=== gary_poster is now known as gary_poster|away === thumper is now known as thumper-afk === defunctzombie is now known as defunctzombie_zz === thumper-afk is now known as thumper === lynxman_ is now known as lynxman === andrewdeane is now known as awd === ubot5` is now known as ubot5 === wedgwood_away is now known as wedgwood === wedgwood is now known as Guest65596 [12:24] how does one install and use the python helpers in charm-tools? I see a charm-helper-sh package but no python equivalent. === wedgwood is now known as Guest36969 [12:43] You can install them into your system python like so: sudo apt-get install python-charmhelpers [12:48] benji: except you need to add the ppa first [12:49] ooh, indeed [12:49] Which is ppa:charmers/charm-helpers, right? [12:52] there's a recipe to build into the juju ppa https://code.launchpad.net/~juju/+recipe/charm-tools but it's broken atm [12:53] i've been trying to fix that, but seems a bit more complicated than i thought [12:54] i *think* that https://code.launchpad.net/~ubuntu-branches/ubuntu/quantal/charm-tools/quantal is correct, just missing pep8 in build-deps [12:54] any of these pyjuju test failures ring a bell for anyone? http://pastebin.ubuntu.com/5673475/ [12:55] whereas https://code.launchpad.net/~ubuntu-branches/ubuntu/raring/charm-tools/raring r18 looks like it reverted some of the previous changes [12:56] mgz: no idea about that, maybe a change in lxc? could you help with the charm-tools packaging issue above? [12:57] probably just a missing builddep? [12:58] mgz: yes, missing pep8 apparently. but the raring branch looks like it removed python-charmhelpers accidentally on r18 [12:58] ...which is what you said : [12:59] there doesn't seem to be a raring build anyway [13:00] I can probably fix this stuff up [13:02] benji: I don't see a python-charmhelpers package in that ppa [13:02] that would be awesome. i had a stab at fixing it before realizing that r18 on the raring branch https://code.launchpad.net/~ubuntu-branches/ubuntu/raring/charm-tools/raring/+activereviews and then got pointed at bzr split documentation and lost all interest :) [13:04] wedgwood: hmm, it may be ppa:juju/pkgs and the package name may be just charm-tools; I'm looking into it real quick [13:06] benji: looks like charm-tools exists in raring, it just doesn't include the python stuff [13:06] wedgwood: yep, it seems to be ppa:juju/pkgs and the package name is python-charmhelpers [13:07] http://goo.gl/DuHzm ? [13:07] i suspect it got unpublished with the latest recipe failure [14:14] http://juju.ubuntu.com/survey is live! === defunctzombie_zz is now known as defunctzombie === wedgwood is now known as wedgwood_away === wedgwood_away is now known as wedgwood [15:18] jcastro: bummer you guys couldn't do the survey using open source like the last one ;) === salgado is now known as salgado-lunch [15:23] oh the limesurvey? [15:23] you can't make me use bad tools. :p [15:23] jcastro: how is it bad? [15:23] like seriously [15:23] Its *at least* as powerful as surveymonkey [15:24] well, surveymonkey is bad, so I guess that's not saying much [15:54] mgz: thanks, I realized last night I didn't have perms to push to lp:ubuntu/charm-tools [15:56] m_3: we can bug someone with the right superpowers to do it though [15:56] and the recipe build can just come out of a branch under ~juju [15:57] I'm nearly there with the various packaging fixup pain [15:58] mgz: and yes those tests look familiar... I put 'export DEB_BUILD_OPTIONS=nocheck' into ~/.pbuilderrc to focus on the packaging... /me guilty look [15:58] :) [15:59] and I got a different set of failures on the buildds... at least the tests pass when run in branch, and the build I cared about worked [16:01] bummer... I hate chasing inconsistencies === salgado-lunch is now known as salgado [17:00] SpamapS, hi [17:06] koolhead17: howdy! [17:07] SpamapS, am good. === gianr__ is now known as gianr === salgado is now known as salgado-dr [17:55] m_3: good thing you waited to fix your blog, now you get 50% lower prices on aws! [18:01] my AWS bill has dropped consistently over the 2 years I've run a t1.micro for fewbar.com .. I think its $13/month .. and thats including about 800MB of backups going to S3 weekly. [18:02] (my hpcloud bill is $17/month for my minecraft server) :) [18:06] jcastro: so this wsgi charm... how different from say the gunicorn charm that is? [18:07] sidnei: I was just announcing it, m_3 is the one to talk to [18:07] m_3: ^ [18:08] jcastro: I don't pay anything for it... it's just github pages [18:10] sidnei: it'd definitely overlap... I just wanted to file it as a separate bug to open the conversation about _which_ webserver is the most commonly used for microframeworks plugged into wsgi [18:10] i don't think there's a definitive answer for that, although gunicorn is quite popular [18:10] m_3: so, I think I've now done all the boring bits (bar some posts to the mailing list) === wedgwood is now known as wedgwood_away [18:10] sidnei: ack... that's ok [18:11] mgz: whoohoo! [18:11] m_3: what's up next is merging in your update-alternatives work to the packaging branch, testing it, and getting it in raring [18:11] mgz: thanks man, I wanted a sanity check on those [18:12] mgz: do I have to do something specific to the changelog? that's why I didn't do a real MP before [18:12] we'll add something, but that's pretty simple [18:12] do we leave it alone at 0.6-1ubuntu2 iirc? [18:12] or bump [18:12] coo [18:12] l [18:15] I've bumped to 0.7-0ubuntu1 [18:16] and want to do the install changes on a new version as they're pretty major, so 0.7-0ubuntu2 [18:17] mgz: ack [18:18] m_3: im looking for a solution to using wedgwood_away's lp:charmsupport on the apache/squid/haproxy charms. i think the best option might be to include the package in the same ppa as charm-tools and add the ppa in the install hook. [18:18] as opposed to having a separate ppa that is [18:19] sidnei: hmmmm [18:19] sidnei: will that cause problems with the juju version installed [18:19] sidnei: i.e., if they deployed orig from the distro version of juju [18:20] m_3: meaning charm-tools is in the juju ppa? i thought it had its own ppa [18:20] sidnei: then, the charm adds that ppa, then something during the life of that service updates/upgrades [18:20] oh, nope, I thought it was the same as juju.... /me looking now [18:21] sidnei: https://launchpad.net/~juju/+archive/pkgs [18:21] all together [18:21] looks like also https://launchpad.net/~charmers/+archive/charm-helpers [18:22] * m_3 facepalm [18:22] an alternative might be to just move all of that into the python charmhelpers [18:22] sidnei: I have no idea what we should do with that then [18:23] there was some concern that nrpe stuff shouldn't live there for example [18:23] sidnei: lemme look at lp:charmsupport [18:24] they're really two different things.... helpers are for charms... charm-tools are mostly cli tools for charmers [18:24] the biggest blocker atm for me is that the relation_get function from python charmhelpers can't be changed in a backwards-compatible way [18:25] are you saying that helpers should be split from charm-tools? [18:25] sidnei: yes [18:25] might not solve this particular problem [18:25] but in general, those stand separately [18:26] hmmm, yeah, that's what I was gonna ask... what's in the way of integrating charmsupport into python charm-helpers [18:26] not necessarily no. but i could move the 3 functions from python charmhelpers that im using into charmsupport and ignore python charmhelpers from now on. [18:26] * m_3 facepalm again [18:26] it'd be great to build python charmhelpers [18:26] yes, i join you in that [18:27] to build python charmhelpers as in having a proper set of charmhelpers in a package that charms can use? [18:27] lemme look at relation_get in charm helpers [18:27] which may or may not be the current python charmhelpers that are part of charm-tools [18:27] yeah! [18:28] relation_get in charmhelpers returns the result of relation-get in whatever the default format is, instead of the json-parsed python structure. [18:28] I think helpers were glommed together into the charm-tools project temporarily... [18:28] sidnei: ah [18:28] i don't think i can change it to simply add json.loads() and still be backward-compatible [18:29] right... I should do a big pull and start grepping around for deps on the various charm-helper packages [18:30] prbably have that on disk on one of the testers atm [18:30] actually seems only get_config() does a json.loads(), none of the others do in charmhelpers [18:31] so if we could reconcile that [18:31] sidnei: gotta run to a meeting... lemme know if you see a clear solution, I'll take a look this afternoon [18:31] sorry man [18:31] m_3: no problem. i might throw it in an email, im off next week. [18:34] niemeyer: heya, publish looks awesome, when do you expect to land it? Wondering as far as documenting it for the submissions guidelines and all that. [18:35] jcastro: I've been landing it in pieces.. I just started working on it again 30 mins ago.. if I manage to keep my focus, perhaps sometime tomorrow it should be all in === salgado-dr is now known as salgado [18:41] niemeyer: oh ok, so for 13.04 then, that's brilliant [18:41] uh [18:41] 13.04 what? [18:41] * SpamapS saw something about a final beta freeze yesterday [18:43] I don't know what the status of juju's FFe is offhand [18:44] IIRC, juju is still unseeded. So.. the freeze is nominal [18:44] does juju-core exist in Ubuntu yet? [18:44] is unseeded meaning ... ? [18:45] https://launchpad.net/ubuntu/+source/juju-core [18:45] no existe [18:45] jcastro: unseeded is the new universe man [18:45] oh, I believe this is what mgz/mims were working on today? [18:46] https://launchpad.net/ubuntu/raring/+queue?queue_state=0 [18:46] jcastro: you'll also have to tax an archive admin to pass NEW [18:47] should have done that months ago [18:47] I don't know who's working on the package [18:47] * SpamapS will take some responsibility for ditching in December ;) [18:47] I care more about it being in the ppa for 12.04 tbh [18:47] jcastro: totally, I'd ignore distro for raring [18:48] I don't think I can get away with ignoring it [18:48] :) [18:48] you're going to create a bunch of work for over-taxed individuals because the plan was to work right up until the last final possible moment [18:49] this was what I was upset about.. calling it "13.04" when you have 6 months of new work to do isn't really possible and, big surprise, the schedule slips. [18:50] I more mean about the PPA === wedgwood_away is now known as wedgwood [19:18] jcastro, I asked plars to swing by to help out [19:19] ok [19:19] I'm assuming since the jobs are on jenkins.qa.u.c he might be able to help you ;-) [19:19] m_3: ^^^^ [19:21] jcastro: if you know the job name, I can check to see if it's running on one of the jenkins instances I have access to [19:21] nothing actually runs on jenkins.qa.u.c, it's just for showing results [19:23] jcastro: thanks [19:23] plars: hi [19:24] hi m_3 [19:25] plars: hey, so I've got some instances running that publish to jenkins.qa.ubuntu.com [19:25] plars: the publication mechanism seems to be broken [19:26] plars: can you look at the logs to see what's up with... for instance... precise-openstack-charm-bitlbee [19:26] as a job name [19:26] m_3: ah, which jenkins instance actually runs the jobs? [19:27] m_3: typically, if publication is failing, it will be on the jenkins system that runs the jobs, not the one that it publishes to [19:27] plars: these're running on instances in ec2 [19:27] m_3: and unfortunately, restarting jenkins is usually the only fix (the one running the jobs) [19:29] plars: yup, they're actually ephemeral and come to life every day [19:29] plars: they're configured automatically with juju [19:29] m_3: and they've previously worked, but suddenly stopped? [19:29] plars: james page orig set up creds for publishing to jenkins.qa.ubuntu.com [19:30] plars: yes, they were working in the past [19:30] plars: then they stopped a few months ago due to version changes [19:30] plars: then were running again [19:30] m_3: any chance they are getting shut down before publication has finished? [19:30] plars: and stopped about a month ago [19:31] m_3: I don't have any access (other than the public page) to jenkins.qa.ubuntu.com, so everything I can see is what anyone can see on that [19:31] m_3: If it's an access problem, likely james or IS will have to handle it === bac_ is now known as bac [19:31] m_3: if there's any way to keep the instance alive after it finishes for a bit, for debugging though [19:32] m_3: that might be helpful so that you can see where if it's just dying to quick, or if there's an error on the jenkins running in ec2 when it tries to publish [19:35] plars: he just texted me that his network is out. [19:36] whew [19:36] ok, back [19:36] m_3: wb :) [19:36] plars: sorry [19:36] m_3: what was the last thing you saw from me? [19:36] one sec [19:38] plars: ok, so you were asking if the jobs died before they could publish [19:39] plars> m_3: I don't have any access (other than the public page) to jenkins.qa.ubuntu.com, so everything I can see is what anyone can see on that [19:39] m_3: If it's an access problem, likely james or IS will have to handle it [19:40] m_3: if there's any way to keep the instance alive after it finishes for a bit, for debugging though [19:40] plars> m_3: that might be helpful so that you can see where if it's just dying to quick, or if there's an error on the jenkins running in ec2 when it tries to publish [19:43] plars: ok, yeah I just wanted somebody to check the logs on there while I was publishing [19:44] m_3: I have no ability to do that [19:44] plars: yeah, I can control thelifetime [19:44] plars: ok, cool... I'll ping is [19:44] plars: thanks! [19:44] plars: oh, btw... do you manage the disk space on there? [19:44] plars: i.e., if I wanted to remove old build artifacts [19:44] m_3: nope, that would also be IS, sorry :( [19:44] ack [19:44] plars: thanks === _mup__ is now known as _mup_ [20:24] marcoceppi: you had plans to rename shelr.tv to something else right? [20:24] jcastro: it's already been renamed [20:24] jcastro: [20:24] jcastro: http://jujucharms.com/charms/precise/shelrtv [20:25] ack === defunctzombie is now known as defunctzombie_zz === sander__ is now known as Sander^home [22:29] Tried to follow the EC2 Juju guide and it fails | http://askubuntu.com/q/277638 === thumper is now known as thumper-gym === wedgwood is now known as wedgwood_away === defunctzombie_zz is now known as defunctzombie