[10:35] <pmitros> Is there good text on debugging juju anywhere? I've gone through the Hello World example a few times (wordpress+mysql), and it's never worked. On one machine, it would break while apt was installing packages (specifically, the postfix package was waiting for user input). On my home maschine, I get a bad gateway error from nginx. Some kind of high-level roadmap would be very helpful ("Here are where log files live," "Here's how you downloa
[14:42] <marcoceppi> jcastro: can you delete this? https://code.launchpad.net/~jorge/charms/precise/docs/trunk It breaks `charm list`
[14:45] <jcastro> done
[15:28] <rshade98> I am still having a problem with TLS failure. Any ideas?
[15:50] <hazmat> marcoceppi, see jitsu search
[15:50] <hazmat> rshade98, pastebin?
[15:51] <hazmat> pmitros, which version?
[15:51] <hazmat> pmitros, ie. juju-core or pyjuju
[15:52] <marcoceppi> hazmat: I see there's a search, but no "list" function, any way to list "reviewed" charms outside of charm list?
[15:53] <marcoceppi> Trying to replace `charm list | grep lp:charms`
[16:03] <hazmat> marcoceppi, search -h
[16:03] <hazmat> marcoceppi, owner:~charmers is cheeky but should hit the majority
[16:04] <hazmat> er. owners:charmers
[16:04] <marcoceppi> hazmat: http://paste.ubuntu.com/5619558/ ;)
[16:07] <hazmat> marcoceppi, ugh.. the help in jitsu is broken.. try jitsu help search
[16:07] <marcoceppi> hazmat: Ah, thanks that works much better
[16:08] <hazmat> marcoceppi, its going to break in a week or two when jujucharms.com gets switch over to its new backend, easy enough to fix though
[16:08] <marcoceppi> hazmat: ack
[16:48] <rshade98> hazmat
[16:48] <rshade98> http://paste.ubuntu.com/5619673/
[16:51] <hazmat> rshade98, it looks like you have a cert mismatch, the ca is cached on the client in ~/.juju but the ca cert on the client doesn't match up with the bootstrap'd state server
[16:52] <rshade98> Is there a way to only bootstrap the certs
[16:52] <hazmat> rshade98, not without a destroy-environment and rebootstrap
[16:52] <rshade98> I get a failure with multiple boxes, saying the environment is already bootstraps
[16:53] <hazmat> rshade98, did you switch machines?
[16:53] <hazmat> rshade98, which ever machine you did the original bootstrap on. you have to copy the whole ~/.juju over
[16:53] <rshade98> oh, so it is gone
[16:53] <hazmat> the ca cert file that works, is in the one that bootstrapped
[16:53] <rshade98> I need to destroy the whole environment huh
[16:54] <hazmat> rshade98, in that case destroy && bootstrap is the simplest option
[16:54] <rshade98> is that just deleting the bucket in ec2?
[16:54] <hazmat> alternatively manually creating a new ca, new cert, and installing it onto the bootstrap server
[16:54] <rshade98> I mean s3/ec2
[16:54] <hazmat> rshade98, destroy-environment kills all the machines in the env
[16:56] <rshade98> ok, so that just deletes the bucket etc.
[17:02] <rshade98> now it won't dial the mongos should upload the certs, kill the machine, add the juju groups to it, and start it
[17:41] <arosales> marcoceppi, https://jenkins.qa.ubuntu.com/view/Precise/view/Precise%20Charms/ looking a lot better. Thanks for getting beating that into shape :-)
[17:44] <arosales> s/getting/
[18:50] <hazmat> marcoceppi, you coming out to oakland?
[18:51] <marcoceppi> hazmat: yes
[18:51] <hazmat> cool
[18:55] <hazmat> marcoceppi, it looks like failures generate false success in some cases
[18:55] <marcoceppi> hazmat: have an example?
[18:56] <hazmat> marcoceppi, i thought i did.. but explicitly navigating to the last build looks okay.. i was looking at the wordpress graph test
[18:56] <hazmat> which still fails with memcached.. but the test is passing
[18:56] <hazmat> but the it does look significantly better
[18:57] <marcoceppi> hazmat: yeah, there's still a few broken tests out there. Some corner cases like juju environment not getting set up properly cause the test to pass
[18:58] <sarnold> ooh, I like that jenkins overview of charms :) ec2 and local, very neat
[18:58] <hazmat> marcoceppi, the new juju-deployer version is pretty robust and helpful about doing the proper wait for success with streaming change output (jcore only though)
[18:59] <hazmat> might be able to feed it the test plans, or massage the test plans into a compatible form
[18:59] <marcoceppi> hazmat: juju-developer? You mean the API stuff?
[19:01] <hazmat> marcoceppi, its a tool for deploying complex environments from a yaml/json description.. with inheritance, file value substitution, etc.
[19:01] <hazmat> lp:juju-deployer
[19:01] <hazmat> we use it quite a bit for openstack
[19:02] <marcoceppi> hazmat: that sounds super sexy. Shouldn't be difficult at all to format the test plans to feed that in
[19:13] <rshade98> is juju a per s3 folder environment?
[19:14] <rshade98> it seems to alway have conflicts when they share.
[19:14] <rshade98> and if so, any idea what the best way to randomize the uuid is
[20:23] <_mup_> Bug #1174905 was filed: Not possible to deploy local charms <juju:New> <https://launchpad.net/bugs/1174905>
[20:40] <mwhudson> mramm: hey
[20:40] <mramm> hey
[20:40] <mwhudson> mramm: <generic armhf juju ping goes here>
[20:41] <mramm> I've filed a ticket to get the go 1.1 beta into the archive so that we can start building official images with it
[20:41] <mramm> if you want help trying to compile yourself davecheney (dfc) can help (I CCed him on our last e-mail exchange)
[20:42] <mwhudson> ok
[20:42] <mwhudson> compile go myself or compile juju?
[21:07] <rshade98> is there any way to make the juju commands super verbose
[21:34] <sidnei> rshade98: as in -vvv?
[21:35] <sidnei> rshade98: though you might be actually looking for juju debug-log which tails the services' logs