[01:58] <designated> can someone point me in the right direction for information on how to manage multiple NICs for an openstack deployment using juju?  What needs to be modified in the charm in order to designate NICs to different networks? By default it seems to use a single NIC, is this incorrect?
[13:28] <themonk> i am tring to remove a unit from lxc using $juju remove-unit command
[13:29] <themonk> juju status showing its dying
[13:29] <marcoceppi_> themonk: is the service in an errored state?
[13:29] <marcoceppi_> themonk: ie, the unit that machine is attached to?
[13:29] <themonk> but it is taking like forever
[13:29] <themonk> marcoceppi_, yes it is in error state
[13:31] <themonk> marcoceppi_, the install script failed
[13:31] <marcoceppi_> themonk: that's why it's not going away. Juju blocks future events from occuring when a unit is in an error state. You can either force remove the unit (not always recommended), or just run juju resolved <unit/#> to mark it as resolved and have it finish it's events, the last event queued being the "remove unit" command
[13:34] <themonk> marcoceppi_, i am back
[13:35] <marcoceppi_> themonk: that's why it's not going away. Juju blocks future events from occuring when a unit is in an error state. You can either force remove the unit (not always recommended), or just run juju resolved <unit/#> to mark it as resolved and have it finish it's events, the last event queued being the "remove unit" command
[13:36] <themonk> marcoceppi_, how to force remove it?
[13:37] <themonk> marcoceppi_, though i will try juju resolved first
[13:37] <marcoceppi_> juju terminate-machine --force #; where # is the machine ID
[13:37] <themonk> marcoceppi_, thanks :)
[14:52] <themonk> marcoceppi, is there any way to restart a charm on previously error state unit? without removing it by force
[14:55] <themonk> is there any way to restart a charm on previously error state unit? without removing it by force
[14:58] <themonk> found it juju resolve --retry=true unit-name
[14:59] <marcoceppi> themonk: yeah, just juju resolved --retry unit will work too, no need to explicitly state =true
[15:01] <themonk> marcoceppi, i updated py code in local repository then $juju resolve --retry unit-name will this code change take place automatically
[15:02] <marcoceppi> themonk: no, you have to run upgrade charm
[15:40] <themonk> marcoceppi, after upgrading charm code it got new revision number but after resolve retry new code is not inside charm unit i checked in /var/lib/juju/agents/unit-oxrp-0/charm/hooks/hooks.py
[15:46] <themonk> marcoceppi, is it a bug or i am doing something wrong
[15:48] <bloodearnest> arosales: hangouts wouldn't let me into the cross-team meeting, btw, was full! :)
[16:01] <jcastro> lazyPower, wanna document local/vagrant this afternoon?
[16:01] <lazyPower> jcastro: I've already got a screencast recorded
[16:01] <jcastro> \o/
[16:01] <lazyPower> i need to do the voiceover, the slides, and cut
[16:01] <lazyPower> but i'm about 20% there
[16:02] <lazyPower> i'm leaving in about an hour and a half to head downtown to a meetup with some of the pittsburgh cloud peoples
[16:02] <lazyPower> so this afternoon is not good for me. can we postpone until tomorrow?
[16:02] <jcastro> works for me
[16:03] <jcastro> hey evilnickveitch
[16:03] <jcastro> adding bundle review criteria to this page:
[16:04] <jcastro> https://juju.ubuntu.com/docs/authors-charm-policy.html
[16:04] <jcastro> might make it too big
[16:04] <jcastro> should I split it into a new page?
[16:04] <jcastro> I was thinking under Reference instead
[16:04] <jcastro> something like "Review Criteria"
[16:05] <jcastro> sorry wrong URL, the review stuff I want to split from is here: https://juju.ubuntu.com/docs/authors-charm-store.html#submitting
[16:07] <jcastro> holy smokes review queue!
[16:08] <noodles775> mgz: I still see the same issue - http://paste.ubuntu.com/7005758/ (added paste to bug)
[16:14] <jcastro> marcoceppi, figured out the git thing: https://help.github.com/articles/syncing-a-fork
[16:16] <rick_h_> jcastro: need a https://github.com/juju/juju-gui/blob/develop/HACKING.rst#typical-github-workflow :)
[16:18] <jcastro> ohhh, I am stealing this
[16:21] <lazyPower> #TIL there is an autosquash flag
[16:21] <lazyPower> I've been doing that manually for the longest time
[16:23] <jcastro> yeah we are stealing this
[16:24] <lazyPower> also i just moved the datastax opscenter charm to incomplete. since i'm reworking the cassandra charm - it doesn't need to be reviewed right now. If you attempt to deploy at scale its going to turn into chunky salsa.
[16:24] <lazyPower> thats 1 off the rev queue once it re-syncs
[16:30] <jcastro> hah
[16:31] <jcastro> marcoceppi, ok I landed the doc workflow in the README, lmk if it makes sense to you
[16:42] <lazyPower> Question for anyone thats listening. WHen generating self signed certificates - and using the gui - would you rather see a config box for each option in the cert? (country, state, company name, etc.) or a single line with options in key/value pairing, separated by pipes?
[16:42] <lazyPower> C=US|ST=PA|O=FooCorp
[16:47] <marcoceppi> jcastro: you mean you landed without a review
[16:47] <marcoceppi> lazyPower: both?
[16:47] <jcastro> that was an accident
[16:47] <jcastro> marcoceppi, but I figured it out now
[16:47] <marcoceppi> ;)
[16:48] <lazyPower> marcoceppi: i meant a box per option or a single box
[16:48] <marcoceppi> lazyPower: yeah, why not both? Advanced mode or easy mode
[16:49] <marcoceppi> otherwise, go with what makes sense for the average user
[16:53] <jcastro> marcoceppi, OK, NOW I think I did it right
[16:53] <jcastro> marcoceppi, I was missing the workflow bits rick_h_ showed me, so now we can do feature branches in our own space and then gate trunk
[16:53] <jcastro> that should make evilnickveitch happy
[16:53] <jcastro> let's not mention my accidental cowboy commits though. :p
[16:54] <marcoceppi> jcastro: yup, as it should be
[16:55] <jcastro> fixed my email address on github too, so now I look more respectable: https://github.com/juju/docs/graphs/contributors
[16:55] <marcoceppi> there ya go
[16:55] <marcoceppi> much more reflective
[16:55] <marcoceppi> such contribute
[16:56] <jcastro> much cowboy
[16:56] <evilnickveitch> jcastro, few things are making me happy today, so well done
[16:57] <jcastro> marcoceppi, so I just dogfooded what I wrote and it all works, the only areas I am grey on are the aliases at the bottom
[16:57] <marcoceppi> aliases?
[16:57] <jcastro> marcoceppi, rick's team has some convenience aliases I kept, they look useful
[16:57] <marcoceppi> where?
[16:57] <jcastro> the bottom of the README
[16:58] <marcoceppi> jcastro: ah, I'd suggest making juju-docs-upsteam https://github.com/juju/docs.git instead
[16:59] <marcoceppi> since it'll prompt you for user/pass when you try to push which is a good indicator you're doing something wrong
[16:59] <jcastro> ok I'll polish that later
[16:59] <jcastro> right now I only needed to know how to do this to finish my card, heh
[17:00] <jcastro> evilnickveitch, so wrt a reviewing section under References, is that ok?
[17:00] <marcoceppi> TheMue: I don't think we should have $ in the docs, we made a concious decision that anything wrapped in <code> is to be considered a command, and anything else would use syntax highlighting to suggest otherwise
[17:01] <marcoceppi> $ make copy and pasting more difficult and if we really wanted the $ I'd rather use CSS to stylize command prompt sections with a secial class to prepend the $ then to have it straight in the docs
[17:01] <marcoceppi> and, I'll just put this in the bug report
[17:03] <jcastro> marcoceppi, ok before I start over on a new feature branch I need to make sure I pull from master again correct?
[17:03] <marcoceppi> jcastro: yeah, so basically: git checkout master, git fetch upstream, git merge upstream/master; git checkout -b <feature>
[17:03] <marcoceppi> though, you could skip the fetch/merge and just do git pull upstream master
[17:04] <jcastro> that just wants to merge it into my feature branch though
[17:04] <marcoceppi> jcastro: git checkout master
[17:04] <marcoceppi> the first command puts you back on master
[17:04] <jcastro> aha!, got it
[17:06] <TheMue> marcoceppi: ok, have to change it back then. but found other docs with $, so I harmonized it
[17:07] <marcoceppi> TheMue: which? those should be un$signed?
[17:08] <fginther> anyone see this when bootstrapping to hp cloud? ERROR bootstrap failed: cannot start bootstrap instance: index file has no data for cloud {region-a.geo-1 https://region-a.geo-1.identity.hpcloudsvc.com:35357/v2.0/} not found
[17:08] <marcoceppi> fginther: can you try to bootstrap again with the --debug and --show-log command options?
[17:09] <fginther> marcoceppi, running
[17:10] <TheMue> marcoceppi: https://juju.ubuntu.com/docs/authors-hook-environment.html, first code of relation-get
[17:10] <TheMue> marcoceppi: and it made sense to me for a better separation of what the command is and what the output
[17:12] <fginther> marcoceppi, are there any secrets in that data? (I don't see any with a simple scan)
[17:12] <marcoceppi> TheMue: I can create a special HTML class which will place the $ without affecting copy and paste where it's needed to help distingush command from output
[17:12] <marcoceppi> fginther: the very very first line might provide some private data, everything else should be safe (the giant json blog)
[17:12] <marcoceppi> blob*
[17:13] <TheMue> marcoceppi: but it would place the $ before every line, wouldn't it?
[17:13] <fginther> marcoceppi, http://paste.ubuntu.com/7006105/
[17:13] <marcoceppi> TheMue: not nessiarily
[17:14] <marcoceppi> TheMue: it would be a <span> class you could use inside the <code> block
[17:14] <TheMue> marcoceppi: ah, yes, cool idea. ugly in writing, but the output is important
[17:14] <marcoceppi> TheMue: right
[17:15] <marcoceppi> lmk if that interests you and I'll make a card for it
[17:15] <TheMue> o/ <= see my hand waving
[17:17] <marcoceppi> fginther: have you ever successfully bootstrapped on HP before?
[17:17] <fginther> marcoceppi, nope
[17:17] <fginther> marcoceppi, I'm using 1.17.3-0ubuntu1~ubuntu13.10.1~juju1 on saucy
[17:17] <fginther> marcoceppi, and 1.17.3-0ubuntu1 on trusty
[17:18] <marcoceppi> fginther: do you have an image-metadata key defined in your environments.yaml?
[17:19] <fginther> marcoceppi, no, I only have admin-secret, control-bucket, auth-url, default-series, use-default-secgroup and firewall-mode
[17:20] <fginther> marcoceppi, OS_TENANT_ID, OS_TENANT_NAME, OS_REGION_NAME, OS_USERNAME and OS_PASSWORD are in the env
[17:20] <fginther> oh and OS_AUTH_URL
[17:21] <marcoceppi> fginther: run juju destroy-environment hp-cloud, rm -f ~/.juju/environments/hp-cloud.jenv, change the control-bucket name to something else, then try to bootstrap again
[17:22] <bbcmicrocomputer> so I noticed in Juju 1.17, .juju/ssh/id_rsa.* exists... is Juju using it's own generated keys now?
[17:24] <marcoceppi> bbcmicrocomputer: I guess so?
[17:27] <fginther> marcoceppi, looks like the same error http://paste.ubuntu.com/7006173/. I need to break for lunch, bbiab if you can still assist
[17:27] <fginther> marcoceppi, and thanks for the help
[17:28] <marcoceppi> fginther: sounds good
[18:05] <jhobbs> when I run sync-tools, I'm only getting tools for amd64 and i386; how can I include armhf tools?
[18:14] <mgz> jhobbs: good question. I don't see any obvious references to arch in the code
[18:14] <mgz> but streams.canonical.com does have armhf for the latest versions
[18:15] <arosales> utlemming: fyi joyent provider is hitting https://bugs.launchpad.net/juju-core/+bug/1285803
[18:15] <_mup_> Bug #1285803: [Joyent Provider] metadata mismatch when testing again Joyent Public Cloud <joyent-provider> <test-failure> <juju-core:Triaged> <https://launchpad.net/bugs/1285803>
[18:15] <arosales> utlemming: could you confirm this is not image or stream related
[18:17] <jhobbs> mgz: i found the bug
[18:17] <jhobbs> https://bugs.launchpad.net/juju-core/+bug/1285410
[18:17] <_mup_> Bug #1285410: juju names arm arch 'arm' internally, but 'armhf' in tools <armhf> <constraints> <juju-core:Triaged> <https://launchpad.net/bugs/1285410>
[18:18] <jhobbs> mgz: juju's default filter for tool is amd64/i386/arm
[18:18] <jhobbs> mgz: so armhf doesn't match
[18:18] <jhobbs> mgz: hardcoded too so there is no way to override it from sync-tools..
[18:21] <mgz> jhobbs: well, isn't that fun.
[18:23] <mgz> you can hack a local copy of juju to have the right name in the arch list, run sync-tools with that, will that unstick you?
[18:27] <jhobbs> mgz: yeah i think so - that's what i'm doing now; thanks
[18:36] <fginther> marcoceppi, I was able to find the answer to my problem. I needed to add the image-metadata-url and tools-metadata-url for our project specific data
[18:36] <marcoceppi> fginther: ah, so you're using something other than what we provide in our streams?
[18:37] <fginther> marcoceppi, this appears to be the issue: https://bugs.launchpad.net/juju-core/+bug/1275280
[18:37] <_mup_> Bug #1275280: HP Cloud needs image metadata for newer accounts <juju-core:Invalid> <https://launchpad.net/bugs/1275280>