[06:17] <bashok> Hi Guys, after all the hassles i was finally able to see the bootstrapping process made some progress, however still it was unable to start the bootstrap instance due to this error
[06:17] <bashok> " ERROR juju supercommand.go:282 cannot start bootstrap instance: index file has no data for cloud {blr http://192.168.124.81:5000/v2.0} not found"
[06:17] <bashok> any idea how to solve this
[06:19] <bashok> sarnold, davecheney any thoughts?
[13:23] <ashipika> hi guys.. anybody there?
[13:24] <mgz> asking something like that just just begging for a "no" :P
[13:24] <mgz> this is irc, if you have a question just write it and wait
[13:25] <ashipika> sorry.. sorry.. a moment of panic :) logged into launchpad, all my code was gone.. after a minor heart attach figured out i used a wrong account.. all better now..
[13:25] <mgz> :D
[13:26] <ashipika> sorry again
[13:26] <mgz> happy ending
[13:53] <arges> I've used the 'juju sync-tools --source <path>' sucessfully after downloading tarballs locally. Is there a way to have --source point at a remote URL over HTTP? Thanks
[15:00] <marcoceppi> hey sinzui - is there anything I can do about this other than wait? https://bugs.launchpad.net/charm-tools/+bug/1262836
[15:00] <_mup_> Bug #1262836: charm-tools fails to build on trusty <trusty> <Juju Charm Tools:Triaged> <https://launchpad.net/bugs/1262836>
[15:00] <marcoceppi> I didnt' see a workaround in the bug report
[15:00]  * sinzui looks
[15:01] <marcoceppi> oops, you didn't report that, jjox did
[15:02] <sinzui> hmm, recipes always force native version, so the issue is to change the package version string
[15:07] <sinzui> marcoceppi, This does look like the bzr-plugin forcing the recipe to build as native. Lp always rung with fallback mode
[15:08] <sinzui> marcoceppi, I experienced this in Oct/Nov and found that I could make a version string that was acceptable. While I could get a recipe for juju-core, I abandoned it.
[15:08]  * sinzui looks for old nots
[15:11] <sinzui> marcoceppi, change the plus to a squiggle in the recipe. I seem to have played with that a lot
[15:11] <marcoceppi> sinzui: there's not recipie for this, I've been cutting releases myself
[15:12] <sinzui> marcoceppi, This is what I got to work before abandoning recipes: https://code.launchpad.net/~ce-orange-squad/+recipe/juju-core-unstable-packaging
[15:12] <marcoceppi> using ppa-release similar to the juju-core releases now
[15:13] <marcoceppi> err, ppa-release == backportpackage
[15:13] <marcoceppi> oh, suddenly this bug makes sense
[15:13] <marcoceppi> sinzui: sorry about that, I see the problem now
[15:14] <sinzui> marcoceppi, oh, I saw the recipe was using quit in debian/source/format is your branch doing the same
[15:14] <marcoceppi> sinzui: this is charmtools 0.3
[15:14] <marcoceppi> 1.x isn't backwards compatible I think IS is building it themselves. I'll update the bug
[15:14] <sinzui> ha ha
[15:16] <marcoceppi> let me go get some caffine
[15:32] <yolanda> hi, i'm working on clustering for rabbit charm, and i found an issue retrieving all the peers for a relation
[15:33] <yolanda> i created 4 clustered nodes, if nodes are up, it works ok and i get a list of the 4 nodes when trying to query peers
[15:33] <yolanda> but then i just stop first unit with nova stop, do the query again, and only the first unit (the one stopped) is reported
[15:37] <marcoceppi> yolanda: that's interesting. When you say query you mean with relation-list or in the juju status
[15:37] <yolanda> marcoceppi, in the relation-list
[15:38] <yolanda> i can reproduce it right now
[15:38] <yolanda> i got logs telling i only had first unit when i stop it
[15:38] <yolanda> i started again and got all the units
[15:39] <yolanda> i'm programming rabbitmq clustering and that's an issue, if first unit is stopped for some reason, i cannot add more units
[15:39] <lazypower> jcastro, did you release your mongodb bundle?
[15:40] <marcoceppi> yolanda: that's odd, I've never seen that before and I don't quite know where to start looking. If it's replicatable I'd file a bug against juju-core about it with instrutions to replicate
[15:40] <yolanda> yes, i can replicate it
[15:40] <yolanda> i'll write some bug
[15:40] <jcastro> lazypower, I was just going to check
[15:40] <jcastro> https://jujucharms.com/fullscreen/search/bundle/~jorge/mongodb-cluster/1/mongodb-cluster/?text=mongodb-cluster
[15:40] <jcastro> Kablammo!
[15:40] <lazypower> Great minds think alike
[15:40] <lazypower> BOOM!
[15:40] <marcoceppi> yolanda: thanks! sorry I couldn't be of more help
[15:40] <jcastro> hey wanna give it a shot?
[15:41] <yolanda> no problem
[15:41]  * marcoceppi pings fwereade to see if he can shed light on yolanda's issue
[15:41] <lazypower> You know i do!
[15:50] <yolanda> files this one: https://bugs.launchpad.net/ubuntu/+source/juju/+bug/1267913
[15:51] <_mup_> Bug #1267913: juju relation-list doesn't report full units list when unit is down <juju (Ubuntu):New> <https://launchpad.net/bugs/1267913>
[15:51] <fwereade> marcoceppi, sorry, was meeting, reading back
[15:52] <fwereade> yolanda, thanks for the bug, that's bizarre and terrifying, I'll take a look
[15:53] <yolanda> thx
[16:26] <marcoceppi> Okay, last call for changes to charm-tools
[16:33] <rick_h_> marcoceppi: I request a pony!
[17:05] <lazypower> Hey jorge
[17:05] <lazypower> with your mongodb-cluster bundle, the relationships incoming is a bit wonky and did not draw on the gui
[17:05] <jcastro> yeah so
[17:05] <jcastro> that happened to me yesterday
[17:05] <lazypower> http://i.imgur.com/QUcHZQg.png
[17:06] <jcastro> https://jujucharms.com/fullscreen/search/bundle/~jorge/mongodb-cluster/1/mongodb-cluster/?text=mongodb-cluster
[17:06] <lazypower> WHen i go to draw the relationship, it shows a dialogue displaying the options for the relationship.
[17:06] <jcastro> they're supposed to be there
[17:07] <jcastro> let me try with a simple wordpress bundle
[17:08] <jcastro> I mean, if you look in the bundle itself, the relationships are there
[17:08] <lazypower> I dont doubt it, i'm still investigating.
[17:08] <lazypower> Just reporting my initial findings
[17:09] <jcastro> yeah, I'm just saying, I don't think it's the bundle itself
[17:10] <lazypower> Ah, i think what happened is the relationships were pending this last unit coming online
[17:11] <lazypower> I love juju, its so smart.
[17:11] <rick_h_> lazypower: jcastro yea, relationships in budles are only applied after all units are up I believe
[17:11] <jcastro> ah
[17:11] <jcastro> but what happens if he had to resolve --retry a bunch of things?
[17:11] <rick_h_> not sure, if the bundle process was still processing it might work, but not tested it tbh.
[17:11] <lazypower> I can wipe this and start over
[17:12] <rick_h_> if something throws an error the bundle deploy quits
[17:13] <jcastro> aha
[17:13] <jcastro> so basically, the race condition in the charm probably causes the bundle to quit
[17:14] <lazypower> Thats my guess, but the other problem i ran into is the lxc container already created bug.
[17:14] <lazypower> thats reported as resolved in .17 is it not?
[17:31] <jcastro> lazypower, I tried it on aws
[17:31] <jcastro> so I don't think it's a provider-specific thing
[17:39] <lazypower> Ah ok
[18:07] <marcoceppi> rick_h_: there's no way to re-run deployer?
[18:07] <rick_h_> wipe everything and try again
[18:08] <marcoceppi> rick_h_: is there anything on the roadmap to report when deployer throws an error to the gui? maybe like a "re-run" since deployer will pick up where it left off
[18:08] <marcoceppi> I meah, obviously fixing the charm is a priority
[18:08] <rick_h_> marcoceppi: right now if it dies in an error it should report whatever Juju tells it
[18:08] <rick_h_> in the Gui notification box
[18:08] <marcoceppi> rick_h_: gotchya
[18:08] <rick_h_> if it does not, then there's an issue we need to fix, but we try to help the user know something went bad if we can
[20:19] <jcastro> aha!
[20:19] <jcastro> marcoceppi, found the issue with the local provider
[20:19] <jcastro> I had to blow away .juju/local too
[20:19] <rick_h_> jcastro: which issue is this?
[20:19] <jcastro> marcoceppi, do you have that list of stuff you need to blow away from thumper? we should write that up
[20:20] <jcastro> rick_h_, my local provider stopped working
[20:23] <marcoceppi> jcastro: yeah
[20:25] <marcoceppi> /etc/init/juju-*; /etc/rsyslog.d/25-juju*; /var/lib/juju/containers/*; /var/lib/lxc/juju-*; ~/.juju/<env>; ~/.juju/environments/<env>.jenv
[20:25] <marcoceppi> jcastro: ^
[20:31] <avoine> also /etc/lxc/auto/ I think
[20:31] <avoine> the symlinks in it
[20:32] <jcastro> http://askubuntu.com/questions/403618/how-do-i-clean-up-a-machine-after-using-the-local-provider/403619#403619
[20:32] <jcastro> avoine, gotcha, adding now
[20:35] <avoine> jam: btw I create a bzr branch for that like you ask: https://code.launchpad.net/~patrick-hetu/golxc/fix-1238541
[21:15] <dpb1> marcoceppi:/jcastro: very simple apache2 fix, it's broken right now if you don't set "servername" and are generating self-signed-certs: https://code.launchpad.net/~davidpbritton/charms/precise/apache2/fix-no-servername-case/+merge/201254
[21:15] <jcastro> ack
[21:15] <jcastro> it's coming up on the charm audit anyway
[21:15] <jcastro> nice catch!
[21:16] <marcoceppi> dpb1: thanks
[21:30] <dpb1> :)
[22:07] <achiang> hello, this is more of a generic cloud question than juju-specific, but are there any cloud providers who give free accounts on micro/tiny instances?
[22:07] <achiang> i thought a micro instance on EC2 might have been free, but it looks like that's only a promo price for 1 year
[22:08] <marcoceppi> achiang: not that I'm aware of. AWS has 700 free hours of micro a month (which equates to about 1 micro a month) but the micro is a pretty poor performance instance size
[22:08] <achiang> marcoceppi: is that forever?
[22:08] <marcoceppi> achiang: for the first yea
[22:09] <achiang> or does that only last for 1 year?
[22:09] <achiang> damn, right...
[22:09] <achiang> hoping for a more long term solution
[22:09] <sarnold> canonistack?
[22:09] <marcoceppi> if you work for Canonical, then that does work
[22:09] <achiang> i'm working on something for the general public
[22:09] <sarnold> ah
[22:10] <achiang> too bad heroku isn't a generic OS
[22:12] <achiang> maybe i should play with the vagrant images after all
[22:12] <achiang> but that's slightly more annoying
[22:16] <marcoceppi> it's certainly the cheapest
[22:17] <achiang> going down the heroku path... i could package up my stuff and put it in pip maybe
[22:17] <achiang> or pypi, rather
[22:18] <achiang> and then could specify my package in requirements.txt
[22:18] <achiang> ugh
[22:18] <achiang> but then i have to make a whole django app
[22:18]  * achiang keeps searching for a free lunch somewhere
[22:19] <achiang> can i conceptually think of juju+vagrant as a "cloud environment on your desktop" ?
[22:26] <achiang> who owns content here - https://juju.ubuntu.com/docs/config-vagrant.html
[22:26] <achiang> the vagrant download link is old
[22:26] <achiang> should be this link: http://www.vagrantup.com/downloads.html