[03:04] <Bluefoxicy> Anyone any good with ruby?
[03:04] <Bluefoxicy> Particularly broken ruby stuff?
[13:53] <bluefoxxx> anyone know anything about ruby?
[13:56] <menace> no one ever heard of this thing called ruby.. is it something to eat?
[14:04] <Adri2000> anyone in ~sru could review bug #1048634 ? (waiting for one week)
[14:15] <xnox> How to "add" bug links to the FTBFS rebuilds report? http://people.ubuntuwire.org/~wgrant/rebuild-ftbfs-test/test-rebuild-20121221-raring.html
[14:25] <Adri2000> xnox: tag ftbfs? (just guessing)
[14:25] <Adri2000> I mean, tag the bug in LP with 'ftbfs'
[14:29] <xnox> Adri2000: yeah, looks like it http://people.ubuntuwire.org/~wgrant/rebuild-ftbfs-test/source/build_status.py
[15:52] <hippiehacker> where does the embedded grub config for efi/boot/bootx64.efi come from on the release isos?
[16:40] <hippiehacker> running strings on it shows me what is compiled in
[16:41] <bluefoxxx> Apache still works with a blank /etc/ports.conf
[16:41] <bluefoxxx> in fact, if you remove the default site and add your own, apache2 continues to work, possibly listening on another port
[16:42] <bluefoxxx> with /etc/apache2/ports.conf as-is, it also listens (uselessly) on 80 and 443
[16:42] <bluefoxxx> should I file this as a bug?
[17:19] <FlowRiser> hey all, so i've just made my very own greeter using the qt bindings of lightdm. Now i want to package it, so other users can test it out; I read the documentation and i got the basic concepts down. Can someone explain slowly what packaging is, specifically what are the steps in order to get from source code to a package ?
[17:25] <geofft> FlowRiser: briefly you add a debian/ directory with the appropriate metadata, then run a tool to build a .deb out of that, and usually also a source package
[17:26] <FlowRiser> what does the debian metadata usually contain ?
[17:27] <FlowRiser> geofft, thanks so much for answering, i've been trying to get someone to give me a brief explanation for 3 days now
[17:29] <jtaylor> FlowRiser: http://developer.ubuntu.com/packaging/html/debian-dir-overview.html
[17:29] <FlowRiser> jtaylor, thnaks :D
[17:34] <geofft> there's lots of good documentation. packaging is complicated, in part because it's a powerful system
[17:39] <bluefoxxx> FlowRiser, look into CDBS for debian packaging made easy
[17:39] <bluefoxxx> unless that's been superceded already
[17:39] <bluefoxxx> for a while Debian package management was a land of new tools similar in spirit to Japan's take on Doritos.
[17:39] <jtaylor> don't look into cdbs ._.
[17:40] <geofft> I think the standard recommendation these days is new-style debhelper over CDBS
[17:40] <bluefoxxx> jtaylor, superceded then I guess?
[17:40] <bluefoxxx> ah
[17:40] <bluefoxxx> see this is what I mean.
[17:40] <geofft> (CDBS definitely made sense over old-style debhelper though)
[17:40] <jtaylor> its still used, but not recommended
[17:40] <bluefoxxx> it's like in Japan there are new flavors of doritos twice a week
[17:41] <geofft> yeah, part of why Debian packaging is complicated is because there's still a lot of work to figure out the best way to do it.
[17:41] <geofft> see also source format 3.0. These sorts of things _shouldn't_ be things a new packager has to really care about, but they still show up
[17:41] <bluefoxxx> geofft, there's also legacy from all the stuff that almost made it but sharply fell into obscurity
[17:42] <bluefoxxx> like delta-RPMs on RHEL
[17:42] <bluefoxxx> there were a few UDS where one of the main topics was Delta-debs
[17:42] <jtaylor> delta-debs work fine
[17:42] <jtaylor> (in debian)
[17:42] <bluefoxxx> oh
[17:42] <bluefoxxx> at some point like a year later someone told me the whole idea was stupid and pointless because "Internet is fast now"
[17:42] <bluefoxxx> apparently they changed their mind again.
[17:42] <jtaylor> thats not true
[17:43] <jtaylor> I love debians delta debs
[17:43] <jtaylor> many users do not have fast internet
[17:44] <bluefoxxx> what's the current accepted way to build a local caching server?
[17:44] <bluefoxxx> I'm using apt-cacher here because it works
[17:44] <bluefoxxx> someone said use http_proxy, and there's like 4 servers specifically designed to cache packages for apt in Universe
[17:44] <jtaylor> I'm using apt-cacher-ng
[17:45] <jtaylor> with http_proxy for chroots and apt config proxy for the main machine
[17:45] <bluefoxxx> yeah oh.
[17:45] <bluefoxxx> I put my proxy config in /etc/apt/apt.conf.d/01proxy
[17:46] <bluefoxxx> the installer (for mini.iso at least, for 12.10/Quantal) puts it in /etc/apt/apt.conf
[17:46] <bluefoxxx> it's the only thing that goes in apt.conf
[17:47] <bluefoxxx> jtaylor, I'm doing enterprise linux system administration these days, so I've been looking at ... certain things.  Notably puppet.
[17:47] <bluefoxxx> the concept of a WSUS-alike for Linux has been on my list of things needed for Enterprise for a while though
[17:48] <bluefoxxx> WSUS is Microsoft's server for caching Windows Update patches locally, so you only downtoald them once per LAN.
[17:48] <bluefoxxx> Microsoft's tools also allow you to control which patches go to which systems, though that can be done through other means on Linux.  Not sure if there's a *great* way to do that yet though.
[17:50] <bluefoxxx> in any case, all babbling aside, at some point I need to put together a list of "must haves" and start pestering people on the mailing lists, filing bugs, figuring out how to fix things and submit debdiffs, and put in blueprints and somehow convince people that certain things are significant
[17:51] <bluefoxxx> (I figure one day I should make a contribution to society, once)
[18:03] <JanC> bluefoxxx: I'm sure several companies are using something like WSUS with Ubuntu/Debian right now
[18:03] <bluefoxxx> JanC, yeah the question is how easy is it to find/set up
[18:03] <bluefoxxx> in the case of apt-cacher-ng, that works fairly well
[18:03] <bluefoxxx> though, it's not in main.
[18:04] <JanC> "not in main" is not really all that important
[18:04] <bluefoxxx> You also can't get a high level view of all the Debian/RHEL/etc systems on your network, arrange them in groups, and set which packages (and which versions!) they get (=centralized package management), which is a major function of SCCM
[18:05] <bluefoxxx> JanC, being in main is mainly symbolic
[18:05] <bluefoxxx> except for maybe puppetmaster-passenger
[18:05] <bluefoxxx> which is both universe and horribly broken (although puppetmaster is main)
[18:06] <bluefoxxx> and we're back to me not knowing enough about the crap I'm working with, 'cause if I did I'd have fixed it by now :|
[18:08] <JanC> that's a common issue with open source solutions  ;)
[18:09] <JanC> anyway, there are some solutions that are used by large deployments, and those are probably able to do what you want
[18:11] <JanC> e.g. FAI & GOsa
[18:17] <andrei_> Hello, I would like to become an ubuntu contributor. So far I have read ContributeToUbuntu and registered with Ubuntu in Launchpad. I'm having a bit of trouble searching for available projects? Also, is it possible or should I join a specific team? Thank you!
[18:23] <bluefoxxx> JanC,  been looking into FAI.  GoSa dunno yet.
[18:24] <bluefoxxx> JanC, also see Foreman
[18:24] <bluefoxxx> Though my senior coworkers are TERRIFIED of Foreman
[18:24] <JanC> I don't know Foreman
[18:25] <bluefoxxx> http://theforeman.org/projects/foreman/wiki/Screencasts Third video says it all
[18:25] <JanC> but FAI, GoSa & co. are used in Germany by some really huge deployments
[18:25] <bluefoxxx> It'll automatically push a PXE at a piece of bare hardware, or set up a virtual machine
[18:25] <bluefoxxx> configure it, create a Puppet node for it, install Puppet, and push all the configuration settings
[18:25] <JanC> OTOH, I don't know how much manual configuration they need
[18:25] <bluefoxxx> Suddenly:  brand new server :)
[18:26] <bluefoxxx> There's something slightly unnerving about software that can wipe and re-install a server or mess with your VM provisioning without a human having to physically push a big red button
[18:26] <JanC> bluefoxxx: that sounds like it somewhat overlaps with the Ubuntu voodoo stuff ☺
[18:27] <bluefoxxx> Voodoo is another thing I have to look into I guess
[18:33] <JanC> I mean MAAS & Juju & such
[18:35] <penguin42> bluefoxxx: Generally it's much worse if it does that to hundreds of servers at one click :-)
[18:44] <maswan> bluefoxxx: fyi: we do fairly large fai deployments on ubuntu for hpc use at $WORK: fai-chboot -e $host on the install server, then ssh $host reboot, and it gets reinstalled from scratch
[18:46] <bluefoxxx> nice
[18:47] <bluefoxxx> maswan, we run around and put CDs into VMware virtual drives, then manually configure everything here
[18:47] <bluefoxxx> mostly on CentOS.
[18:47] <bluefoxxx> the RPM database keeps breaking and I have to rebuild it and flush the yum cache
[18:48] <maswan> bluefoxxx: ugh. we have a bit of centos too, due to political stuff, I forget how that's setup but I think it is minimal kickstart + FAI
[18:48] <bluefoxxx> maswan, we can skip the argument about which distro is better
[18:48] <bluefoxxx> It's sufficient to say RHEL and its derivatives are simply OLD
[18:48] <maswan> bluefoxxx: My point was that we use FAi for that too
[18:48] <bluefoxxx> and the EL package availability is very narrow scope
[18:49] <maswan> (but a different tree of fai nfsroot/config/etc)
[18:50] <bluefoxxx> a colleague had indicated that FAI was more a Debian thing and Kickstart was more of a RH thing
[18:50] <maswan> yeah, it is
[18:51] <maswan> But since we came from the debian/ubuntu side and then added some RH, we kept the kickstart for minimal OS and then did it FAI:y for package install and conf
[18:51] <maswan> It is quite a step to setup the first time, but it is really good to have automatically deployable intsances once you scale up a bit.
[18:51] <maswan> For that it doesn't really matter which of the deployment solutions you choose
[18:52] <bluefoxxx> yeah that's what I"m getting with puppet
[18:52] <bluefoxxx> I"ve written like 400 lines of puppet code
[18:52] <bluefoxxx> just for a configuration module for puppet
[18:52] <maswan> Yeah, that works too.
[18:53] <bluefoxxx> granted, it configures the client and server with various options, is supposed to port to debianites/rhelites, lets you use the distro repositories or the puppetlabs stuff, etc.
[18:54] <bluefoxxx> well, I'll have to compare FAI with the Foreman and see if anything is appropriate here
[18:55] <bluefoxxx> maswan, so many options.  I didn't put that much thought into it before :|
[18:56] <maswan> bluefoxxx: Yeah, it was easier back when we choose FAI, it was the only debian:y solution out there at, hmm, woody I tihnk it was. :)
[18:57] <bluefoxxx> My ideal scenario is a distribution which handles all the "getting set up" part for you.  Something that will easily put in a running FAI/Kickstart/Foreman/etc, Puppet/Chef/Ansible/etc, deploy a caching server for Deb and RPM, patch management, domains/LDAP (hey, Samba 4!) for authentication, integration with existing network environment (active directory server -> authentication)...
[18:57] <bluefoxxx> without doing 10 hours of configuration just to get to the point where you can start configuring the software to do something besides just basically idle
[18:58] <bluefoxxx> but then, I am a lazy mook
[18:58] <bluefoxxx> Thing is there's apparently a ton of ways to do this :|
[18:58] <bluefoxxx> suddenly, not so simple a problem
[18:59] <maswan> Yeah. And new ones are frequently invented too.
[19:02] <penguin42> bluefoxxx: Isn't this what preseeding is supposed to be about?
[19:03] <bluefoxxx> ... ten million ways to do this
[19:03] <penguin42> yep
[19:09] <andrei_> Sorry for the rather newbie message but, I have been trying to find a project(blueprint as it is called) to work on and I'm not sure even how to find one on which I can work.
[19:09] <andrei_> I have never used launchbad before.
[19:12] <bluefoxxx> andrei_,  put in google:  blueprints for ubuntu 13.04
[19:13] <bluefoxxx> the first result is https://blueprints.launchpad.net/ubuntu/raring ... not comprehensive by far, but a good place to start :)
[19:14] <andrei_> bluefoxxx: thank you
[19:22] <andrei_> if I decide to work on a project, what is the next step ? ( assuming it has an assignee )
[19:28] <bluefoxxx> I don't know.  E-mail the person it's assigned to and ask.  They probably know what they're doing.