[02:40] <_mup_> Bug #903014 was filed: relation-broken runs in inappropriate context <juju:In Progress by fwereade> < https://launchpad.net/bugs/903014 >
[02:43] <_mup_> Bug #903016 was filed: relation-broken hooks will not always be fired <juju:In Progress by fwereade> < https://launchpad.net/bugs/903016 >
[02:45] <_mup_> Bug #903017 was filed: unit workflows stop/start inappropriately <juju:In Progress by fwereade> < https://launchpad.net/bugs/903017 >
[02:47] <_mup_> Bug #903018 was filed: charm upgrade is dangerous <juju:In Progress by fwereade> < https://launchpad.net/bugs/903018 >
[04:04] <EvilBill> I keep seeing that the juju client is in macports - yet it's not listed on the macports site, and a "port search juju" comes up empty. Am I missing something?
[04:33] <EvilBill> I found something on it… https://trac.macports.org/ticket/30237
[04:35] <EvilBill> Not directly installable though.
[06:54] <SpamapS> EvilBill: its pure python, so once you have libzookeeper, I believe you can just use setup.py to install it
[06:55] <EvilBill> SpamapS: Hm, I'll try that.
[06:55] <EvilBill> grabbing libzookeeper from macports now
[06:55] <SpamapS> EvilBill: actually you'd also need txzookeeper too
[06:55] <EvilBill> I'll try that :)
[06:55]  * SpamapS thinks at some point we need to have releases to coordinate these ports/targets
[06:56] <EvilBill> I agree.
[06:56] <EvilBill> things need to coalesce a bit.
[06:56] <EvilBill> documentation too.
[06:57] <EvilBill> blah, txzookeeper isn't in macports, and libzookeeper failed to build.
[06:57] <EvilBill> Will look at this later, then
[06:59] <SpamapS> EvilBill: there's also virtualbox. :)
[06:59] <EvilBill> True :)
[06:59]  * SpamapS has been native on his macs for over a year now. :)
[06:59] <EvilBill> I've got other linux machines around here too.
[06:59] <SpamapS> I still flip back to OS X to update the OS on my iphone tho. :-P
[06:59] <EvilBill> regular lab environment.
[07:00] <EvilBill> But I use a Mac as my "desktop" environment to protect myself from myslef.
[07:00] <EvilBill> If I am on a linux desktop, I tend to try out new stuff constantly
[07:00] <EvilBill> in search of the newest shiny
[07:00] <EvilBill> and eventually I break it.....
[07:00] <EvilBill> and then my main desktop environment is screwed up, and I spend time not working to FIX it.
[07:02] <SpamapS> yeah I used to have that running Debian sid on my workstation.. my laptop was always the more stable box. :-P
[07:03] <SpamapS> Now I have my smaller laptop (MacBookAir 11") on Ubuntu 11.10, and my dev laptop (MBP 15") is precise.. works pretty well.
[07:03] <EvilBill> Yep.
[07:03] <EvilBill> Nice.
[07:03] <SpamapS> the Air, running Unity 2D, gets 5.5hrs battery life.. just barely less than it did in OS X (about 6hrs)
[07:03] <EvilBill> NICE.
[07:04] <EvilBill> I need to setup a ubuntu machine with a gui again.
[07:04] <EvilBill> I have a bunch of server-type stuff
[07:05] <SpamapS> marcoceppi: btw, thanks for the updates, merged
[07:05] <SpamapS> EvilBill: I'm kind of addicted to apple's hardware. Having an apple store 1 block away from the house is really 90% of the problem. ;)
[07:06] <EvilBill> SpamapS: I'd be SO broke if that happened.
[07:07] <SpamapS> Yeah like, when it came time to buy a new, smaller laptop.. I could have probably gotten a much better deal on a Lenovo .. but the air is *so* pretty.. and *so* thin.. and I *so* just got to walk out of the store with it. ;)
[07:12] <EvilBill> Yep.
[07:12] <EvilBill> I have way too much apple gear too.
[07:12] <EvilBill> Hell, I have a CUBE.
[09:10] <niemeyer> Good mornings!
[09:35] <koolhead11> hi all
[09:40] <niemeyer> koolhead11: Hey there
[09:41] <koolhead11> niemeyer: hello.
[09:41] <koolhead11> i upgraded my juju from PPA
[09:41] <koolhead11> sudo add-apt-repository ppa:juju/pkgs && sudo apt-get update && sudo apt-get upgrade juju
[09:41] <koolhead11> but i can still not see the details error log
[09:41] <koolhead11> *detailed
[09:43]  * koolhead11 is yet not able 2 deploy the simplest charm he wrote :(
[09:45] <niemeyer> koolhead11: Sorry, I'm out of context.. why isn't it working?
[09:46] <koolhead11> niemeyer: https://code.launchpad.net/~koolhead17/charm/oneiric/boa/trunk
[09:46] <koolhead11> i am trying to execute it vila LXC
[09:46] <niemeyer> koolhead11: You mean via the local provider?
[09:46] <koolhead11> niemeyer: yes
[09:47] <koolhead11> niemeyer: also http://paste.ubuntu.com/767706/
[09:47] <koolhead11> why is it searching 4 example inside oneiric :(
[09:51] <niemeyer> koolhead11: Because that's the default Ubuntu series.. you've provided none, so it picked the default oe
[09:51] <niemeyer> one
[09:52] <niemeyer> koolhead11: local:foo is the same as local:oneiric/foo when "oneiric" is the default
[09:52] <koolhead11> niemeyer: i have directory structure like this /home/atul/example/oneiric/charm
[09:53] <koolhead11> so when i say juju deploy --repository=example local:boa
[09:53] <niemeyer> koolhead11: So it looks like you're passing the wrong path in the command line
[09:54] <koolhead11> it means /home/atul/example/oneiric/boa
[09:54] <niemeyer> koolhead11: --repository=/home/atul/example is probably what you want
[09:54] <koolhead11> niemeyer: ok let me try that way
[09:58] <koolhead11> niemeyer: no luck . :( http://paste.ubuntu.com/767711/
[09:59] <kickinz1_> quit
[09:59] <koolhead11> kickinz1_: /quit will work :D
[10:02] <niemeyer> koolhead11: Can you please paste "cat ~/example/oneiric/boa/metadata.yaml"?
[10:02] <koolhead11> sure
[10:03] <koolhead11> niemeyer: http://paste.ubuntu.com/767715/
[10:03] <niemeyer> fwereade: Morning
[10:03] <fwereade> heya niemeyer
[10:03] <fwereade> niemeyer, nice weekend?
[10:04] <niemeyer> fwereade: Sorry for being slow on the stop stuff.. I've been thinking about it, and would actually like to talk to you today at some point
[10:04] <niemeyer> fwereade: It was ok.. was traveling back home from SF
[10:04] <fwereade> niemeyer, don't worry about it, it emerges that there's enough subtlety to the general restart problem that I've been quite occupied already ;)
[10:05] <fwereade> niemeyer, sounds like a fun conference
[10:05] <niemeyer> fwereade: It was really nice
[10:10] <fwereade> niemeyer, and ofc I'd be happy to talk about it whenever makes sense for you, but be warned I'm a bit slow/stupid today, I was up late massaging the restart-transitions branch into digestible chunks
[10:12] <niemeyer> fwereade: Ah, same thing here.. I woke up at 6AM, which is around noon in the timezone I was a couple of days ago, so my body has no idea about what's happening whatsoever.
[10:13] <fwereade> niemeyer, heh :)
[10:14] <niemeyer> koolhead11: That's quite weird..
[10:14] <niemeyer> koolhead11: It looks fine
[10:14] <koolhead11> niemeyer: exactly :D
[10:14] <niemeyer> koolhead11: What does "dpkg -l juju" say?
[10:14] <koolhead11> 1 sec
[10:15] <koolhead11> niemeyer: ii  juju           0.5+bzr427-1ju next generation service orchestration system
[10:16] <koolhead11> am using Oneiric as local LXC for my work
[10:16] <koolhead11> interesting thing is the same directory i have kept mysql charm and it runs without any error
[10:19] <niemeyer> koolhead11: I think you're hitting a bug in the charm detection logic, but I can't quite spot where yet
[10:20] <koolhead11> niemeyer: would you suggest me to ask a question under launchpad section for the anser?
[10:20] <koolhead11> *answer
[10:21] <koolhead11> also was wondering do i need to install any pkg on barebone machine to get DEBIAN_FRONTEND=noninteractive working?
[10:24] <niemeyer> koolhead11: No, that's just a hint for the packaging infrastructure to not ask any questions
[10:24] <niemeyer> koolhead11: Can you please run "which juju", just in case?
[10:24] <koolhead11> niemeyer: /usr/bin/juju
[10:25] <niemeyer> Ok
[10:25] <niemeyer> koolhead11: I guess we'll have to debug it..
[10:26] <niemeyer> koolhead11: Hmm
[10:26] <niemeyer> koolhead11: I have a guess, actually
[10:26] <koolhead11> niemeyer: you want me to pastebin all steps and infos so it becomes easy to reproduce?
[10:27] <koolhead11> and on other side my  default mysql charm is working prefectly
[10:27] <niemeyer> koolhead11: Can you please add this to your metadata.yaml:
[10:27] <niemeyer> revision: 0
[10:28] <niemeyer> koolhead11: Right below "name:" (not that it makes a difference.. we just put it there usually)
[10:28] <koolhead11> niemeyer: ok. BTW i created a blank revision file inside that boa charm
[10:28] <koolhead11> am adding as u suggested
[10:31] <koolhead11> niemeyer: http://paste.ubuntu.com/767737/
[10:32] <niemeyer> koolhead11: Cool, at least we know it's looking at it
[10:33] <koolhead11> niemeyer: am confused earlier it said sumthing else in error
[10:35] <niemeyer> koolhead11: let's try this: please open a python prompt
[10:35] <niemeyer> koolhead11: and type this:
[10:35] <niemeyer> from juju.charm.provider import get_charm_from_path
[10:36] <koolhead11> k
[10:36] <niemeyer> koolhead11: Now run:
[10:36] <koolhead11> hmm
[10:37] <niemeyer> d = get_charm_from_path("/home/atul/example/oneiric/boa")
[10:37] <niemeyer> print d
[10:39] <koolhead11> niemeyer: http://paste.ubuntu.com/767746/
[10:39] <koolhead11> i have a empty file with name revision created there as well
[10:40] <niemeyer> koolhead11: Ok, we got it
[10:42] <niemeyer> koolhead11: Can you please file a bug pointing out that this error is being eaten up silently?
[10:42] <niemeyer> koolhead11: The revision file should have an int in it
[10:42] <niemeyer> koolhead11: You can use 0 there
[10:43] <koolhead11> niemeyer: in the file simply write 0 ?
[10:43] <niemeyer> koolhead11: Yep
[10:44] <koolhead11> niemeyer: and what will be the bug description i will file?
[10:44] <koolhead11> empty revision file results in this
[10:46] <koolhead11> niemeyer: your so damm right. i can execute the charm for Boa
[10:46] <koolhead11> :D
[10:47] <niemeyer> koolhead11: Sweet
[10:49] <koolhead11> SpamapS: hazmat yay!! my simplest charm is working !! :D
[10:50]  * fwereade cheers
[10:51] <koolhead11> fwereade: :D
[10:52] <koolhead11> niemeyer: i am not sure what i should write in bug description. Its just my foolishness of not mentioning anything in revision file i suppose :D
[11:02] <niemeyer> koolhead11: No, it wasn't your fault at all
[11:02] <niemeyer> koolhead11: It can't silently say that the charm isn't found when it's clear that there's a charm in that location
[11:03] <niemeyer> koolhead11: As a hint, the summary may be "juju fails silently with empty revision file"
[11:03] <koolhead11> niemeyer: ok doing right away.
[11:03] <niemeyer> koolhead11: Them copy both the pasted where you run the command and get a "no charm" error, and then python prompt dump
[11:03] <niemeyer> s/and then/and the
[11:03] <koolhead11> k
[11:04] <niemeyer> koolhead11: Thanks a lot
[11:04] <koolhead11> niemeyer: thanks 2 you as well :D
[11:17] <_mup_> Bug #903149 was filed: juju fails silently with empty revision file. <juju> <juju:New> < https://launchpad.net/bugs/903149 >
[11:32] <TheMue> Oh, just found out how nicely Emacs integrates with Bazaar.
[11:33] <H3llGhost> Hello
[12:06] <TheMue> niemeyer: Just comparing EC2 in Py and Go. What does var _ juju.Environ = (*environ)(nil) do? AFAIK I would say it create an unnamed nil reference to an environment. But why is that needed?
[12:09] <niemeyer> TheMue: http://golang.org/doc/go_faq.html#guarantee_satisfies_interface
[12:11] <TheMue> niemeyer: thx, ic. never checked it this way, only later when using an interface implementation.
[12:15] <jrgifford> EvilBill: you have a CUBE?! whoa.
[12:15]  * jrgifford </end-offtopic-stuff>
[12:23] <TheMue> niemeyer: Could you shortly describe the state of the Go port compared to the Py code? I'm currently comparing the EC2 providers (trying to understand the bootstraping).
[12:24] <niemeyer> TheMue: It's in its early infancy at the moment
[12:24] <niemeyer> TheMue: We're just catching up some momentum in the last couple of weeks
[12:25] <TheMue> niemeyer: How's the roadmap of the Go port? Do we first provide some supporting tools do we start from ground e.g. with initializing and bootstrap?
[12:26] <niemeyer> TheMue: Yeah, we have to get it to the point of initializing and bootstrapping
[12:26] <niemeyer> TheMue: After that it's a lot easier to fork off development
[12:26] <TheMue> niemeyer: Btw, tutorial worked fine immediately, as expected. ;)
[12:27] <niemeyer> TheMue: That said, there are already areas that may be explored in parallel
[12:27] <TheMue> niemeyer: OK
[12:28] <TheMue> niemeyer: I started with bootstrap just to keep the order of the steps I've done during the tutorial.
[12:29] <niemeyer> TheMue: Started in which sense?
[12:30] <TheMue> niemeyer: I'm trying to understand the bootstrap and how the first instance is setup. So I followed the Py code to the EC2MachineProvider. And here I also wanted to see, how far I can find similar in the Go port.
[12:31] <niemeyer> TheMue: Aha, cool
[12:31] <niemeyer> TheMue: The Go code is lacking at that point already
[12:31] <TheMue> niemeyer: Yep, I've seen
[12:31] <niemeyer> TheMue: rog actually has logic for that
[12:32] <niemeyer> TheMue: But we stepped back one of his branches, and now we're moving forward in slower increments with better testing
[12:33] <TheMue> niemeyer: The Py code is using generators very much. Nice.
[12:33] <niemeyer> TheMue: I don't find it so nice, but we're fixing that.. :)
[12:34] <TheMue> niemeyer: Hehe, it's not everyones logic.
[12:36] <TheMue> niemeyer: Btw, I would like some more verbose naming for public types, funcs etc. ReadEnvironmentsFile() instead of ReadEnvirons().
[12:36] <niemeyer> TheMue: It's not a generator really.. it's manual coroutine-like yielding
[12:36] <TheMue> niemeyer: Old Smalltalker school.
[12:37] <niemeyer> TheMue: Our convention is somewhere in the middle
[12:37] <niemeyer> TheMue: Long names do really help readability if you go overboard with them
[12:38] <niemeyer> TheMue: I can show you snippets with that convention being used where it ends up looking like reading a book
[12:38] <TheMue> niemeyer: I found them always helpful during long term maintenance of code. It depends on the visibility.
[12:39] <TheMue> niemeyer: Inside of funcs or even loops I often use simple one char vars too.
[12:40] <niemeyer> TheMue: Depends on a lot of things.. If we use Environs everywhere, calling it Environments doesn't help much, for instance
[12:40] <niemeyer> TheMue: I appreciate readable code, though.. so let's see if we can get to agreement as we go
[12:41] <TheMue> niemeyer: Don't think it will be a problem.
[13:19] <_mup_> txzookeeper/trunk r46 committed by kapil.foss@gmail.com
[13:19] <_mup_> [trivial] update pypi license metadata
[13:22]  * koolhead11 finds db-relation-changed most tricky part of writing charm!! :P
[13:25] <hazmat> g'morning
[13:26] <koolhead11> hola hazmat
[13:26]  * hazmat catches up
[13:27] <hazmat> koolhead11, congrats on the charm
[14:07] <_mup_> Bug #903213 was filed: need supporting code to help upstartify services <juju:In Progress by fwereade> < https://launchpad.net/bugs/903213 >
[14:10] <koolhead11> hazmat: thanks!! :)
[14:24] <fwereade> hazmat, if you have the appetitie for it, I have a horrifying pipeline of reviews
[14:25] <fwereade> hazmat, um... think how much worse it would be if there were fewer than 9 branches ;)
[14:25]  * fwereade hangs head in shame
[14:26] <fwereade> hazmat, it's basically the monster that grew from restart-transitions broken into 4; the upstartify-agents followup broken into 2; and then the other branches I'd already had dangling off those 2
[14:27] <fwereade> hazmat, for convenience, they go in this order: http://paste.ubuntu.com/767923/
[14:29] <hazmat> fwereade, thanks, the roadmap is very helpful
[14:35] <hazmat> as are the small branches
[14:45] <fwereade> hazmat, I also ended up writing a short essay clarifying just what the hell I was doing in restart-transitions
[14:45] <fwereade> hazmat, http://paste.ubuntu.com/767947/
[14:46] <fwereade> hazmat, it's a general discussion of all 4 initial branches
[14:46] <fwereade> hazmat, or, what *became* the 4 initial branches
[15:00] <niemeyer> Lunch time..
[15:02] <jcastro> marcoceppi: hi!
[15:02] <jcastro> mainerror: wow, looks like nijaba was busy while I was away
[15:02] <marcoceppi> jcastro: Morning!
[15:02] <jcastro> I meant marcoceppi
[15:02] <marcoceppi> Yeah he was
[15:02] <jcastro> oh cool, freeciv.
[15:02] <jcastro> looks like I have a bunch of blogging to do
[15:02] <mchenetz> Hi everyone...
[15:03] <mchenetz> I was playing around with Juju this weekend and i had a few observations...
[15:03] <jcastro> marcoceppi: is roundcube ready?
[15:03] <marcoceppi> jcastro: It's in the "charm store" nijaba has some improvements to it that I have yet to review. Primarily the peer relation thing
[15:03]  * jcastro nods
[15:04] <jcastro> man, your awesomeness knows no bounds.
[15:04] <jcastro> mchenetz: just ask!
[15:04] <hazmat> fwereade, That's awesome, minus the branch review that should get into the internal docs.
[15:04] <mchenetz> The, "add-relation haproxy:reverseproxy wordpress:website" command does not actually work that with local provider. It creates server entries in haproxy that forward to local host instead of the server.
[15:06] <mchenetz> I am thinking there has to be better differentiators from when something is actually hosted remotely vs local
[15:13] <mchenetz> I am working on some charms for Security... IPS with Snort, Event correlations engines, etc...
[15:16] <jcastro> that sounds great
[15:17] <mchenetz> That is just the start. I am very excited about Juju. :-)
[15:17] <jcastro> https://juju.ubuntu.com/Charms
[15:17] <mchenetz> Right now i am working on getting my openstack servers running...
[15:17] <jcastro> there's a link there to the spreadsheet, if you want to claim a charm and update it's progress
[15:18] <jcastro> https://docs.google.com/a/canonical.com/spreadsheet/ccc?key=0AoW1nhI7IMt3dFRvSFdkZmNqQ0t3RjZ2QTR2Z19teWc&hl=en_US#gid=0
[15:18] <mchenetz> Okay, i will do
[15:18] <jcastro> we're doing that so people don't spend like a day doing something that someone else is working on, etc.
[15:18] <jcastro> marcoceppi: and limesurvey, done?
[15:19] <mchenetz> That makes sense...
[15:19] <marcoceppi> jcastro: mhumm
[15:20] <jcastro> marcoceppi: heh, it would be cool to get some other team in the project to run a survey
[15:20] <marcoceppi> jcastro: What?
[15:20] <mchenetz> Added snort to the spreadsheet
[15:21] <jcastro> marcoceppi: I'll find a team in ubuntu that needs to run a survey, and have them try the charm
[15:21] <marcoceppi> Ohh, that would be awesome
[15:22] <TheMue> jcastro: Thx for the shared folder.
[15:24] <koolhead11> jcastro: owncloud2 is still not in our repo, in order to write its charm, can i use it from source/github?
[15:24] <jcastro> sure
[15:25] <koolhead11> jcastro: i have added my name against it there in the spreedsheet
[15:25] <jcastro> IMO having a config to use the latest stable upstream version should be a feature of every charm.
[15:25]  * jcastro hopes SpamapS doesn't hear
[15:25] <koolhead11> jcastro: :P
[15:42] <nijaba_afk> jcastro: hard to do if there is no way to get she sha1 or md5 of the latest tarball.  Sourceforge does not help in that way....
[15:42] <jcastro> ah bummer
[15:42] <m_3> morning
[15:43] <jcastro> m_3: welcome back!
[15:43] <SpamapS> jcastro: wait the spreadsheet should not be the canonical source of who is working on what. launchpad bugs should be. The spreadsheet is just an easier way to gather the new charm efforts.. or so I thought.
[15:43] <m_3> jcastro: you too
[15:43] <jcastro> SpamapS: you are correct.
[15:44] <SpamapS> jcastro: oh, and welcome back. :)
[15:44] <jcastro> mchenetz: hey can you file a bug on each of the charms you want to work on?
[15:46] <SpamapS> jcastro: perhaps we need to change the process for the new-charm tag so that people can file a bug with that tag on it immediately, and then we'll review when they change the status from New to Confirmed...
[15:46] <nijaba> heya SpamapS
[15:48] <jcastro> SpamapS: I don't think we have enough people doing that to warrant a new tag
[15:48] <jcastro> but yeah
[15:51] <SpamapS> jcastro: we should have an explosion of "I'm working on X" in the next couple of months.
[15:51] <SpamapS> jcastro: so I'd like to make sure if nothing else that we all agree on what people should do. I didn't even know the spreadsheet was public. ;)
[15:51] <SpamapS> Figured it was just shared with a few of us charmers or something.
[15:51] <SpamapS> nijaba: good morning. :)
[15:53] <nijaba> good morning marcoceppi
[15:53] <marcoceppi> morning nijaba
[15:54] <jrgifford> marcoceppi: hey, you're back!
[15:54]  * marcoceppi curses IRCCloud
[15:55] <jrgifford> don't suppose there are any decent alternativess for irc cloud yet?
[15:55] <nijaba> SpamapS: so, you think that sharing an sshkey via zk, then sending a private key via ssh is more secure than sharing the private via zk directly?
[15:56] <nijaba> SpamapS: I figured that if I can get access to the ssh key in zk, it is the same than getting access to the ssl private key, no?
[15:57] <nijaba> SpamapS: at leas this is the way I implemented it in roundcube charm (waiting for review/promulgation)
[15:57] <jcastro> jrgifford: I have a charm for alice IRC, but it's broken right now, waiting on upstream to stick it in cpan
[15:59] <jrgifford> jcastro: awesome.
[16:11] <SpamapS> nijaba: I'm concerned about *storing* the private key in ZK, vs storing the public key there.
[16:11] <SpamapS> nijaba: at least if its just the public key, they can only MITM future connections, they can't decrypt past traffic.
[16:12] <nijaba> SpamapS: ah, so the pub key would be used to accept conection from a master to get the key.  smart...
[16:12] <nijaba> SpamapS: I'll change that then
[16:14] <SpamapS> nijaba: granted, the public ssh key could be used to login and steal the private key, so I may be overthinking it.
[16:14] <hazmat> mchenetz, re local vs. remote.. the idea is to minimize the differences, that particular problem is a bug in the wordpress charm.. there's an additional cli hook api for resolving the public/private addresses in a provider agnostic manner
[16:15] <SpamapS> nijaba: this is why, IMO, juju cannot be trusted for secure work until we have *tight* ACL's on the data.
[16:16] <nijaba> SpamapS: well, not to login, but to get logged in.  So one would have to insert a host as a juju "slave" and generate a peer-joined event to get the key.  not 100% secure, but harder than just grabbing it from zk
[16:18] <SpamapS> nijaba: quite trivial actually. All nodes have full access to ZK right now.
[16:19]  * SpamapS wonders how that bug got bumped down in priority given the production focus for this cycle
[16:58] <mchenetz> hazmat: No problem... Just an observation.
[17:04] <SpamapS> mchenetz: could you report that as a bug, https://launchpad.net/charm/+source/wordpress/+filebug
[17:17] <mchenetz> Spamaps: no problem, I will report it
[17:20] <SpamapS> mchenetz: thanks!
[17:27] <mchenetz> Spamaps: you had asked me the other day about my observations about Local Provider vs Vagrant... Here you go..
[17:27] <mchenetz> Vagrant: Able to easily deploy on multiple OSs
[17:28] <mchenetz> Vagrant: Able to maintain state after reboots
[17:28] <mchenetz> Vagrant: Base OS is accessible on creation
[17:28]  * SpamapS puts an extra layer of asbestos on juju
[17:29] <mchenetz> Local-provider: Easy to create on linux (Other OS's require work)
[17:29] <mchenetz> Local-Provider: Base OS is not accesible until initial charm is installed (At least not to my knowledge)
[17:29] <mchenetz> Local-Provider: Does not maintain state (on this version)
[17:31] <negronjl> SpamapS: approved lp:~clint-fewbar/charm-tools/add-tests
[17:31] <SpamapS> mchenetz: THANKS, great thoughts, quite helpful
[17:31] <SpamapS> negronjl: w00t
[17:32] <zirpu> negronjl: did you post your slides from mongosv?
[17:32] <negronjl> zirpu: not yet but, I will. Trying to figure out the best way to do that ... any suggestions ?
[17:34] <zirpu> slideshare? it's pretty common now. i think it accepts pdfs.
[17:35] <fwereade> hazmat, jimbaker, bcsaller: standup?
[17:36] <SpamapS> mchenetz: So, to address some of the things..    multi-OS .. somebody just needs to write a provider that uses virtualbox and we'll that bit.. but agreed, we're targetting Ubuntu right now. :)
[17:36] <SpamapS> mchenetz: The base OS is meant to be very "standard" because charms do all the customization, so I'm not sure thats a problem in juju.
[17:37] <SpamapS> mchenetz: the ability to reboot is landing in a series of reviews right now actually
[17:37] <mchenetz> Spamaps: I am just think of what a development team would use... Most the devs i see work on a mac or pc and then will push stuff to the cloud.
[17:38] <mchenetz> So they could be working on windows or osx... Having the Vitualbox image allows them to work in a similar environment..
[17:40] <mchenetz> The one nice thing they did was expose the inited directory to the VM so that users can easily copy files into the OS directory and they can be seen on the VM.
[17:40] <mchenetz> I think that is a great feature for people that don't know that much about VMs and then really don't have to.
[17:40] <SpamapS> yeah containers are nice that way
[17:41] <mchenetz> Any quick tip on, "pty-allocation-request-failed-on-channel-0"
[17:42] <jimbaker> fwereade, works for me
[17:44] <bcsaller> yeah, I'm around too if there are things we have to cover
[17:46] <m_3> mchenetz: I've been able to recover from that by wiping my lxc cache
[17:46]  * m_3 looking for the cache
[17:46] <fwereade> hazmat, jimbaker, bcsaller: I think I've invited you all
[17:47] <mchenetz> m_3: how do i go about doing that?
[17:47] <m_3> mchenetz: I'd start with a 'juju -elocal destroy-environment'
[17:47] <mchenetz> m_3: i just created the environment. :-(
[17:47] <m_3> mchenetz: then rm -Rf /var/cache/lxc/
[17:48] <mchenetz> m_3: okay i will try. Thanks...
[17:48] <m_3> mchenetz: might not be the same problem you're seeing, but that worked for me
[17:49] <m_3> mchenetz: I'd have a local env that gets corrupted after a couple of weeks of solid usage... clearing the lxc cache resets me to working local env
[17:50] <m_3> mchenetz: remember that the first 'juju deploy' after a cache wipe is going to take a _long_ time
[17:50] <marcoceppi> What's the policy for charm-contributors again? "Open" group, right?
[17:50] <m_3> marcoceppi: code of conduct signed
[17:50] <mchenetz> m_3: sounds good trying now
[17:50] <marcoceppi> m_3: o/ and thanks
[17:51] <zirpu> is the 1st environment in ~/.juju/environments.yaml the default?
[17:52] <marcoceppi> zirpu: No, if you have multiple environments and don't have a default: key set or use the -e flag juju will throw an error
[17:54] <m_3> zirpu: there's a 'default:' top-level entry in environments.yaml...  I see code that gripes if there're multiple envs w/o a default
[17:54] <zirpu> you mean an environment var set?
[17:55] <zirpu> like: export JUJU_ENVIRONEMTN=foobar
[17:55] <zirpu> (minus typos)
[17:56] <m_3> zirpu: at the moment you have to explicitly set a default through the e
[17:57] <m_3> 'default' key in environments.yaml or via the command line option '-e' on all commands
[17:57] <EvilBill> Jrgifford: Yeah, I have a cube. Hotrodded with an Nvidia gfx card and 1.5GB RAM. Though I don't use it and it sits in the garage...
[18:01] <m_3> zirpu: a shell env variable such as JUJU_ENV was discussed on the list... dunno if it's going to be implemented though
[18:08] <mchenetz> Spamaps: Bug report submitted
[18:08] <mchenetz> Bug id:903312
[18:09] <mchenetz> m_3: that worked... Thanks
[18:09] <mchenetz> Although, that corruption error could be a potential problem and/or bug for local provider
[18:16] <m_3> mchenetz: yes agree... filing the bug now so we at least have a place to capture the lxc cache reset
[18:22] <mchenetz> m_3: sounds good
[18:25] <_mup_> juju/ssh-known_hosts r437 committed by jim.baker@canonical.com
[18:25] <_mup_> Added support for ssh_keys in format_cloud_init
[18:34] <_mup_> Bug #903318 was filed: juju ssh fails on local provider: pty-allocation-request-failed-on-channel-0 <juju:New> < https://launchpad.net/bugs/903318 >
[18:39] <m_3> mchenetz: ^^ is the bug... please verify what you did to get it working again.
[18:40] <mchenetz> m_3: exactly... I would attempt to perform an SSH to a service/0 or whatever and i would get pty-allocation-request-failed-on-channel-0
[18:41] <m_3> mchenetz: thanks
[18:41] <mchenetz> The fix would be to do, juju detroy environment, and rm -Rf /var/cache/lxc/
[18:52] <niemeyer> WOohay.. lbox submit works
[19:10]  * SpamapS confirms that bug and marks it High
[19:22] <jcastro> marcoceppi: do you have a snippet/function handy for the config switch to pulling an upstream version handy?
[19:22] <marcoceppi> The phpMyAdmin implementation for doing a switch is crazy ugly. You can look at it for ideas, it's implemented in lib/common.sh
[19:23] <marcoceppi> In a nut shell install_phpmyadmin is called during config-changed which, depending on the current installed version and the value in each hook, will call install_upstream or install_apt
[19:24] <marcoceppi> Then there are two environments in env/ which load the appropriate variables for each case :\
[19:24] <marcoceppi> At one point it does a barrel roll, then viola!
[19:24]  * SpamapS would have liked to see an immelman
[19:25]  * marcoceppi does an Immelman turn
[19:25] <jcastro> thanks!
[19:27] <marcoceppi> Np, I'd like to make this a lot more cleaner but still trying to hash out the best way to do it
[19:27] <marcoceppi> and by a lot more cleaner, I mean a lot more clean
[19:36] <jcastro> SpamapS: I think I am getting this bug: https://bugs.launchpad.net/juju/+bug/891419
[19:36] <_mup_> Bug #891419: Juju fails test suite when building on precise <juju:Fix Released> < https://launchpad.net/bugs/891419 >
[19:36] <jcastro> when trying to bootstrap, my precise is up to date as of today
[19:37] <jcastro> SpamapS: nevermind.
[19:38] <SpamapS> jcastro: mind never'ed
[19:38] <EvilBill> jcastro!
[19:38] <jcastro> hiya bill!
[19:38] <EvilBill> how's it going?
[19:39] <jcastro> about to test the limesurvey charm so I can blog it
[19:39] <EvilBill> cool
[19:39] <EvilBill> btw
[19:39] <EvilBill> stop saying that you have a macports client.
[19:39] <EvilBill> there isn't one that usable that I can find
[19:39] <EvilBill> I just checked out the SVN version of macports, and its' not in there.
[19:40] <jrgifford> EvilBill: speaking of mac clients, i haven't found anything either.
[19:40] <EvilBill> hey jrgifford
[19:40] <jcastro> lynxman: heya, mac client?
[19:41] <EvilBill> I see this:
[19:41] <EvilBill> https://trac.macports.org/ticket/30237
[19:41] <EvilBill> but that isn't something can can be "port install'd"
[19:46] <lynxman> jcastro: waiting on approval...
[19:46] <lynxman> EvilBill: it can be port installed :)
[19:46] <EvilBill> lynxman: Great! How?
[19:46] <lynxman> EvilBill: just download the Portfile from there and add a local repository on your machine
[19:46] <EvilBill> I tried that last night, it blew up on some zookeeper dependency.
[19:47] <lynxman> EvilBill: because it depends on several portfiles, one of them being zookeeper
[19:47] <EvilBill> I'm putting ports on this new mac here at work, and I'll try it on this
[19:47] <lynxman> EvilBill: there's 7 tickets open for this :)
[19:47] <lynxman> EvilBill: so you'll need to fill all the dependencies, unfortunately
[19:47] <EvilBill> awesome, how can I get onboard to add my voice to the crew?
[19:48] <lynxman> EvilBill: I've been talking with the #macports guys for a while, but it's a very very very slow process
[19:48] <EvilBill> Great.
[19:48] <EvilBill> I see that the juju portfile you submitted has been there for two months.
[19:49] <EvilBill> and the version it refers to is 0.5+bzr398
[19:49] <lynxman> EvilBill: if you want to get all portfiles, tickets 30221 30222 30223 30236 31570 30237
[19:49] <lynxman> EvilBill: that's the one in Oneiric release
[19:49] <jcastro> I thought there was some new sexy replacement for macports all the kids are all excited about
[19:49] <EvilBill> ok
[19:49] <marcoceppi> How hard would it be to build it into a "native" mac osx application?
[19:50] <hazmat>  jcastro there is.. homebrew
[19:50] <EvilBill> marcoceppi: I have no idea. My guess is you'd need to package up all the dependencies… like zookeeper, twisted, yaml, etc, into one monolithic bundle.
[19:50] <EvilBill> yeah, homebrew is up and coming, but no idea how well that thing works.
[19:50] <marcoceppi> EvilBill: sounds like a pain :\
[19:50] <EvilBill> marcoceppi: Yeah, I tend to agree.
[19:50] <lynxman> marcoceppi: quite difficult, specially following all the packaging guidelines and meeting all requirements
[19:51] <lynxman> marcoceppi: and I'm in no way a OSX packaging expert :)
[19:51] <hazmat> homebrew does /usr/local installs afaicr, while macport/fink use sandbox dirs off /opt i believe
[19:51] <lynxman> hazmat: correct
[19:52] <hazmat> hm..homebrew chilling with 400 outstanding pull requests, its the most active of the bunch that i know
[19:53] <hazmat> and it already has zookeeper and zookeeper python
[19:53] <hazmat> https://github.com/mxcl/homebrew/blob/master/Library/Formula/zookeeper.rb
[19:53] <hazmat> the rest of juju is basically pure python libs
[19:54] <jcastro> https://bugs.launchpad.net/charm/+bug/903361
[19:54] <_mup_> Bug #903361: Charm needed: Alice IRC <new-charm> <juju Charms Collection:New for jorge> < https://launchpad.net/bugs/903361 >
[19:54] <jcastro> my charm is finally ready!
[19:55] <jcastro> (for review)
[19:55] <jcastro> spent this whole time wrestling building the thing from source, ends up it was in the archive as of 11.10. :-/
[19:56] <hazmat> hmm.. given that.. maybe the osx story just becomes use pip/easy_install to get juju
[19:56] <hazmat> post homebrew install zookeeper
[19:59] <jcastro> ok so who wants to put this in homebrew?
[19:59] <jcastro> we need something for OSX people
[20:02] <EvilBill> in my experience, macports is still the thing people go to
[20:02] <EvilBill> our guys around here use it a lot
[20:02] <EvilBill> I say lean on the macports guys
[20:02] <EvilBill> if it's ready to go, it shouldn't be a big deal to get it in
[20:02] <mchenetz1> I feel homebrew is the new macport
[20:02] <jrgifford> maybe out west. around here, most of them are homebrew people.
[20:02] <mchenetz1> I use brew now
[20:03] <mchenetz1> Macports is old school
[20:03] <hazmat> EvilBill, https://trac.macports.org/ticket/30237
[20:03] <hazmat> that's the juju portfile
[20:03] <EvilBill> hazmat: thanks
[20:03] <hazmat> it doesn't seem like we've addressed some of the original upstream comments
[20:03] <hazmat> er. distributor
[20:04] <mchenetz1> Maybe its because I do a lot of rails development and brew seems big in the rails world
[20:15] <lynxman> hazmat: they have been addressed, I had to open a new ticket
[20:17] <lynxman> hazmat: the whole process is a bit of a pain though :/
[20:20] <SpamapS> mchenetz: would be nice if there was just one. :)
[20:21] <SpamapS> lynxman: is anybody even responding to these? they're all so old
[20:23] <marcoceppi> Looks like Pecl and Pear's queue for acceptance. Dead
[20:25] <EvilBill> lynxman: hey, you submitted byobu. Awesome.
[20:42] <EvilBill> ok, here I go trying to build juju via macports.
[20:44] <EvilBill> and libzookeeper failed to build
[20:44] <EvilBill> *headdesk*
[20:44] <mchenetz1> Evullbill: tell me what you did after your done and I will try it too...
[20:49] <EvilBill> mchenetz1: I can't even get that far. Libzookeeper (a dependency, but one that macports knows about) blows up on build.
[20:50] <mchenetz1> I will try to figure it out when I get home...
[20:50] <mchenetz1> If I figure it out before you then I will tell you what I did...
[20:51] <EvilBill> Thanks.
[20:51] <EvilBill> I'm going to get a sandwich before my next meeting.
[20:52] <_mup_> Bug #903392 was filed: Vagrant as a cross-platform local provider? <juju:Triaged> < https://launchpad.net/bugs/903392 >
[20:54] <mchenetz1> Vagrant as a local provider would be great... I think the two projects will work well together..
[20:56] <m_3> mchenetz1: might be lightweight ways to just wrap the current local provider
[20:56] <m_3> but either way... something there
[20:57] <m_3> lower that barrier to people getting to take juju out for a spin
[20:57] <mchenetz1> Sounds good... I am very interested in this piece of juju
[20:57] <mchenetz1> I think that is one of the biggest issues right noe
[20:58] <mchenetz1> Now
[22:28] <EvilBill> looks like I picked the wrong week to quit sniffing glue...
[22:37] <_mup_> juju/ssh-known_hosts r438 committed by jim.baker@canonical.com
[22:37] <_mup_> Tests pass with dummy provider - a start