Time Nick Message 00:42 larsks Grrr. Well, I've traced through the code and I figured out where adminPass is going and the answer is its broke as hell. 01:30 westmaas uh oh 01:36 larsks westmaas: Look in nova/virt/disk/api.py. It tries to copy /etc/passwd and /etc/shadow out of the target filesystem and put them back. 01:36 larsks Works great with Linux, not all with Windows. And it tries that if when using a config_drive (which is empty, so no /etc/passwd there either). 01:36 westmaas ah ha. 01:36 westmaas doh. 01:37 larsks Trying to track down what commit correspond to my version of that file so that I can work up a patch. 01:37 westmaas that'd be awesome. 01:56 pdurbin http://devops-boston.eventbrite.com does look great. thanks, guys. can be hard to get away for stuff like this though. 03:01 pdurbin the fu to get vagrant centos base boxes to support hostonly networking (the stuff i was talking about in yesterday's log): https://github.com/phase2/veewee-boxes/blob/master/definitions/CentOS-6.0-x86_64-netboot/postinstall.sh#L33 03:10 pdurbin it works! so that veewee box definition was used to build this base box: http://dl.dropbox.com/u/9227672/CentOS-6.0-x86_64-netboot-4.1.6.box . and the IPs that were defined in the Vagrantfile (192.168.206.130, 192.168.100.130) now work on centos, just like how the were working on ubuntu 03:19 pdurbin oh good, they log their channel. here you go... the whole conversation: http://irclogger.com/.vagrant/2012-10-31#1351737777 03:20 pdurbin not to self: bug smerrill in about a week for a centos 6.3 base box with the proper guest additions for hostonly networking: http://vagrantup.com/v1/docs/guests.html 03:21 pdurbin sjoeboo: or maybe you can get me spun up on packaging base boxes. but i think you're on vacation so maybe i'll read the docs :) 03:21 pdurbin or see how far i can get with openstack on centos 6.0... but i suspect trouble there 03:45 larsks pdurbin: OpenStack + CentOS works pretty well :). 03:46 larsks In other news: It lives...IT LIVES!!! Patches to openstack-nova that does sane things with adminPass. So tomorrow I should have my windows image doing the right thing. 03:47 pdurbin larsks: cool. just posted the fix to that github issue: https://github.com/mitchellh/vagrant/issues/997#issuecomment-9970185 03:48 larsks Our Puppet manifests do that on all of our CentOS 6 systems. 03:48 larsks That bit of udev is just too clever for its own good. 03:53 pdurbin heh 03:54 pdurbin also from #vagrant, this talk is recommended. inspiring open source story, they say: http://confreaks.com/videos/1081-cascadiaruby2012-an-open-source-story 12:43 pdurbin good talk. i approve 13:28 pdurbin fixin' django apps 13:29 pdurbin logging out. don't want to get *too* involved :) 14:12 pdurbin Learning to Scale a Public Cloud - http://vimeo.com/52266808 (openstack talk) 14:22 pdurbin "Bio: Gabe Westmaas manages teams at Rackspace that write software for the OpenStack nova and glance projects." 14:22 pdurbin westmaas: looking forward to watching! 14:33 westmaas pdurbin: 85% "umms" 14:33 pdurbin heh 14:33 pdurbin i've been meaning to give some tech talks 14:37 westmaas I need to practice more, that much is clear 15:07 Barb___ Hello folks..first time here, just saw Philip's abcd message :) 15:23 pdurbin Barb___: hi! 15:23 pdurbin crimsonfubot: lucky harvard abcd 15:23 crimsonfubot pdurbin: http://www.abcd.harvard.edu/ 15:24 pdurbin Barb___: not everyone here knows what abcd is :) 15:25 Barb___ good point! 15:52 * pdurbin finds himself linking someone to this nmblookup fu: http://irclog.perlgeek.de/crimsonfu/2012-05-01#i_5524882 16:21 vvuksan boston folks http://devops-boston.eventbrite.com/ 16:22 semiosis hi vvuksan, i copied Wuher's message in ##infra-talk about it here yesterday :) 16:22 vvuksan sweet 16:22 vvuksan thx 16:22 semiosis yw, wish i could go... but i'm in miami 16:22 vvuksan :( 16:54 larsks Anyone here know if it's possible to mount a whole-disk filesystem under Windows? OpenStack "config drives" have no partitions which at least initially seems to confuse Windows. 16:59 gridiron hmmm...does this help? http://serverfault.com/questions/444446/mounting-whole-disk-filesystems-in-windows-2008 16:59 gridiron Otherwise...I got nothin. 17:03 larsks gridiron: That would be the question I posted :). 17:04 pdurbin larsks: any docs besides this? http://docs.openstack.org/developer/nova/api_ext/ext_config_drive.html 17:05 larsks pdurbin: Not sure exactly what you're asking me... 17:05 larsks There are not many docs on the use of config drives, which is why I went spelunking through the source yesteday... 17:05 larsks It's a trap! 17:05 pdurbin westmaas: more docs please ^^ 17:08 pdurbin ventz: "Introduction to using MySQL Master-Master in an active/passive configuration behind mysqlproxy to enable zero downtime database schema changes (and thus, zero downtime, continuous deployments)." -- http://devops-boston.eventbrite.com free! tomorrow! 17:09 gridiron Sorry larsks... copied the wrong tab. 17:09 pdurbin and a talk on foreman (which we use) for libvirt and openstack provisioning... 17:09 gridiron It was a link for Linux Reader... 17:10 gridiron and now I can't find it...grrrr 17:11 gridiron Anyway... The reader allows it to mount but is read only was the gist 17:12 gridiron you need read write I take it? 17:12 pdurbin huh, and a talk on orc? a language i haven't heard about... http://orc.csres.utexas.edu ... unless it's something else. "Orc, a tool for providing zero-downtime deployments of contract-compliant web applications" 17:13 ventz pdurbin: what time? 17:14 ventz ughh, whole dya 17:14 ventz i really want to go to this, but i have to do 1 smr. deployment + 2 calls 17:14 pdurbin it sounds awesome 17:14 ventz yea it does 17:15 larsks gridiron: I don't necessarily nead to write it, I guess, although I was hoping there was a solution that doesn't involve third-party software. 17:15 larsks I may need to teach OpenStack about "fdisk" :(. 17:17 pdurbin larsks: you have a lot of demand for windows VMs? 17:17 larsks We would like our Windows admins to be able to take advantage of OpenStack for service prototyping in the same way that we can for Linux hosted services. 17:17 larsks It seems like a reasonable request :). 17:17 larsks I'm getting a little jaded by the OpenStack code, though. 17:31 pdurbin i don't like the sound of that. the code thing :) 17:33 pdurbin ventz: so you got your splunk api code working? 17:33 larsks Just this particular feature seems poorly thought out. Obviously the rest of it seems to work out okay :). 17:35 pdurbin larsks: and you're still thinking these VMs will be ephemeral? you have the Windows admins or whoever backup their stuff? and let them destroy and re-create their VMs at will? 17:36 pdurbin i promise to stop asking you this every week :) 17:39 larsks Maybe? Ask me in six months. 17:41 pdurbin magoo: a little bird tells me you use vagrant 17:46 ventz pdurbin: yes 17:46 ventz pdurbin: i'll share it 17:47 pdurbin ventz: cool. i think i'm going to focus on logstash anyway but would be nice to have it handy. Pax has a good point about using python though 17:50 ventz one thing i've always loved about perl has been the ease of regexs. That said, the "programing" side could really use some actual "programing" instead of a layer of hacks on top of eachother 17:50 ventz and that's one thing I like about ruby and python 17:50 ventz i wish perl did this -- well v6 will, but yea, it's far far away from where it should be 18:08 pdurbin ahhh. `yum install puppet` on fedora gives me syntax highlighting in vim 18:42 pdurbin "In this test I peaked out at about 10 Gb/s" -- guy at work testing a storage array 18:51 semiosis OH in ##infra-talk just now: <ells> anyone used: http://ansible.cc/ 18:51 semiosis wasn't someone here loving it the other day? 18:51 pdurbin semiosis: that was larsks 18:52 pdurbin semiosis: i've been meaning to ask larsks about https://github.com/larsks/crackpot 18:53 semiosis great name 18:59 ironcamel ventz: you wished perl did what? 18:59 pdurbin ironcamel: i was kinda hoping you hadn't seen that ;) 19:00 ironcamel why? 19:00 ironcamel i'm confused :) 19:00 pdurbin your love of perl is strong. even stronger than mine :) 19:01 ironcamel was it just trolling? 19:01 pdurbin no, ventz is a perl guy. it's all good :) 19:01 ironcamel ok, i'll move along then :) 19:01 ironcamel let me know if there's actually something that wanted to know how to do in perl 19:02 ironcamel i will show you how :) 19:02 ventz i love perl so much that I decided in college to take a "software dev" class, and my language choice was perl 19:02 ventz it was great at first -- quick parsing, modules, etc... 19:03 ventz half way through, i realized it had become a little harder to replicate stuff that came for "free" with other languages 19:03 ventz by the end, i had 80,000+ lines of code 19:03 ventz i learned a very valuable lesson 19:05 pdurbin ironcamel: we love you for this :) 19:05 pdurbin ventz: what language would you choose today for that? 19:05 ironcamel i hope the lesson you learned was to use CPAN, instead of writing 80,000 lines of code 19:10 ventz ironcamel: 80,000 lines were w/ almost every module i could find :) 19:10 ventz the annoying part at the time was writing a synchronous web server -- ended up using IO::Select 19:11 ventz and the other thing which was very broken at the time was the THaw/Freeze modules so that you can serialize a whole object 19:11 ventz i remember trying the top 3-4 and all being broken. Since another group used perl, we ended up emailing the devs and got emails along the lines "yea...some stuff doesn't work the way it should yet" 19:12 ventz pdurbin: i would probably use python or ruby 19:12 pdurbin ventz: ok 19:12 ventz at least 10% of the code was trying to use an object system which allowed object protection from read/write/change 19:13 ventz that comes by default with any half working OO language 19:13 pdurbin ventz: this was pre-Moose, right? 19:13 ventz moose was around, but it was still "iffy" 19:13 ventz pdurbin: one sec, let me find the code. I still have it somewhere 19:13 ventz this was the course I took: http://www.ccs.neu.edu/home/matthias/670-s07/ 19:14 ironcamel ventz: how does python allow object protection from read/write/change? 19:14 ironcamel also, if you are implementing a web server from scratch, you are doing something really wrong, imo 19:15 ventz ironcamel: not web server per say - it was a custom app server for a game we were building 19:15 ironcamel ah, ok 19:15 ventz that was what was so twisted about this class 19:15 ventz everything was as "generic as possible" with something that was as custom as possible 19:16 ventz ah, here it was: 19:16 ventz http://www.ccs.neu.edu/home/matthias/670-s07/Projects/13.html 19:16 ventz ironcamel: ^^ see for server part 19:16 ventz we had 7 days to write that 19:17 ventz server, client, and glue in between 19:18 ironcamel Looks like a blast. If you didn't have to get it done in 7 days. 19:58 westmaas ironcamel: do you highlight on perl? 19:59 ironcamel westmaas: do you hilight on ironcamel? 19:59 ironcamel westmaas: yes, i do :) 19:59 ironcamel if you ever need me, you know what to yell 19:59 westmaas perlcamel 19:59 pdurbin westmaas: openstack-nova-2012.2 is what i want, right? 20:00 westmaas you doing essex? 20:00 westmaas if so then yes, I believe thats correct 20:00 westmaas like I said I don't do major releases :) 20:00 pdurbin i thought 2012.2 was the latest 20:01 pdurbin crimsonfubot: lucky openstack release naming 20:01 westmaas oh yeah 20:01 westmaas sorry 20:01 crimsonfubot pdurbin: http://wiki.openstack.org/ReleaseNaming 20:01 westmaas I got confused 20:01 westmaas it is 20:01 pdurbin westmaas: you're killing me ;) 20:02 pdurbin cool. so with vagrant i can go from base centos to openstack-nova-2012.2 installed in less than 7 minutes 20:11 pdurbin larsks|away: you linked to https://github.com/puppetlabs/puppetlabs-openstack at http://irclog.perlgeek.de/crimsonfu/2012-10-30#i_6113435 but does it support centos? it looks like not.. 20:21 pdurbin ok, openstackers... here's my proof of concept for OpenStack on CentOS managed with Puppet on Vagrant - https://github.com/pdurbin/openstack-vagrant-centos-puppet 20:34 pdurbin please note that it doesn't actually do anything yet, per the readme :) 20:34 shuff that's fine 20:34 pdurbin except install folsom nova 20:34 pdurbin shuff: try it out! vagrant up! 20:35 shuff mostly what i want right now is the ability to say "hey, we already deployed CIAC, why don't you look at what the RC guys are doing with OpenStack" 20:35 pdurbin shuff: you know who's way ahead of us? larsks 20:35 pdurbin crimsonfubot: lucky harvard ciac 20:35 crimsonfubot pdurbin: http://www.asian-studies.org/grants/main.htm 20:36 pdurbin shuff: CIAC? 20:41 shuff Cisco Intelligent Automation for Cloud 20:41 shuff http://blogs.cisco.com/tag/ciac/ 20:41 shuff Cisco's private cloud offering 21:21 comptona pdurbin: I'm running OpenStack on CentOS 6 using the puppetlabs-openstack modules 21:21 comptona (just got around to scrollback) 23:19 magoo pdurbin: yup, proud vagrant user 23:25 pdurbin magoo: that's awesome. i just got the vagrant religion a week ago. love it 23:26 pdurbin comptona: thanks much! tweeted at you :) https://twitter.com/philipdurbin/status/264117601100238849 23:26 magoo it has been awesome for us. I got Crowdtilt on it from the beginning so our dev process has been simplified from the getgo 23:27 magoo pdurbin: do you use chef? 23:28 comptona pdurbin: =) 23:28 comptona it took me a while to work out all the kinks, hit me up if you have any questions 23:29 pdurbin comptona: so! what you *could* do... if you're into it... is fork my repo! and add a working config! :) 23:29 pdurbin comptona: this one - https://github.com/pdurbin/openstack-vagrant-centos-puppet 23:30 comptona pretty sure I can't contribute back any code since it's $work 23:30 pdurbin magoo: right? that makes sense, right? since it's vagrant-based 23:30 comptona but they can't stop me from talking! =) 23:30 pdurbin comptona: ah. bummer. that's fine. i'll figure it out 23:31 magoo i didn't understand that at all haha 23:31 pdurbin comptona: this example thing looks promising... will probably try that next: https://github.com/puppetlabs/puppetlabs-openstack/blob/master/examples/site.pp 23:31 magoo but sure :P 23:31 pdurbin magoo: ask ironcamel. he used to work on openstack :) 23:32 comptona pdurbin: yeah, the puppetlabs openstack modules worked great for me 23:32 pdurbin and no, we don't use chef, but to agoddard's dismay 23:32 comptona set up keystone and added service accounts, all that good stuff 23:32 pdurbin comptona: awesome. seems the horse to bet on 23:32 comptona I had to make a few changes that I'm working on getting permission to contribute, but mostly template stuff 23:33 comptona like changing the LDAP template so I could auth against active directory 23:33 magoo pdurbin: you work for Rackspace? 23:34 pdurbin magoo: nope. harvard 23:36 magoo gotcha, must have forgot what room I was in ;) 23:37 semiosis not everyone here is from harvard though 23:37 pdurbin comptona: cool. now that i look at this thing i realize i'm a little lost puppet-wise. i was expecting just a "modules" directory 23:37 semiosis magoo: in fact, there are some rackspace people here who aren't on the east coast at all! 23:38 pdurbin comptona: i have a "modules" directory here anyway... and it's what i'm used to from our puppet at work - https://github.com/pdurbin/greptilian-vagrant 23:39 pdurbin sjoeboo: and you, our puppet master, are on vacation. hope you're having fun! :) 23:40 magoo It's a small crowd out here in Blacksburg so I'm not surprised 23:40 pdurbin comptona: see, larsks's repo looks more the way how ours is organized: https://code.seas.harvard.edu/puppet/puppet/trees/develop/modules/openstack 23:41 comptona I'm sorry, I don't understand what you're asking me 23:41 pdurbin comptona: heh. all good. all good :) 23:42 pdurbin just tell me to go to puppet camp 23:42 comptona no, I know puppet pretty well so I can answer questions about it, I just don't follow :) 23:42 semiosis pdurbin: usually top level would have manifests/ and modules/ and possibly also templates/ 23:44 pdurbin so can i make the puppetlabs-openstack repo a git submodule of my repo? 23:44 comptona ah, I see now 23:44 pdurbin i do that with the epel repo: https://github.com/pdurbin/greptilian-vagrant/tree/master/modules 23:44 pdurbin er. on the relavant repo too: https://github.com/pdurbin/openstack-vagrant-centos-puppet/tree/master/modules 23:44 comptona you could do it that way, or you could use the "puppet module install" command to grab that puppet module and all its dependencies 23:45 comptona the puppetlabs openstack stuff is broken out into a separate repo/puppet module for each service (keystone, swift, nova, etc.), so there's a bunch of dependencies 23:46 comptona what I did was ran "puppet module install puppetlabs-openstack" to get the files I needed, and then committed them into a separate "modules_external" directory that I then added to my modulepath in puppet.conf 23:47 comptona that's how I handle all external modules; "modules/" is just for the stuff we write ourselves 23:48 pdurbin comptona: ok. interesting. thanks 23:49 pdurbin i think i'd like to use git submodules if i can... at least in this demo repo... like i do with epel 23:49 comptona sure, I imagine that would work still 23:49 pdurbin i set puppet.module_path = "modules" in my Vagrantfile: https://github.com/pdurbin/openstack-vagrant-centos-puppet/blob/master/Vagrantfile 23:52 pdurbin semiosis: right. that's what i'm doing. top level manifests and modules 23:53 pdurbin "To install the released versions from the forge: puppet module install puppetlabs-openstack" 23:54 pdurbin i guess i'll try that 23:54 pdurbin oh wait. hmm. git instructions right below that. "rake modules:clone" 23:56 pdurbin uh oh. kids crying. and they need a bath