit postgresql:db` [19:56] mhall119: but yea the un-holly that was summits base theme checkeout that was 3 layed bzr forks at dirrent directory roots but overlayed and THEN more un-holly [19:56] yea its bad [19:56] I'm working on a 95% solution, that requires every project to add their own extra 5% [19:57] m_3: still no [19:57] hazmat: whats the story on bug 926550 ? Without it, we can't do proper testing of a precise SRU [19:57] <_mup_> Bug #926550: No way to test proposed updates to juju < https://launchpad.net/bugs/926550 > [19:57] imbrandon: I'd expect submodules could work for themes if necessary [19:58] mhall119: try using `postgresql:db` instead of just `postgresql` [19:58] m_3: indeed, i actually have a branche in LP where i had it all cleaned up like that anc cjohnston did not have time to "learn a new way" before UDS [19:58] imbrandon: gotcha [19:58] given it was only liek 2 weeks [19:59] but still not a hard change [19:59] right [19:59] brb [19:59] btw anyone mentions i actually admited to cleaning up django code and i will deny it [20:00] hazmat: Ugh.. http://askubuntu.com/questions/140818/what-is-the-best-way-to-use-the-mysql-charm-in-juju-with-dynamic-database-credie .. bad answer! [20:00] hazmat: the mysql charm *already* has that capability [20:00] hazmat: and we don't want to be encouraging forks! [20:01] SpamapS, then please correct away, it didn't have that ability to my knowedge [20:01] m_3: I ran juju add-relation summit:db postgresql:db [20:01] didn't work [20:01] drop the :db bits [20:01] SpamapS, it can create multiple dbs per rel? [20:01] imbrandon: I tried it without the :db bits [20:01] didn't work [20:02] db-admin: [20:02] interface: mysql-root [20:02] ok what dident [20:02] I tried it with *just* the :db on postgresql [20:02] didn't work [20:02] hazmat: gives you root, which is what he needs. [20:02] SpamapS, good point [20:02] add-relation summit postgresql [20:02] imbrandon: it gives the same error [20:02] it sends connection info to my summit hook just fine [20:02] ok then the metadata.yaml is likly off [20:03] it's just that the connection info doesn't work [20:03] SpamapS: +1 [20:03] can you pb the error and full metadata.yaml please :) [20:03] mhall119: strange... can you paste your metadata on each one? [20:03] * SpamapS answered, and -1'd [20:04] m_3: one minute [20:04] SpamapS, updated my answer as well [20:05] m_3: http://paste.ubuntu.com/1001655/ is the summit metadata.yaml [20:05] SpamapS: while i 1000% agree about no need for the shell etc, why are we discouraging forks, i missed that bit cuz everyone used to be all "everyone should have their own cusom local charms" iirc [20:05] mhall119, that's not valid yaml [20:05] i hope not anyways.. multiple keys of the same value? [20:05] hazmat: what? [20:05] db: && db: [20:06] imbrandon: because forks mean only the users of the new fork get the new functionality. [20:06] mhall119: with http://paste.ubuntu.com/1001657/ and http://paste.ubuntu.com/1001658/, I can use http://paste.ubuntu.com/1001666/ to drive it (note the postgresql:db in the add-relation) [20:06] imbrandon: people should have their own custom charms for the bits that are custom to their infra [20:06] SpamapS: and thats gonna happen if we fight it or use the energy elsewhere :) [20:06] SpamapS, forks are the lifeblood of evolution :-) [20:06] the dna even [20:07] Well I'm not saying people should just accept what is there as-is [20:07] mhall119: I see what you're trying to do, but we need to use a different interface to connect to both interchangably [20:07] I'm saying, people should be encouraged to *submit a patch* [20:07] not to just go off in the weeds with their own custom thing [20:08] the whole point of the charm store is collaborating around best practices [20:08] sure but what he had was not patch worthy and a local fork would have been just fine [20:08] mhall119: the trick would be to extend the mysql and pgsql interfaces to be generic relational-db with an extra db-type parameter [20:08] is what i mean [20:08] imbrandon: it was patch worthy, and we already had that patch, when we added phpmyadmin, we added it to mysql.. and now everybody has access to it. [20:08] m_3: ew [20:08] not the ":problem" of dynamic credentials [20:09] mhall119: then the client charm can do different things with different db types [20:09] imbrandon: that is the problem. He wants dynamic creds.. thats what a user with the grant option is for. [20:09] mhall119: otherwise, stick to a single type of backend [20:10] m_3: wouldn't you rather have a strong "I speak to these DB types" in metadata.yaml ? [20:10] ok maybe this is bad example cuz we;re on the tech bits a little much i was generalizeing a little more [20:11] SpamapS: then juju would need to learn every new dbms and not justa new charm that way wouldent it [20:11] imbrandon: You can have a requires: interface: db-with-no-charm and when that db gets charmed, viola, you can relate to it. :) [20:12] hrm [20:12] imbrandon: but I see the point [20:12] SpamapS: depends... sometimes no. in particular for a framework like django, I might lean towards a more generic interface [20:12] a generic one is not totally useless [20:12] anyway, I have to go afk for a bit [20:13] esp largely interchangeable ones in common webapps that abstract them with pdo anyhow on the code side [20:13] m_3: Yeah I do see that value.. where django can at least *try* to talk to anything its ORM can talk to, and we might not have charms for all of those yet. [20:13] jcastro, negronjl the review-queue web thing should be good now [20:13] and would just flip flop or be a real mess like drupal 8 and use all 4 types at the same time for diffrent tables and data [20:13] SpamapS: but yes, in general I do agree that your app should be opinionated about the relation store it uses... then it can optimize accordingly. Probably not a place to be super general [20:13] lol [20:14] hazmat: url path to review queue? [20:14] m_3, http://jujucharms.com/review-queue [20:14] gracias [20:14] m_3, see channel title ;-) [20:14] hey, fixed! [20:14] I changed it when I broke it [20:14] fixing [20:15] hazmat: doh === jcastro changed the topic of #juju to: Reviewer: ~charmers || Review Queue: http://jujucgarns.com/review-queue || Charms at http://jujucharms.com || Want to write a charm? http://juju.ubuntu.com/Charms || OSX client: http://jujutools.github.com/ [20:15] hrm wtf am i gonan do with drupal 8 , thinking about that its gonna be a real mess SpamapS [20:15] m_3, its tricky.. i never read that.. === hazmat changed the topic of #juju to: Reviewer: ~charmers || Review Queue: http://jujucharms.com/review-queue || Charms at http://jujucharms.com || Want to write a charm? http://juju.ubuntu.com/Charms || OSX client: http://jujutools.github.com/ [20:16] ok so now we have a queue, if you guys could post on the list on how we should resolve getting this down hardcore at first and then move to a more normal mode [20:16] (see my post about the queue from early today) [20:16] as it will seriously use mysql and pgsql and sqlite and memcache all at the same time for what in d6 and 7 are one single database [20:17] if you tell it to [20:18] bah i worry about that in 2 years when its released [20:20] imbrandon: specific interfaces (mysql,pgsql) can live right alongside more generic ones... we just have to keep the docs/examples/readmes straight so people know what to use [20:21] yea, plus i hope drupal learns of their madness before release on that one [20:22] they were trying like alot of apps do and solve a problem that needed to be done at another level of the stack inside the code [20:22] e.g. db HA in this case [20:22] makes me cringe eveytime i see stuff like that [20:23] https://tahoe-lafs.org/trac/tahoe-lafs <-- I just sent a mail to this guy for a charm [20:24] I can't believe we didn't think of it before, it's pretty clever [20:25] jcastro, sounds good about a charm for tahoe-lafs - one of my friends works on that project (zooko) [20:26] jimbaker`: he's the guy I just mailed! [20:27] jimbaker`: it's interesting because if you have multiple cloud providers ... it's like what Tahoe is for! [20:28] ok so we're down to 24 items now [20:28] I removed a bunch of mostly dead ones where the person wasn't responding [20:28] and told them to add the group if they need a review [20:28] so these should be the real deal [20:28] Why does GlusterFS show up if it's "In Progress"? [20:29] Libcloud is composed of multiple components, currently those are: [20:29] Compute - libcloud.compute.* [20:29] Storage - libcloud.storage.* [20:29] Load balancers - libcloud.loadbalancer.* [20:29] DNS - libcloud.dns.* [20:30] jcastro, very cool [20:31] marcoceppi: yeah, it shouldn't be... we've had problems in the past with removing things once they were in such a list... perhaps that's what's going on [20:32] is it promulgated? [20:32] it's showing up under ~marcoceppi in the web UI [20:32] jcastro: it's not promulgated because it's not done yet :) [20:32] oh ok so you don't need a review? [20:32] well, the glusterfs-server one is, the client still needs a few tweaks [20:32] nope [20:32] unsubscribe ~charmers [20:32] that's why it's in progress [20:33] Ah, just unsubscrived charmers [20:33] ah I see what I did there, I just swapped out the tag with the group [20:34] sorry about that [20:35] imbrandon: http://i.imgur.com/XKj2d.jpg <-- look at that sexy beast [20:35] jcastro: wait, so why'd you remove new-charm from Gluster? [20:36] m_3: we have the group now [20:36] we don't need the tag [20:36] jcastro: oh, sorry I'm behind [20:36] no worries [20:36] the post this morning on the juju list should explain it all [20:36] but note that it still shows up in the review queue [20:37] the TLDR is, if it needs a review, subscribe ~charmers [20:37] yeah [20:37] we just did that so we wouldn't accidentally lose one [20:37] but I went through and just fixed all the open bugs [20:37] Which direction will the weighting of resources go? [20:37] jcastro: so should I close #1003116 as invalid? [20:37] <_mup_> Bug #1003116: bugs that're In Progress shouldn't show up in the review queue < https://launchpad.net/bugs/1003116 > [20:37] * MarkDude was readling ML and sees this piece as important to getting it in Fedora repos [20:38] m_3: yeah that was just me messing it up [20:38] MarkDude: what do you mean? [20:38] jcastro: ok, thanks... now I'll go read the mail :) [20:38] m_3: though it probably shouldn't do that, dunno [20:38] now I'm thinking otherwise [20:38] bkerensa: do you really need me to take a pic of my ipad2 and ipad3 side by side here on the desk :) [20:38] prioritizing resources [20:38] m_3: "let's discuss next call" for that one [20:39] jcastro: ha! [20:39] I don't feel strongly about it, I mean, adding the group is fine I think [20:40] and if the person wants more time to work on it without being in the queue just remove the team and tell him, readd it when you're done [20:40] imbrandon: your ipad2 and ipad3 cost how much? I got like $5k worth of free stuff last year alone :P [20:40] that's how distro does it [20:40] MarkDude: I need more specifics, you mean as far as the python version? [20:40] bkerensa: maybe true but i got $500 that i actually use daily :) [20:41] resource maps [20:41] * imbrandon runs from those [20:41] libcloud [20:41] remembering that I am a bit like Jono on technical matters :D [20:41] * imbrandon runs from everyone [20:42] MarkDude: think of it like a internal spreadsheet that juju will keep to its self with cross refs of all the cloud providers and what they provide etc etc [20:43] SpamapS: bikeshed, our own calendar or just share the calendar with distro? i'd rather not proliferate more calendars in the project [20:43] past that it dont matter to anyonr but the core devs peepes [20:43] they have at most 3 people per day [20:44] Well from what I have gathered - this will make it a bit easier to *sell* to Fedora [20:44] * MarkDude understands linking it to USD is a method that will help management folks [20:45] And that this is a debate that has devs on one side, and managers on the other [20:46] I still have no idea what you want, do you mean like an easy explanation to explain to them why they'd want juju? [20:49] Well from my read of recnet emails, it looks like there is a debate [20:50] and that those wishing to tie resource maps to USD will win [20:50] * MarkDude has the simple explanation understood [20:50] sorta at least [20:51] * MarkDude is just not sure if this debate will be over soon [20:51] Or if you guys will drag it out a bit [20:51] Like Fedora is doing with naming debate..... still [20:51] MarkDude: oh, some of that to is just have its hashed out on this side of the fence, putting it to a any supported world bank currency even if we choose another would be dead simple [20:51] MarkDude: ok I need a link so I can catch up. :) [20:52] https://lists.ubuntu.com/archives/juju/2012-May/001542.html [20:52] towards the end [21:00] progmaticly currency converion is easy MarkDude so whatever is decided if fedora needed to change it to soemthing else its dead simple once the hard bits of calc are in place and thats given no matter what they choose unless its monoply money [21:00] That what I figured [21:00] The devs can be sold on juju [21:01] this part will help me get the manager types [21:01] or at least those that listen to them ;) [21:01] btw when you send that out or blog post it PLEASE cc me :) [21:02] not ment in a bad way but that is sooo going to be my post-uds train wreck i GOT to watch happen :P [21:03] Sure wooooowoooo trainwreck >>> https://docs.google.com/document/d/1h5kHxrn-DxotB-YCNHbdgfdhewkwVXCJ0D6WilkiDVU/edit [21:04] it will fly but still be very fun to see :) [21:06] * MarkDude has been given a blessing by default from some in leadership over this [21:08] MarkDude: a group of 11, when there is 12 in the group someone is always Lazurus [21:08] :) [21:08] lol [21:08] * MarkDude has had 10+ viewers of the document, and NO COMMENTS [21:09] Those with common sense are going to stay away. Hopefully for me to write my reality based objections [21:09] * MarkDude just assumes a few of you might find the hotdog debates funny [21:13] just change it to "Maddog" instead of Hotdog, in honor of Jon Hall [21:15] You like how I included my best link? [21:15] LUGOD, Maddog is the 1st video [21:15] * MarkDude is second [21:16] Better than ANYTHING else I have done in FOSS [21:38] jcastro: I know it seems like "moar calendars" but I'd rather have one calendar for ~charmers people to subscribe to, since I imagine a lot of ~charmers have no interest in Ubuntu dev [21:38] jcastro: but, yeah, green, red.. just pick a color for the shed :) [21:38] anyone seen a kfreebsd/ubuntu in the wild after debian started support for it ? [21:39] imbrandon: no, but upstart would be the major hurdle there [21:39] inotify *and* ptrace [21:40] zfs [21:40] mmm [21:40] threads mmm [21:40] so fast webserver mmmm [21:41] and m_3 is probably cursing me for highlights [21:41] lol [21:41] SpamapS: would dtrace and a userspace inotify take their place ? [21:41] given i have no real idea [21:42] er way dtrace is solaris [21:42] bah i give up [21:44] imbrandon: I've been told that dtrace != ptrace so no [21:45] imbrandon: and the inotify bits will likely just need ifdefs and then some assumptions in places after writing jobs will have to be replaced with 'initctl reload-configuration' [21:45] SpamapS: ok i JUST realized something that puts to rest in my mind resistance is futile and we should give in to libcloud and wrap it in twisted ( there is even an example to do it in the docs of libcloud ) lead dude on libcloud == cloudkick founder , cloud kick == rackspace now , rackspace == openstack [21:46] dtrace is awesome.. but osx, freebsd, solaris only [21:46] well.. oracle has a port in progress to linux.. but its not there yet [21:46] ahh yea i knew it was killer on solaris and osx i thought it was something else on bsd [21:46] imbrandon: zfs is meh.. there are ways to get that on Linux IIRC. Threads are not any better from the measurements I've seen. The webserver stuff I'm not aware of. [21:47] crossbow networking, zones containers, lets pick apart the corpse ;-) [21:47] SpamapS: yea they are, i've seen real scientificly done well not just back of the napkin that its MUCH better in bsd [21:47] not zfs the threads specificly for serving [21:47] but yea first hand [21:48] and really loook no further than ftp.cdrom.com [21:48] one lone machine [21:48] or was for years [21:48] Hah that still exists? [21:48] heh [21:49] no idea but i did do a shit ton of reasearch into it first hand about 3 years ago and its all still relevant [21:49] web serving on bsd is by far the best hands down, its just everything else that makes it impractical to do on a scale we need [21:50] thus a bsd+ubuntu would rock for that [21:50] imbrandon: you'd need *at least* a 10% improvement to justify the amount of work necessary [21:51] i'll put it like this i was saturating a 100MB core switch with a 600mhz celeron 1u server on a single ide drive on realse day for gutsy [21:51] serving ISO's as fast as i could on a bsd box to test some :) [21:51] imbrandon: nope, I've turned off mmm highlights _long_ ago :) [21:51] is that really important at this point? [21:51] SpamapS: well yes and no [21:52] SpamapS: for most no, for things the scale of wikimedia , very much so [21:52] and yes it was well above 10% [21:52] i do need to redo those control test though on modern stories [21:53] maybe when it gets cold outside and i need a heater in my garage like sabdfl, i'll see if i can peek out the newly installed google fiber :) [21:54] imbrandon: I'm betting nginx closes the gap and makes it a lot less about the kernel [21:54] imbrandon: given the scale of.. Facebook.. Google.. et. al. I think the linux kernel *probably* has scale figured out [21:55] btw you ever had a competent ccna dude tell you that you was killing the whole row of racks in the DC's core sw, hahah intresting moment it is [21:55] imbrandon: I mean, FB rewrote PHP before they tried to run on FreeBSD [21:55] SpamapS: actually your likely right, that makes a huge diff and thats why is cuz nginx does it more like bsd and apache uses builtins [21:55] ahhh [21:56] SpamapS: ahahahahhahhahha [21:57] imbrandon: yeah, nginx and varnish both do things "the new way" not just loading everything off to libc [21:57] just a php to c compiler :) roadsend is rewriting php, and qucurious already did in java for php5.4 100% coverage AND suposidly preforms identical to php + apc [21:57] err php to c converter, stillneed gcc to compile hiphop code [21:57] :) [21:58] roadsends php llvm compiler is native tho [21:58] but not done, only one done is the java one, and hiphop is "done enough" [21:58] to use for most things except wordpress [21:59] ( wordpress uses a few newer apis it hasent grown yet ) [22:00] they even released a jit hiphop runtime too reciently , not had time to even fire it up tho [22:00] but its said to work "inplace" like the php-cgi kinda [22:02] i wonder if i'm too damn old to go back and turn a cs to a doctorate and do some cool un-holly php compiler stuff [22:02] :) [22:03] hazmat: bug 926550 ? I absolutely have to upload that in the SRU or it will never pass verification [22:04] <_mup_> Bug #926550: No way to test proposed updates to juju < https://launchpad.net/bugs/926550 > [22:04] hazmat: you are marked as IN Progress on that [22:05] SpamapS, indeed i am [22:05] * hazmat checks the branch [22:05] hazmat: got something? I seem to recall a very simple change :) [22:06] SpamapS, it is pretty simple http://paste.ubuntu.com/1001944 [22:07] SpamapS, needs another test i suppose.. but how to test it functionally? [22:07] oh.. i guess just try it and see [22:08] un momento [22:08] <_mup_> juju/proposed-support r487 committed by kapil.thangavelu@canonical.com [22:08] <_mup_> merge trunk [22:08] hazmat: you can actually fully test it, you will just have "proposed" enabled completely [22:08] SpamapS, yup [22:18] * SpamapS sets maintainer on all his charms... [22:20] I have a challenge for any bored charmers [22:21] I want to see 3 services tied together in a way their authors never intended, but that is useful [22:21] like, vsftp <-> mysql <-> ??? that shows uploading to vsftp and then the file is owned by the same user who can log in to the web app [22:22] like, maybe a pam_mysql subordinate or something [22:22] imbrandon, for your amusement http://lists.fedoraproject.org/pipermail/advisory-board/2012-May/011612.html [22:23] SpamapS, that sounds like a great challenge. That sortof thing is what draws me to juju :) [22:24] MarkDude: shall I assign the bug to you then? ALLLLLLRIGHTYTHEN [22:24] ;) [22:24] Well not so mcuh [22:24] * MarkDude is good at talking, bbq, and wearing a penguin suit [22:25] sometimes talking while wearing a penguin suit [22:25] MarkDude: that sounds like 3 services that weren't intended to be integrated [22:25] Think Jono, but less charming [22:25] lol [22:25] juju deploy markdude-in-a-penguin-suit [22:25] juju deploy bbq [22:25] good point [22:25] juju add-relation markdude-in-a-penguin-suit bbq [22:25] awesomeness [22:25] I imagine you'll call that relation 'spatula-zipper' or something of that sort [22:26] that or things that *just sorta happened* [22:26] We'll cal that the "unholy" stack [22:26] juju destroy-enviroment # party killer [22:26] Never was my intent to be known like this. Minus the bbq thing ;) [22:27] Daviey: every party needs a pooper thats what we invited juju fer [22:27] heh [22:28] Daviey: perhaps MaaS could grow support for match-light charcoal? [22:31] * MarkDude wishes there were a way I could juju deploy sense-of-humor to some Fedorans [22:31] SpamapS: sounds like a power control provider.. so sure. [22:31] MarkDude: closest thing you will get is a subordinate for 'sl' [22:31] Daviey: it is, though cooling is an issue on systems plugged into it. :) [22:32] maybe next versions [22:32] time to leave the coffee shop, later folks [22:32] maas match-light --constraint=196-arms [22:33] SpamapS: if match power control is bucket of water, it's less idempotent.. which will be an issue. [22:33] maybe bucket o water + hair dryer. [22:33] imbrandon: did you see that Debian may drop Wordpress [22:34] Daviey: I think in order to make charcoal idempotent, you have to be omnipotent [22:34] no but not suprising and its been talked about before [22:34] imbrandon: how's that "pull from upstream" charm going then? ;) [22:35] SpamapS: imho probably for diff reasons i think it should too , drupal , wordpress, phpmyadmin et al really move to fast for debians pkg model [22:35] SpamapS: its done and on github, i'm doing some more "cool stuff" before i get it promed [22:35] hrm, i'm not sure everyone will be happy if phpmyadmin went to juju only. [22:35] imbrandon: its really about upstream commitment.. if the security patches aren't going to be backported by upstream to anything.. then its not a good fit for distro [22:36] Daviey: yea i dont think juju only is the way either, dont know a great way, but i can agree with their arguments [22:36] Even though mysql isn't disclosing their security bugs, they're at least releasing them as patch-only releases [22:36] SpamapS: right [22:37] thats really the issue w/ wordpress et.al [22:37] SpamapS: i hoped the mysql issue was making progress? [22:37] Daviey: the only progress is that we're crossing our fingers that they don't break stuff [22:37] and even if they did no one would use it, the web changes to much even for a secure wp to be used over 5 years for the 80% market [22:37] Daviey: *everybody* is in the same boat [22:38] SpamapS: not really we can choose to do ci and make the commit ment to make sure the newest is best not just dput and pray but thats against alot of history too [22:39] i think pulling it from the archive with a eeasy installer like composer to get apps ( e,g npm for nodejs ) is gonna be the way to go [22:39] drupal has already moved to composer as has zendframework and symphony and such [22:40] imbrandon: I'm saying we're crossing our fingers on the mysql stuff [22:40] oh yea [22:40] mysql is ALMOST as much magic to me as X [22:40] CI is actually exactly what we'll be doing with the juju charms [22:40] i can configure MYSQL tho [22:40] :) [22:41] SpamapS: right thats what i was getting at , i'm just still sick as hell and i'm sure making even less sense than i do most dayss [22:41] i am feeling better tho and on my 4th gatorade today [22:41] heh [22:42] no caffeine for me is kickin my rear [22:42] yeah I get a bad headache when I try to ignore the siren call of that sweet sweet adenosine disruptor :) [22:43] btw i tell yea i snagged a lil dell insp 1101 [22:43] imbrandon: are you just a swirling vortex of moderately powered hardware or what? [22:43] not very much horsepower at all but i'm plsently suprised and unity isnt half bad on it [22:44] nah, i just got rid of that macbook i had at uds, it was a pos [22:44] then gave my kid brother the other MBP [22:44] so i only have my mini no and the lil dell + ipad [22:44] now* [22:45] less management and not too bad off, not a mba, but i can take the $$ trade off considering i got it basicly free for helping a friend out the other day for a half hour [22:46] keeping 2 os's updated on 3 machines and that was just my daily workstations not servers etc was quickly a pita [22:48] archived my osx into a vm that is off till needed and lil-dude ( yup thats the hostname ) and my mini are running my interpretations of dell dev distro :) [22:48] whatever it was called [22:48] * m_3 likes "swirling vortex of moderately powered hardware"... thinks SpamapS is feeling poetic or something [22:49] i still think a "dev" desktop machine in the cloud would be an ok charm [22:49] hehehe [22:50] i have noticed though unity is far less annoying on the 10.1 screen [22:50] well apart from the icon launcher is HUGE if not hidden [22:51] but not making me want to kick a cat like on my desktop [22:51] that and the only hotkey i need to rember is super [22:51] heh [22:52] * m_3 remaps super [22:52] SpamapS: what list is dd taling about it on [22:52] m_3: i did on my desktop so it would work with my muscle mem of OSX [22:53] imbrandon: yup [22:53] i flip ctl and super [22:53] actually looking over at the windows keyboard on ther it would be alt and super [22:54] but on mac its ctl and super [22:54] I actually have a mac keyboard on my pugetsystems.com desktop [22:54] so it matches my laptop [22:54] and I don't have to confuse the poor muscles [22:54] so bottom left of mmy keyboard is ctl->alt->super->space maped in all os;s [22:54] yup [22:55] thats exactly why i buy the bluetooth mac keyboards they work with everything [22:55] and the reason i had it seperate at uds as the old macbook keymap was diffrent but the bt keyboard and the mbp and the usb all match [22:55] heh [22:56] other wise it woulda been kinda silly to have a sep keyboard , i'm sure it looked now that i think about it [22:56] heh [22:56] yeah, I really like these little mac bt kbds [22:57] yea i'm on like #5 now, 3 still in full time use [22:57] 2 have seen better days and are donor boards for keys and such [22:58] imbrandon: debian-devel [22:58] the one i use on my desktop in the other room i have a bar that connect it to the magic trackpad and gives it induction recharging [22:58] cool [22:58] so its the same form as the full usb one [22:58] when connected [22:58] ponly trackpad and no number pad [22:59] m_3: I emerged from my mothers womb laden with the sorrows of the world on my back, each one crying out for prose and poem that might transform them to joy like so many caterpillars turn to butterflies [22:59] * SpamapS farts [22:59] m_3: yea if you have a trackpad as well , try one , they are very nice [22:59] http://www.amazon.com/Twelve-South-12-1101-MagicWand-Connects/dp/B004L9M0AO/ref=sr_1_3?ie=UTF8&qid=1337727556&sr=8-3 [23:00] mine is acutlly another brand but almost identical, it just is a bar that snaps on the back at the top over the battery bar on both of them [23:03] hahah and their pictures of it down below have the trackpad hooked on the left and mouse on the right, like i do anyhow but just not hooked up on this desktop with a caption of "photoshop with two hands" ... thats too funny [23:03] * m_3 snaps repeatedly [23:03] oh wait... need to find the beret [23:03] imbrandon: nope, haven't tried the trackpad [23:04] was hoping to get the little mouse with the trackpad on its back working... but no love last I tried [23:04] yea i got it before i got a real mbp so i could have multi-touch on the older white macbooks and was hooked [23:04] oh it is [23:04] yeah, multi works great on the mbp [23:04] and its the whole top of it [23:05] imho i would get the pad before the mouse and stick with a mighty mouse or normal roller mouse [23:05] cuz the magic mouse refuses to click AT ALL if your hand are dirty in the slightest [23:06] sweaty , anything [23:06] i'm constantly needing to wipe the top glass off for it to fully work [23:06] I was shocked the other day when 3 finger drag on Ubuntu sent my firefox window flying off the screen [23:06] man that was right over the plate... I'm gonna let it pass tho :) [23:07] heh try 3 finger push then SpamapS you'll love that [23:07] its the expose iirc , it is on osx at least [23:07] i think unity too [23:07] whats this osx you speak of? [23:07] ;) [23:08] :) the thing that unity got all the muti touch goodness from [23:08] :)( [23:08] hehe [23:08] the thing my wife makes me keep on the tv [23:08] I should switch it while she's out of town... see how long it takes to notice [23:09] no but seriously think about how you drag in the browser with two fingers without thinking now, when you learn them all even the 4 finger combos , man its very very nice, almost like an emacs mouse user [23:10] I try to ignore that the mouse even exists really [23:10] like i drag 4 down and it shows "Launchpad" aka the same thing you get in unity when you press super, whatever thats called [23:10] aka all my app icons [23:11] and then up to see all my running apps and all their windows at once [23:11] imbrandon: how'd you launch the unity launcher? [23:11] yea i use as many mouse and trackpad gestures as most linux peeps traditionally use in emacs keyboard shortcuts to exit when they are done :) [23:12] m_3: on osx the equiv is 4 fingers down [23:12] pull [23:12] oh, sorry I thought you meant on unity [23:12] it launches "Launchpad.app" [23:12] i did [23:12] i hit super to do that [23:12] just super [23:12] I've been trying to figure out the command or signal to get the launcher to launch [23:12] besides the super key [23:13] oh thats all i know so far [23:13] not dug enough,if i find it tho i'll holler [23:14] but i've begun to just hit super and type, kinda anying i cant then just arrow down to the icon i need to tab then arrow [23:14] I'd like command-space for that, a key-combo and not a single key [23:14] but i'm getting used to it [23:14] oh yea, cmd-space spotlight is killer [23:14] the one that's killing me right now is the timing for ctrl-A [23:14] but imho the dash and launchpad are more alike than cmd+space spotlight although technicly your right [23:15] it's like it's not engaging the control key half the time I press ctrl-A [23:15] ctl+a bckspace ? [23:15] this is for tmux [23:15] ahh hrm, might be your key too, tried another ? [23:15] ctrl-a to prefix everything [23:16] m_3: there's something wonky with keyboard + trackpad on the macs [23:16] but with the capslock mapped to control, I'm a _lot_ faster in doing ctrl-a [23:16] yea , thats what i use it in too ctl+a then backspace == one wondow back [23:16] over and over till i get where i want [23:16] m_3: my Air especially has something weird with timing.. left-Alt-tab shouldn't bring up HUD, but it does unless I hold it down for a while [23:16] it's like it's not getting the timing right... but it's not consistent [23:16] ha! [23:16] hahah [23:16] yes, I've noticed timing issues in general for hud and dash [23:17] I want to turn off HUD.. its not configurable tho :-/ [23:17] ^5 SpamapS [23:17] it's like I've got to press alt as 'ta-dah' to get the hud up [23:17] bout time [23:17] lol [23:17] win 16 [23:18] m_3: esc+y [23:18] better than /win N [23:18] esc+N [23:18] 11 starts with q [23:18] err not esc+N , esc N [23:19] seperate [23:19] but yea, i've grown very used to that key combo too and less wrong window messages that way tooo :) [23:19] imbrandon: I don't have that bound in irssi that way [23:19] ahh i thought it was default that way [23:19] could be :) [23:19] ive had the same irssi since breezy tho [23:20] configs that is [23:20] my irssi goes back to when bitchX's upstream did like online hari-kari and removed all trace of bitchx [23:20] heh [23:20] ha [23:21] I even re-did my bitchx theme .. osbxwannabe [23:21] i flirted with xchat and bip for a few weeks [23:21] http://irssi.org/themefiles/osbxwannabe.png [23:21] hahaha.. wow.. look at that [23:21] thought about quassel, but you just really can't beat irssi at this point [23:21] but cant shake irssi no matter how much i try [23:21] nice [23:22] Last-Modified: Sat, 17 May 2008 15:39:06 GMT [23:22] I bet thats when irssi.org migrated, I think I did that theme way before 2008 [23:22] :) [23:22] hi guys! I'd like to know if any of you is interested in driving a session in the Ubuntu User Days, which is for users who are very new to Ubuntu. it can be about juju (as long it's very very basic), or what you want to do it. the User Days are going to take place between June 23-24. [23:23] * SpamapS dives into the nearest trashcan [23:23] funny thing is i think i saw one yesterday on flickr of mine from close to the same time, /me looks [23:23] * m_3 scatters [23:23] JoseeAntonioR: jk... sure, I can do one [23:23] :P [23:23] JoseeAntonioR: this is an IRC event right? [23:24] * m_3 checks what he's committing to (for once) [23:24] yea in the classroom like normal [23:24] m_3: yep! [23:24] imbrandon: hey, and what about you? [23:24] a second juju one ? [23:24] m_3: is there any particular topic you're interested in? [23:24] JoseeAntonioR: just call it charmschool or intro to juju or something similar [23:24] imbrandon: you can choose another topic, if you want to [23:24] or did you mean something else ? [23:25] hrm , i'll do one on from not ahving juju installed to a full working drupal install with custom themes and all [23:25] that should be enough diffrent [23:26] imbrandon: remember it's for new users in Ubuntu, they're just familiarizing with the environment [23:26] oh right [23:26] let me think on it a day [23:26] and i'll poke ya tomarrow [23:26] imbrandon: there are some suggestions here: https://wiki.ubuntu.com/UserDaysTeam/CourseSuggestions [23:26] if you still have a spot [23:26] great, thanks! [23:27] kk [23:27] we sure will :) [23:28] JoseeAntonioR: how about "command line basics aka for the ultra new users [23:29] imbrandon: that would be great, that session always has a great audience [23:29] kk cool [23:30] imbrandon: when you're ready, just tell me the session and the time you want it to be [23:30] JoseeAntonioR: i normally just tell daniel or whom ever to pencil me in as a floater so others can have their prefered times as any will work ok for me [23:31] and i'll fill in with one thats left [23:33] imbrandon: if that's fine for you, that'll work for us too [23:33] kk yea , i'll just check it the day before and see where ya put me [23:33] ty [23:34] thanks to you! [23:34] command line basics, right? [23:34] yup [23:34] great, thanks! [23:38] zomg my music went from "Halestorm - I miss the missery" to "Alabama - Lady Down on Love" banshee needs to grow a brain like iTunes :) [23:38] man thats not even cool, i was rockin out too [23:38] lol [23:42] got two 4x8 sheets of holed pegboard and some 1x1's to add a tall cube like backing on the right side and behind my monitors "walls for hangin stuff" that i wanna really feel better so i can get put up [23:43] my psudo desk frankenstin is quickly becomming a fortress [23:44] imbrandon: I printed some ubuntu posters to cover my empty walls [23:44] they look pretty cool [23:46] m_3: one other thing i would add is if you do want a mouse and not the pad, get one with the blue led's in them if you are mobile more than not, its not all hype and those blut track lights really do pick up on most any surface even shiney ones MUCH better [23:46] JoseeAntonioR: :) [23:47] imbrandon: yeah, I haven't been doing much blender recently... that's my sole reason to mouse really [23:47] MS calls their bluetrack and logitec its TrackAnywhere or somerthing, but they are all blue leds [23:48] yea and they arent much more if any then the other mice [23:48] maybe 5$ diff or something [23:49] doh, I just realized the need-maintainers.txt thing was pulling from the wrong dir.. a static copy of the charm store [23:49] sometimes they are branded at "couch mice" too [23:49] * SpamapS updates [23:49] oops [23:50] someone tried to explain that is was because the blue light wave is larger than red , and while i know thats tecnicaly true i'm not sure mice are that sphostica but maybe [23:51] even more oops was fork bombing the t1.micro that I have it running on with 81 bzr pull's [23:51] oooooops [23:51] HAHAHAHHAHA [23:52] oh man then form Beaste Boys to Alan Jakson, this thing is gonna make me make my own playlists isnt it [23:52] bleh [23:53] Heh, I just heard Coldplay's crappy ass "You gotta fight" [23:54] oh i'[ve avoided that so i would not want to murder them [23:55] there are a few covers that are as good or better, but man coldplay should not have even considered beaste boys, and stuck with like a vanalla ice songs or somethin [23:55] i rag on them alot tho as it was my ex's fav bands [23:55] heh [23:55] :) [23:57] now Chris Cornell: Billy Jean, better than MJ and thats hard to do. Johny Cash: Hurt, is so much win i dont even wanna hear the NiN version ever again [23:57] m_3, looks like the oneiric branches are back [23:58] m_3, who did the magic?