[01:08] <imbrandon> black on black gives me a heart attack, and the silence makes it deadly ...
[01:17] <lifeless> broken theme
[01:17] <lifeless> lets see
[01:19] <imbrandon> heya lifeless
[01:44] <imbrandon> lifeless: ahh u figured it out, someone trip over the wrong cat5 cable ? heh
[02:08] <imbrandon> lifeless: i got to looking at the structure of the .app and it occured to me that the same .app could in theory be used for Mac and Linux e.g. the file system layout/api is Whatever.app/Contents/{Resources,MacOS,Linux} etc
[02:08] <imbrandon> and the binary actually lived in the OS named subdir
[02:10] <imbrandon> i guess there could be a gnustep/nextstep api writen ( or already written for win as well )
[02:13] <lifeless> imbrandon: dunno what was up, I just asked IS who fixed ;)
[07:00] <imbrandon> ugh wtf, so unattended upgrades decided to update the fileserver while i'm midst rsyncing 100gb+ of crap to it and sshd decided it was done for the evening and wasent going to return for the remainder of the show
[07:01]  * imbrandon is glad said fileserver is 2 rooms away
[08:52] <imbrandon> mmm smell of a fresh ubuntu install on my main machine bright and early on a tuesday morning ... gonna be a good week ( i hope )
[08:52] <imbrandon> now just the matter of to upgrade to quantal or stay on precise
[08:52] <imbrandon> heh
[08:55]  * imbrandon just found out an old buddy of his quit his ( rather high paying ) job to form a startup and is the infrastructure is managed ( mostly ) with juju on aws, think its time i had a beer with him and see what he's upto :)
[09:02] <imbrandon> jamespage: got through all of the video too btw, apart from some audio glitches ( that obviously you had no control over and arent terrible anyhow ) it was an outstanding overview i think, in fact the best "here is juju" all in one place i've seen of the admiditly few juju vids i've watched ( mostly just jcastro at oscon and the UDS ones )
[09:02] <imbrandon> :)
[10:50] <jamespage> imbrandon, thanks for the feedback!
[11:08] <imbrandon> you know what ...
[11:09] <imbrandon> we need to marry one ( more than one eventually ) one of the configuration management tools into juju as a first class citizan ...
[11:10] <imbrandon> we keep saying we're not chef etc etc and they can be used inside juju ... so lets do it ...
[11:11] <imbrandon> SpamapS m_3 ^
[11:12] <imbrandon> ( even cloud-init has puppet and capistrano baked in )
[11:17] <imbrandon> i mean think about it, what can we do that shows juju can orchstrate not just wordpress but , deployment automation ( cloud-init , preseed , cobbler ) ; config managements tools ( puppet , chef , bcfg2 ) ; and Build Automation ( jenkins/hudson , ant, maven ) as well as make them scale-out , until now we've shown nicely how we can make largely default apt-get software scale out, but what about all these custom tools people use now individually and 
[11:18] <imbrandon> make them first class citizens $somehow and _show_ we're not replacing them we marrying them. I mean we are right ? We would be crazy to reinvent the wheel on some of this stuff
[11:27]  * imbrandon waits for the world to wakeup and get told he's nuts
[11:27] <imbrandon> ( or "show me the code" )
[11:27] <imbrandon> heh
[12:16] <imbrandon> heya SpamapS here is a few targets for the generic monitoring interface: http://paste.ubuntu.com/1084350/
[12:16]  * imbrandon quickly raises his hand and says "1 ... 2 ... 3 ... not it !"
[12:49] <imbrandon> default: sample; environment: ; sample: ;     type: openstack
[12:49] <imbrandon> config: chef-solo
[12:53] <hazmat> mgz, just noticed you pushed some revs to the ostack branch, could you lbox propose -cr on it again to update the code review
[12:54] <imbrandon> new rpms's up https://github.com/jujutools/rpm-juju/blob/master/juju-0.5.1-1.noarch.rpm?raw=true
[12:55] <imbrandon> morning hazmat
[12:55] <hazmat> imbrandon, greetings
[13:15] <mgz> hazmat: will do, need to land one more thing first
[14:02] <imbrandon> hazmat / SpamapS  or someone willing to laugh at my python got a moment to tell my why this function does not do what i want
[14:02] <imbrandon> http://paste.ubuntu.com/1084504/
[14:03] <imbrandon> i get 0.0
[14:03] <hazmat> imbrandon, lack of paths
[14:03] <imbrandon> ahh kk
[14:03] <hazmat> maybe
[14:03] <imbrandon> well i exe the same thing in the the dir as the py file and it does as expected
[14:04] <imbrandon> so the commands them selfs are ok, did not think about paths
[14:04] <mgz> well, it's a little hillarious using os.system('cat ...') rather than open('...').read()
[14:05] <imbrandon> mgz: heh give me php, js, ruby, or a hand full of other things i'm ok , python == total newb :)
[14:05] <imbrandon> so very very likely
[14:06] <imbrandon> i'm sure there are bzr modules i could import as well
[14:07] <imbrandon> :)
[14:42] <_mup_> Bug #1015207 was filed: juju setup fails, ERROR Invalid SSH key - 12.04 LTS <juju:New> <MAAS:New> <juju (Ubuntu):Confirmed> < https://launchpad.net/bugs/1015207 >
[16:00] <dpb___> what should i be looking for if I don't see anymore juju debug-log after 2012-07-10 10:43:36,009 Machine:0: unit.deploy INFO: Started service unit lds-quickstart/2
[16:00] <dpb___> for lxc/local ^^
[16:31] <dpb___> found logs in my local directory... should be enough to go on for now
[17:15] <imbrandon> 07/03/2012 Amazon Web Services charges $2.24
[17:15] <imbrandon> man i love bills that small , wish they all were :) heh
[17:23] <tncardoso> are all files in the charm directory uploaded to the server being deployed? I want to send along with the charm some auxiliary files (like a default configuration file).
[17:24] <imbrandon> yes
[17:24] <imbrandon> tncardoso: thats pretty common case, like in my charms i create a /templates directory at the top level
[17:25] <imbrandon> then the hooks can read in that file, edit values, and write it back out to where it needs to go
[17:25] <imbrandon> etc
[17:25] <tncardoso> imbrandon: ok, thanks :)
[18:12] <dpb___> anyone seen this one: (in unit.log): Failure: txzookeeper.client.ConnectionTimeoutException: could not connect before timeout
[18:12] <dpb___> basically, my instances sits in *pending* forever.
[18:12] <dpb___> this is on a lxc environment
[18:12] <SpamapS> dpb___: is zookeeper running?
[18:13] <SpamapS> dpb___: and do you have any local firewalls that might be blocking connecting to it?
[18:17] <imbrandon> quad core 2.7ghz with 8gb ram and ruby STILL takes forever to compile ...
[18:21] <benji> imbrandon: SSD?
[18:23] <dpb___> hmm
[18:25] <imbrandon> benji: no :(
[18:25] <dpb___> SpamapS: yes, zookeeper appears to be running
[18:25] <benji> imbrandon: you have my condolences
[18:26] <SpamapS> dpb___: and no firewalls?
[18:26] <SpamapS> dpb___: like, can you type 'telnet 192.168.122.1 <whatever-port-zk-listens-on>' ?
[18:26] <dpb___> this machine does have a public and private interface.  I'm not sure about any firewalls, is there a way I could empirically test?
[18:26] <imbrandon> heh, benji its about the only thing i ever wait on it seems, everything else is nice and quick
[18:26] <imbrandon> :)
[18:26] <dpb___> ok.. let me see.
[18:26] <dpb___> clientPortAddress=192.168.122.1
[18:26] <dpb___>  clientPort=60457
[18:26] <SpamapS> dpb___: does teh lxc container actually boot up? You can try ssh'ing to it (you can get the IP from /var/log/syslog as dnsmasq will log it)
[18:27] <dpb___> SpamapS: seems I can  telnet 192.168.122.1 60457
[18:28] <benji> yeah, you probably get a lot of benefit from disk caching
[18:30] <SpamapS> imbrandon: try compiling with eatmydata
[18:30] <imbrandon> i'm due for an upgrade anyhow, def making sure i double the ram and have the ssd primary drive option this time
[18:30] <SpamapS> imbrandon: I don't *think* compilers fsync() .. but its worth a try ;)
[18:30] <imbrandon> SpamapS: eatmydata ?
[18:30] <SpamapS> imbrandon: yes, turns off fsync()
[18:30] <imbrandon> ahh
[18:30] <imbrandon> nice
[18:30] <SpamapS> replaces it with a noop
[18:31] <SpamapS> CC="eatmydata gcc"
[18:31] <imbrandon> :)
[18:31] <SpamapS> dpb___: from your machine, or from the container that was booted?
[18:31] <imbrandon> ruby-1.9.2-p320 - #extracting ruby-1.9.2-p320 to /usr/share/ruby-rvm/src/ruby-1.9.2-p320
[18:31] <imbrandon> ruby-1.9.2-p320 - #extracted to /usr/share/ruby-rvm/src/ruby-1.9.2-p320
[18:31] <imbrandon> shit sorry
[18:31] <imbrandon> ruby-1.9.2-p320 - #configuring
[18:31] <imbrandon> ruby-1.9.2-p320 - #compiling
[18:32] <dpb___> SpamapS: I don't know that there is a container.  This is from my machine.
[18:32] <dpb___> SpamapS: I do lxc-list and i only see things that are "Stopped"
[18:33] <dpb___> SpamapS: ok, I am seeing some dhcp stuff in syslog... let me see If I can get the iP
[18:48] <dpb___> SpamapS: So, telnet 192.168.122.1 60561 is not working on the created instance
[18:48] <dpb___> SpamapS: I can ping that IP
[18:55] <SpamapS> dpb___: check iptables
[18:55] <SpamapS> dpb___: ufw does cause problems
[19:10] <dpb___> SpamapS: that was it.  Thanks!!
[19:14] <SpamapS> dpb___: no problem...
[19:15] <SpamapS> jcastro: ^^ perhaps we should add a note to the local provider docs about ufw
[19:16] <jcastro> sure
[19:16] <jcastro> hrmph, I haven't had a ufw problem with 12.04 and the local provider
[19:16] <jcastro> is there a ufw step somewhere that I am missing?
[19:17] <SpamapS> not sure
[19:18] <SpamapS> jcastro: I would actually expect ufw to always cause an issue
[19:18] <SpamapS> because its a remote host hitting an unaccounted for port
[19:21] <jcastro> https://juju.ubuntu.com/docs/getting-started.html
[19:21] <jcastro> ok so in which step are we turning off ufw?
[19:21] <jcastro> also, is it just turning it off or allowing the local stuff?
[19:26] <SpamapS> jcastro: I think ideally we'd provide some kind of ufw command to allow the lxc network to access us
[19:26] <SpamapS> I don't know ufw well enough to know that, but something like "hey ufw, let 192.168.122.0/24 access us"
[19:46] <jcastro> SpamapS: indeed
[19:46] <tncardoso> is it common to lock on: 2012-07-10 16:33:06,852 Machine:0: unit.deploy INFO: Started service unit serv/0 - what this mean? Where can I find more about the deploy lifecycle?
[19:52] <imbrandon> its not locked there, thats means its started
[19:52] <imbrandon> no much more to output at that point until another event happens
[19:56] <tncardoso> imbrandon: humm, but the debug hook didnt show anything
[19:57] <tncardoso> imbrandon: for example, trying to ssh gives the following info:   Waiting for unit to come up
[19:59] <imbrandon> how long ? i've seen it take nearly an hour and half on local provider for the first instance
[20:03] <tncardoso> imbrandon: 30 minutes
[20:03] <tncardoso> imbrandon: in this step it is downloading the server image?
[20:03] <tncardoso> imbrandon: *is it
[20:04] <imbrandon> yes
[20:04] <imbrandon> among other things, the first instance to be fired up takes a *long* time sometimes
[20:05] <imbrandon> grabbing about 300mb+ of data
[20:05] <imbrandon> etc
[20:05] <imbrandon> after that it should be fast
[20:05] <imbrandon> even if you destroy env etc
[20:06] <tncardoso> imbrandon: got it.. thanks a lot
[20:06] <tncardoso> :)
[20:06] <imbrandon> np
[20:44] <jcastro> http://engineering.twitter.com/2012/07/caching-with-twemcache.html
[20:44] <jcastro> this could be a cool charm, I've filed a bug on it
[20:44] <jcastro> if someone is looking for something to do
[22:12] <negronjl> SpamapS: ping
[22:18] <jimbaker> SpamapS, just want to double check that i can merge the reuse-security-groups branch into trunk, now that it's been approved
[22:48] <imbrandon> I some opinions / expertise here so i dont go looking like an idiot with this post, as i'm sure its going to draw some attn, what would be the "best" way i can show that juju could be used to orchrastrate db replication ... over wan ... across providers.
[22:48] <imbrandon> ( this is only a part of the post but still ... )
[22:49] <imbrandon> SpamapS: your a DB guy right ? ^^ opinon ?
[22:53]  * imbrandon has a blog post 3/4 written about using juju to create your own PaaS-like envrionment instead of ending up like Pentrest and others on AWS or Heroku  
[22:54] <imbrandon> all the goodies from a PaaS , none of the black box :)