[00:31] <m_3> crazy how many times I write config-git instead of config-get... need a new vim alias
[00:51] <SpamapS> btw, charm-tools is in precise now, woot. :)
[01:04] <m_3> negronjl: ping
[01:05] <negronjl> m_3: pong
[01:05] <m_3> hey man
[01:05] <negronjl> m_3: what's happening
[01:05] <m_3> ok, so I finally got a bunch of node apps in front of mongo
[01:06] <m_3> adding extra mongo nodes winds up with _everything_ bound to the last mongo node that comes up
[01:06] <negronjl> m_3: hmm ...
[01:06] <m_3> I guess it's re-firing relation hooks... which it should
[01:07] <m_3> but how can you tell that the last node up isn't what you wanna bind to?
[01:07] <negronjl> m_3:  your app should connect to the output of relation-list
[01:07] <negronjl> m_3: not the output of relation-get <>
[01:07] <m_3> replset isn't enough info along
[01:07] <m_3> alone
[01:07] <m_3> ok
[01:07] <m_3> so pull the whole set at once
[01:08] <negronjl> m_3:  the mongo driver takes care of the rest ... just pass all of the mongodb nodes to your app
[01:08] <negronjl> m_3:  which app did you use ( a charm ? ) if so, which one so I can re-create
[01:08] <m_3> but it's gotta put it into facter or something to get all the ip addrs... otherwise I just get names
[01:08] <negronjl> m_3: names is fine
[01:08] <m_3> sorry, 'relation' names, not dns
[01:09] <m_3> lp:~mark-mims/charm/oneiric/node/trunk
[01:09] <negronjl> m_3:  can you email me the description of your test so I can re-create and try to fix what I can ?
[01:09] <m_3> it's just hitting one db instance at a time... I can refactor that
[01:09] <m_3> but it'll prob need something like facter
[01:09] <m_3> no biggie
[01:09] <negronjl> m_3:  let me try and recreate and I'll report back.
[01:09] <m_3> what are you using to test the frontend?
[01:09] <negronjl> javascript
[01:09] <negronjl> m_3: ^^
[01:10] <m_3> a charm? or just an external app?
[01:10] <negronjl> m_3: manual testing using javascript ( mongo has a javascript interface )
[01:10] <negronjl> m_3:  Let me create a simple web-app that is DB intensive and I'll get back to you to test and such
[01:10] <m_3> gotcha... ok, well I'll just plan on refactoring the app
[01:11] <negronjl> m_3:  cool .. the more tests ( and diff. kinds of test ) the better.
[01:11] <m_3> I didn't get the replicated data when hitting the last node joining though... but there're too many things that went wrong with the test to say so reliably :)
[01:12] <negronjl> m_3:  no worries ... I'll try and make a better test
[01:12] <m_3> there's an outdated blog post about the app itself http://markmims.com/cloud/2011/09/07/node-mongo-ubuntu.html
[01:12] <m_3> might explain the app quickly
[01:12] <negronjl> m_3: I am thinking of combining a test of mapreduce and take the output and put it into mongo or, do a mapreduce job within mongo itself to test.
[01:12] <m_3> simple though
[01:12] <m_3> ha
[01:13] <m_3> cool, yeah two birds with one stone
[01:13] <negronjl> m_3:  exactly :)
[01:13] <m_3> ok, maybe half dozen birds with one stone
[01:13] <negronjl> m_3: I hope :)
[01:13] <negronjl> m_3:  gotta jet out for a bit.  email me if you have anything else
[01:13] <m_3> later man... emailing stack script now
[01:13] <negronjl> m_3:  cool .. thx
[01:35] <negronjl> m_3: ping
[01:53] <m_3> yo
[01:53] <m_3> negronjl: ^
[01:59]  * m_3 dinner bell
[05:33] <george_e> I'm reading this page: https://juju.ubuntu.com/docs/getting-started.html
[05:33] <george_e> ...and it mentions something about control-bucket and admin-secret in environments.yaml.
[05:33] <george_e> What are those two fields and what do they mean?
[18:39]  * SpamapS is on his way to give a juju presentation to the ventura county linux users group :)
[18:39] <marcoceppi> SpamapS: good luck!
[22:54] <george_e> In environments.yaml, if the key 'default' is missing, which deploy environment is chosen as the default? The first?
[23:09] <marcoceppi> george_e: I believe so
[23:09] <marcoceppi> Either that or it throws an error
[23:10] <george_e> Ah, okay.
[23:11] <marcoceppi> george_e: http://paste.ubuntu.com/758722/
[23:11] <george_e> Ah, thanks.